American Society of Cytopathology 58th Annual Scientific Meeting Platform and Poster Presentations

Supplement to Cancer Cytopathology

PP #1

Loss of P16 Protein Expression as a Biomarker for Malignant Mesothelioma

Hakan Aydin, MD, Deborah Chute, MD, Bin Yang, MD, PhD

Anatomic Pathology, Cleveland Clinic, Cleveland, Ohio

Introduction: The cytologic diagnosis of malignant mesothelioma (MM) in serous effusions is challenging due to significant morphologic overlap with reactive mesothelial cells. Numerous immunohistochemical markers are utilized to discriminate mesothelial cells from adenocarcinoma, but few ancillary tools are available to discriminate between benign and malignant mesothelial proliferations. Recent studies indicate that alterations of the p16 gene, either by deletion or promoter hypermethylation, are associated with malignant transformation of mesothelial cells. However, the differences in p16 protein expression between benign and malignant mesothelial cells have not been well investigated. We investigated the utility of p16 immunohistochemical staining in differentiating benign from malignant mesothelial proliferations both in surgical and cytologic specimens.

Materials and Methods: Forty-four cases of mesothelial proliferation were retrospectively identified from our hospital archives and were included. Twelve cases were benign peritoneal mesothelial inclusion cysts. Thirty-two cases were malignant mesotheliomas, encompassing 18 surgically resected specimens and 14 cell blocks from pleural or peritoneal effusions. Expression of p16 was evaluated immunohistochemically with the p16 antibody kit (MTM, Germany) using the BenchmarkXT® instrument (Ventana Medical Systems, Inc., Tucson, AZ). The results were evaluated according to staining intensity (0-negative, 1+ weak, 2+ moderate, and 3+ strong) and percentage of positive mesothelial cells.

Results: Among 13 cases of benign mesothelial proliferation, strong (3+) and diffuse (>95% of cells) p16 immunoreactivity was seen in all cases. In contrast, loss of expression of p16 protein was seen in all 32 cases of malignant mesothelioma, including 78.2% (25/32) cases with total p16 negativity and 21.8% (7/32) cases with partial p16 immunoreactivity (20-50% of cells, 2-3+). The difference in p16 immunoreactivity between benign/reactive mesothelial proliferations and malignant mesotheliomas was statistically significant (p<0.001).

Conclusions: Strong expression of p16 is seen in benign/reactive mesothelial proliferations. In contrast, loss of p16 protein is a hallmark of malignant mesothelioma cells. Immunohistochemical staining with p16 can be a useful ancillary tool in differentiating benign and malignant mesothelial proliferations both in surgical and cytologic specimens.

PP #2

High Sensitivity and High Specificity of P16/Ki-67 Dual-Stained Cytology for High-Grade CIN - Results from Screening and Triage Trials in Over 28,000 Women

Christine Bergeron, MD¹, Dietmar Schmidt, MD², Hans Ikenberg, MD, PhD³, Ruediger Ridder, PhD⁴

¹Laboratoire Cerba, Cergy Pontoise, France; ²Institute for Pathology, Mannheim, Germany; ³Cytomol, Frankfurt, Germany; ⁴mtm Laboratories, Heidelberg, Germany

Introduction: The detection of p16 over-expression is a sensitive indicator of cell-cycle deregulation and thus for the presence of cervical dysplasia. Simultaneous co-detection of the Ki-67 proliferation marker within the same cell may enhance the specificity of the test and thus further improve the diagnostic performance of cervical cytology. We investigated a new immuno-cytochemical dual staining approach for p16/Ki-67 biomarker co-detection in cervical cytology specimens in a prospective, pan-European screening study and in a large cohort of retrospectively collected ASC-US and LSIL cases. Sensitivity and specificity of this dual-stained cytology test for detection of underlying high-grade CIN (CIN2+) was compared to Pap cytology (screening) and to HPV testing (screening and ASC-US/LSIL triage).

Materials and Methods: For the prospective screening trial, 27,349 women attending routine cervical cancer screening visits were enrolled. Pap, HPV, and dual-stained cytology testing was performed, and all women with any positive test result (except for HPV test positivity in women aged <30 years as the only positive test) were referred to colposcopy/biopsy follow-up. For the retrospective triage trial, residual material from ThinPrep LBC vials (362 ASC-US, 415 LSIL cases) was used to investigate the p16/Ki-67 dual staining. The presence of one or more dual-stained cervical epithelial cell(s) defined a positive test result, independent from morphology interpretation. Pathologist majority consensus diagnoses on biopsy served as gold standard for both studies.

Results: In screening, sensitivity of p16/Ki-67 dual-stained cytology for HGCIN was found significantly higher (90.1%) than Pap cytology (66.4%; p<0.0005). Specificity was identical for both tests (95.3% vs. 95.4%, respectively). The sensitivity of HPV testing was 96.4%, but with a substantially lower specificity (90.2% over all ages) as compared to the cytology based tests. In women aged <30 years, specificity of dual-stained cytology was 92.3% compared to 81.4% for HPV testing. In addition, dual-stained cytology showed high sensitivity in both the ASC-US and LSIL cases of the screening population as well as in the retrospective study population, while reducing the number of false-positive results by approx. 50% vs. HPV testing in both studies.

Conclusions: Dual-stained cytology testing was shown to combine both high sensitivity and high specificity for the detection of women with underlying CIN2+ in primary screening, and in the triage of Pap cytology results categorized as ASC-US or LSIL. In addition to the potential for improving current screening and ASC-US triage algorithms, this new test may for the first time provide a triage option for LSIL cytology.

PP #3

Morphologic Features which Affect Validation and Proficiency Test Performance of Biopsy-Proven HSIL Pap Tests

Carol Filomena, MD, Jamie Covell, CT(ASCP), Rhonda Metzler, CT(ASCP), Robert Goulart, MD

GYN PT and Assessment Committee, American Society for Clinical Pathology, Chicago, Illinois

Introduction: Before Pap tests are entered into the American Society for Clinical Pathology (ASCP) GYN Proficiency Test (PT) program, they are thoroughly evaluated and must attain a minimum performance of 90% target diagnosis agreement after 40 individual reads and biopsy confirmation of all SIL and cancer cases. Despite this vigorous process, a small subset of validated slides does not perform as well in actual testing. When the target diagnosis agreement level falls below 90% for any particular slide, it is immediately removed from the pool of active test slides. Similarly, a small subset is removed relatively late in the validation process (between 20-40 reads) when they fall below the minimum agreement requirement. Comparison of the morphologic characteristics of “high” vs. “low” performing slides could provide information concerning subtle differences that result in less reproducible interpretations. The target diagnostic category of HSIL was chosen for study as it is both clinically significant and relatively common.

Materials and Methods: Two HSIL ThinPrep® Pap test groups were identified, each 30 cases (slides). The first group consisted of high performers (HP) with ≥90% target diagnosis agreement throughout the validation process and in active testing. The second consisted of low performers (LP) with <77% agreement which failed validation or were soon removed from the active test pool. Each slide was re-evaluated by five committee reviewers in a blinded fashion. Slides were classified as “satisfactory for evaluation (SAT)” if they were technically adequate and met TBS criteria. Review characteristics were: number and predominant HSIL cell pattern (groups vs. single cells); relative number of LSIL cells; nuclear membrane irregularity, hyperchromasia, nuclear chromatin granularity, N/C ratio, nucleoli, and the presence/absence of atrophy, atypical keratinized squamous cells, inflammation, organisms (Trichomonas and Candida), and endometrial cells. Reviewers were also instructed to note which slides they perceived might perform poorly. A consensus interpretation (majority) was assigned for each category.

Results: All 60 slides were considered SAT with the exception of one slide in the LP group. No slide contained significant atrophy. The only two HSIL cell characteristics that demonstrated differences between the HP and LP groups were: 1) degree of cellularity and 2) presence of nucleoli. Thirty-five of 60 slides (58%) had >10 HSIL cell groups. Of these, 24 (69%) were in the HP group vs. 11 (31%) in the LP group. Twenty-eight of 60 slides (47%) had greater than 20 single HSIL cells. Of these, 20 (71%) were in the HP group versus 8 (29%) in the LP cohort. Nucleoli were identified in six (10%) of cases, all in the HP group. In regard to the prediction of slide performance, a total of 17 (28%) cases were predicted to perform poorly. Of these, a high majority 15 (88%) were in the LP group.

Conclusions: An increased quantity of HSIL cells, either in cell groups or as single cells, correlated with completion of the validation process and high field PT performance (reproducibility of HSIL target diagnosis). A high level of performance was also seen in the slides with nucleoli present in the majority of HSIL cells, albeit an infrequent finding. Although only one-half of the 30 low performing cases could be predicted by the reviewers, designation as such was highly accurate.

PP #4

Effects of Increasing Cytotechnologist Workload Using the Location Guided Imaging System FocalPoint GS on SurePath Pap Tests

Angelique Levi, MD¹, Kevin Schofield, BS, CT¹, Tarik Elsheikh, MD², Malini Harigopal, MD¹, David Chhieng, MD¹

¹Pathology, Yale University, New Haven, Connecticut; ²Pathology, Ball Memorial Hospital, Muncie, Indiana

Introduction: Recently, the FDA has approved the use of the location-guided imaging system, FocalPoint GS (Becton Dickinson, Burlington, NC) on SurePath Pap tests (Becton Dickinson) for primary screening. An important part of the FDA approval was the acceptance of a higher screening limit of 200 slides per day compared to the 100 slides per day limit for manual liquid-based Pap test screening. Studies examining the effect of increasing workload on the performance of individual cytotechnologists (CTs) are limited, and to our knowledge have not been done previously using FocalPoint GS.

Materials and Methods: Using FocalPoint GS, the performance of 3 CTs with varying levels of experience were evaluated. They CTs were not aware of the study design of the 3 phases, but were informed it was a study assessing screening productivity. The study consists of 3 phases, each phase lasting 10 days, and each day consists of 7.5 hours of screening. In phase I, the CTs are asked to screen at their usual pace. In phase II, the CTs are asked to screen as fast as they can without diminishing the quality of their work. In phase III, the CTs were asked to screen 10% more than their daily workload from phase II. For each slide, 10 fields of view (FOVs) selected by the imaging system are screened, and if any cellular abnormalities are present, even if reactive changes are favored, a full manual review of the slide is performed. All Pap tests underwent 100% rescreening by the remaining CTs in the laboratory, who were not involved in the study. Data is collected from all 3 CTs from phase I of the study and for 1 CT from phase II of the study. Results from the remaining CTs for phase II and all CTs for phase III are incomplete at this time.

Results: The total slides screened by each CT in phase I is as follows: CT 1, 823; CT 2, 816; CT 3, 665. The total number of slides screened by CT 1 in phase II is 940. Comparing phase I (all 3 CTs) to phase II (CT 1only), the average daily productivity (# of slides screened per day) increased from 76 slides per day to 94 slides/day (24% increase), and the hourly screening rates increased from 10 slides per hour to 12.5 slides per hour (25% increase). The average screening hours per day was 7.5. The average minutes per slide in phase I was 5.9 (CT 1, 5.4; CT 2, 5.5; CT 3, 6.7) compared to phase II was 4.8 (CT 1). The detection rate of abnormal Pap tests (defined as ASCUS or higher) for each CT in phase I is as follows: CT 1, 13.6%; CT 2, 10.0%; CT 3, 10.4%. The abnormal detection rate for CT 1 in phase II is 13.5%.

Conclusions: The preliminary results of this study indicate that an increased average CT workload (phase I compared to phase II) with FocalPoint GS appears to have been accomplished mostly by a reduction in the amount of time spent reviewing 10 FOVs, and likely a reduction in the percentage of cases that underwent full manual review (data still pending). The abnormal detection rate decreased slightly from phase I to phase II for CT 1. This trend may be more apparent when the phase I to phase III data from all 3 CTs is compared. Additional CT screening data to be analyzed include false negative fraction and discrepancy rates in each phase of the study. As the average CT workload increases to the projected number of greater than 100 slides per day with Focal Point GS in phase III of this study, CT screening performance may change in statistically significant way.

PP #5

Performance of Direct Smears Compared to Liquid-Based Preparations of Thyroid FNA Samples: Analysis of 47,076 Responses in the College of American Pathologists Interlaboratory Comparison Program in Non-Gynecologic Cytopathology

Andrew Fischer, MD¹, Amy Clayton, MD², Joel Bentz, MD³, Patricia Wasserman, MD⁴, Michael Henry, MD², Rhona Souers, MS⁵, Ritu Nayar, MD⁷, Ann Moriarty, MD⁶

¹Pathology, University of Massachusetts, Worcester, Massachusetts; ²Division Anatomic Pathology, Mayo Clinic, Rochester, Minnesota; ³Laboratory Medicine Consultants, Ltd, Las Vegas, Nevada; ⁴Pathology, Long Island Jewish Medical Center North Shore, New Hyde Park, New York; ⁵Statistics, College of American Pathologists, Northfield, Illinois; ⁶Pathology, AmeriPath Indiana, Indianapolis, Indiana; ⁷Pathology, Northwestern University, Chicago, Illinois

Introduction: Controversy exists whether thyroid fine needle aspirate (FNA) are best processed with direct smears or with liquid-based preparations (LBP). We compared the performance of direct smears and LBP of thyroid FNA slides circulated in the College of American Pathologists Interlaboratory Comparison Program in Non-Gynecologic Cytopathology (CAP NGC). Problem areas for participants were identified for both preparations.

Materials and Methods: A database of participant responses for thyroid FNA slides prepared by LBP (Papanicolaou-stained ThinPrep, SurePath, and cytospins) or Papanicolaou-stained alcohol fixed smears with 10 or more responses from at least three different labs was analyzed. Discrepancies between the reference diagnosis and participant response were calculated at the level of the following three general categories: Negative (including specific reference diagnosis of benign/goiter and thyroiditis), Suspicious (including specific reference diagnosis follicular and Hurthle cell neoplasm), and Malignant. For the specific reference diagnoses of benign/goiter and papillary thyroid carcinoma, participant specific diagnoses were tabulated.

Results:47,076 thyroid FNA responses between 2001 and 2009 included 44,478 for direct smears and 2,598 for LBP. As shown in Table 1, responses to the general category of Negative were discrepant in 14.9% of direct smears compared to 5.9% for LBP (p<.001). Responses to the category of Suspicious had a high but comparable discrepancy rate for direct smears and LBP (37.1% vs 41.1; p=.22). Responses to the category of Malignant had a discrepancy rate of 7.3% for direct smears compared to 14.7% for LBP (p<.001). Cytotechnologists had a higher discrepancy rate than pathologists for all general categories (Table 2 and 3). Overall 7.8% of direct smears with a benign/goiter specific reference diagnosis were misdiagnosed as follicular neoplasms compared to 1.3% of LBP. Overall 7.2% of LBP with a specific reference diagnosis of papillary thyroid carcinoma were diagnosed as benign/goiter compared to 4.8% of direct smears.

Table 1. Discordant response rates by preparation type (total responses)

General diagnostic category	Smear number (%)	LBP number (%)	P-value
Negative	25,518 (14.9)	680 (5.9)	<.001
Suspicious	4,597 (37.1)	229 (41.1)	.22
Malignant	14,363 (7.3)	1,689 (14.7)	<.001

Table 2. Discordant response rates for cytotechnologists

General diagnostic category	Smear number (%)	LBP number (%)	P-value
Negative	8,092 (18.5)	184 (9.2)	<.001
Suspicious	1,418 (44.7)	56 (53.6)	.19
Malignant	4,249 (10.0)	527 (19.4)	<.001

Table 3. Discordant response rates for pathologists

General diagnostic category	Smears number (%)	LBP number (%)	P-value
Negative	12,279 (13.4)	342 (5.0)	<.001
Suspicious	2.289 (33.9)	123 (35.8)	.67
Malignant	7,121 (6.4)	811 (12.8)	<.001

Conclusions: In the CAP NGC Program, LBP performed better than direct smears for cases with a benign general category and a specific reference diagnosis of benign/goiter, but LBP performed worse than direct smears for cases with a malignant general category and a specific reference diagnosis of papillary thyroid carcinoma. Findings in poor performing slides are reviewed.

PP #6

Responding to Practice Changes: The Creation of a Cytotechnologist Career Ladder at Mayo Clinic

Angela Sorenson, MA, SCT(ASCP), Lisa Colborn, MBA, CT(ASCP), Renee Root, BS, CT(ASCP), Sally Nelson, BS, CT(ASCP), Johnita Ihrke, BS, CT(ASCP), Kevin Halling, MD, PhD, Michael Henry, MD, Amy Clayton, MD

Cytopathology, Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota

Introduction: Decreasing cervical PAP testing and increasing pathologist workload has created new opportunities for cytotechnologists (CTs) to participate in nontraditional roles within anatomic pathology practices. Little has been published describing how these roles function within the practice setting.

Materials and Methods: Personnel and practice records from Mayo Clinic (MC) Cytology and Molecular Cytology Laboratories were reviewed for a ten year period (2000-2010). Staffing levels, CT responsibilities, test volumes and requirements for the CT's expanded roles are summarized.

Results: In the year 2000, MC CTs routinely performed gynecologic and non gynecologic cytology screening, EUS fine needle aspiration (FNA) adequacy assessment and digital image analysis (DIA). Over the next decade CTs assumed additional responsibilities including: Fluorescence in situ hybridization (FISH) analysis, FNA screening, ER/PR/HER2 quantitative immunostain review, specimen problem solving, circulating tumor cell analysis, histologic tumor identification for microdissection, acid fast bacillus stain screening, novel cytology test development, education for new CT roles and quality monitoring. Non gynecologic specimen volumes increased from 16,481 to 20,101 (a 22% increase) and FNA volumes increased from 3,427 to 7,421 (a 116% increase) during the ten year period. The total number of other testing (FISH analysis, acid fast bacillus screening, etc.) performed by CTs in just 2009 was 24,034. While gynecologic cytology volumes have decreased by 59% during the same time period, the number of cytotechnologists employed in the laboratory increased from 24 to a maximum of 49 (42 currently) (Table 1). The increased complexity of CT management and testing responsibilities necessitated the creation of a career ladder within the MC Cytology and Molecular Cytology laboratories (Table 2).

Conclusions: As testing and management responsibilities for MC CTs has expanded, the development of a hierarchical career ladder has become necessary to establish a known standard of progression for CT's job responsibilities. The addition of 18 CTs in 10 year's time supports the attraction and career satisfaction by CTs into the expanded roles. The additional opportunities available to CTs are an attractive retention tool for existing employees looking to further their career and to new applicants who may not have these opportunities elsewhere. The experience at MC with the expanded CT role has been supported positively by the anatomic pathology practice.

Table 1. Cytotechnologist Positions over Time

Year
	2000	2001	2002	2003	2004	2005	2006	2007	2008	2009	2010
Position
Cytotechnologist	18	21	21	20	24	19	19	17	21	12	12
Senior Cytotechnologist	0	0	0	0	0	5	9	10	8	13	12
Lead	2	2	2	2	2	4	5	4	4	3	3
Specialist	1	1	1	2	3	4	5	6	7	7	7
Development Technologist	1	1	1	1	1	3	4	5	5	4	4
Assistant Supervisor	1	1	1	2	2	1	1	1	2	2	2
Supervisor	1	1	1	1	1	2	2	2	2	2	2
Total Certified CT's	24	27	27	28	33	38	45	45	49	43	42

Table 2. Description of Cytotechnologist Career Ladder Roles

Role	Description	Required CT Experience	Other Requirements
Cytotechnologist	Cytology screening; exposure to ancillary testing	CT certification
Senior Cytotechnologist	Screening; additional ancillary testing	3 years	Required to pass SCT(ASCP) or  MB(ASCP) exam
Lead	Troubleshooting; point person for CT's	3 years
Specialist with Education/Training Focus	Trains new CT's; provides education to CT's,  pathologists and residents	3 years	Required to pass SCT(ASCP) or  MB(ASCP) exam
Specialist with Pathologist assisting Focus	Prescreens FNA specimens; problem solving	3 years	Required to pass SCT(ASCP) or  MB(ASCP) exam
Specialist with Quality Focus	Tracks quality metrics; prepares lab for inspections	3 years	Required to pass SCT(ASCP) or  MB(ASCP) exam
Development Technologist	Develops new lab tests; authors publications	1-2 years	SCT(ASCP) or MB(ASCP) exam  preferred
Assistant Supervisor	Assists managing personnel	5 years
Supervisor	Manages lab workflows and personnel	6 years

PP #7

EGFR and KRAS Mutations in Cytology Specimens from Patients with Lung Carcinoma

Shahreen Billah, MD, John Stewart, MD, PhD, Gregg Staerkel, MD, Yun Gong, MD, Ming Guo, MD

Pathology, The University of Texas MD Anderson Cancer Center, Houston, Texas

Introduction: Molecular testing for EGFR or KRAS mutations for lung cancer patients has become the standard of care for personalized lung cancer therapy. However, our experience of molecular testing in cytology specimens from lung cancer patients is limited. The aim of this study is to examine our practice of molecular testing in cytology specimens from lung cancer patients by comparing the distribution of EGFR or KRAS mutations with cytology diagnoses.

Materials and Methods: From September 2009 to April 2010, we collected 209 cytology cases, including fine needle aspirations (180), pleural fluids (24), pericardial fluids (3), bronchoalveolar lavage (1) and bronchial washing (1), from patients with lung cancer at MD Anderson Cancer Center. Molecular testing with sequencing for mutation detection in EGFR exons 18, 19, 20, 21 and KRAS exons 12, 13 and 61 was performed at the same center. The patients consisted of 114 females and 95 males. The patient's ages ranged from 29 to 91 years with a mean age of 63 years. The frequencies of mutations of EGFR or KRAS were compared with the cytology diagnoses.

Results: Cytology diagnoses consisted of adenocarcinoma (108), non-small cell carcinoma (30), non-small cell carcinoma favor adenocarcinoma (12) or squamous carcinoma (10), squamous (21), poorly-differentiated carcinoma (5) and small cell carcinoma (23). With exclusion of small cell carcinoma, EGFR mutations were significantly higher in adenocarcinoma (29%, 29/100) than in non-adenocarcinoma (7%, 5/75 of cases) (P = 0.002). EGFR exons 19 and 21 were the most frequently mutated genes with significantly higher mutation rates in adenocarcinoma than those in non-adenocarcinoma (P =0.008 and 0.003, respectively). The frequencies of KRAS mutations were comparable between adenocarcinomas (26%, 26/102) and non-adenocarcinomas (21%, 15/72) with the most frequent mutations in KRAS exon 12. No mutations of EGFR or KRAS were observed in small cell carcinoma. Squamous carcinoma and non-small cell carcinoma favor squamous carcinoma showed low EGFR mutation (3%, 1/31) and no KRAS mutations. One case of poorly differentiated carcinoma showed a KRAS mutation. Mutations in KRAS exon 61 was rare and only observed in one case of non-adenocarcinoma. The rates of insufficient DNA for molecular testing were comparable between EGFR (5.9%) and KRAS (6.5%) testing.

Conclusions:

• 1
  The significantly higher frequencies of EGFR mutations, especially in the Exons 19 and 21 in adenocarcinomas and low EGFR mutations in squamous carcinoma in our study indicate a high consistency between cytology diagnoses and molecular testing results in lung cancer patients.
• 2
  Cytology specimens are useful for EGFR and KRAS molecular testing in lung cancer patients with a low specimen insufficiency, particularly when obtaining a core biopsy is not feasible or in the case of insufficient core biopsy sampling.

PP #8

Testing a Scoring System for Predicting Lymph Node Malignancy in Ultrasound Guided FNA Practice

Maoxin Wu, MD, PhD, XiaoYong Zheng, MD, PhD, Hua Chen, MD, PhD, David Burstein, MD

Pathology, Mount Sinai School of Medicine, New York, New York;

Introduction: Application of ultrasound guidance in fine needle aspiration (USGFNA) has enabled Cytopathologists to perform smaller and non-palpable lymph nodes (LN) on regular bases. Pre-FNA factors such as a patient's age, ratio of short to long axis diameter (S/L ratio), internal echogenicity and vascular pattern of a lymph node are reportedly able to predict the benign or malignant nature of a LN (1). This study is designed to test the formula “0.06 x (age) + 4.76 x (S/L ratio) + 2.15 x (internal echo) + 1.80 x (vascular pattern)” generated from the reference (1) as a scoring system for predicting lymph node malignancy in a cytopathologist operated USGFNA practice.

Materials and Methods:83 USGFNA reports of LNs issued between 7/1/2008 to 4/28/2010 were reviewed. Information regarding to patient's age, S/L ratio, internal echo and vascular pattern were collected. The scores were generated based on the formula above. The score of “7” was used as a cutoff for predicting benign (<7) and malignant (>7) LNs. FNA cytology diagnosis, flow cytometric analysis as well as subsequent surgical diagnosis in some cases are served as a gold standard for statistical analysis.

Results:Table 1- Among 83 USGFNA of LNs, 38 cases with scores greater than “7” were proven to be malignant by cytology alone or along with flow and/or subsequent surgical path diagnosis; 8 of them with score greater than “7” but the cytology alone or along with flow results were benign; 37 of them with scores less than “7” were all proven to be benign nodes by cytology alone or along with flow analysis. Thus, this scoring system may achieve 100% sensitivity, 82% specificity, 83% positive predictive value (PPV), 100% negative predictive value (NPV) and 90% accuracy.

Conclusions: This scoring system may serve as a complementary tool in determining how aggressive a FNA procedure should be, how a FNA sample of LN should be triaged for ancillary study and how closely a patient with lymphadenopathy should be followed up.

Table 1. 

USGFNA (N583)	Malignant Pathology	Benign Pathology	Statistics
Score >7 (malignant)	38 (TP)	8 (FP)	PPV (TP/TP1FP) 83%
Score <7 (benign)	0 (FN)	37 (TN)	NPV (TN/TN1FN) 100%
Statistics	Sensitivity (TP/TP1FN) 100%	Specificity (TN/TN1FP) 82%	Accuracy (TP 1 TN)/(TP 1FP1FN1TN) 90%

PP #9

Impact of Loss of Billing for Multiple Units of CPT Code 88172 on Cytology Practice

Swati Mehrotra, MD, Guliz Barkan, MD, Alia Salhadar, MD, Razan Wafai, MD, Umesh Kapur, MD, Eva Wojcik, MD

Pathology, Loyola University Medical Center, Maywood, Illinois

Introduction: The advantages of an onsite cytologist for effective triage of Fine Needle Aspiration (FNA) material are well documented in the literature. Cytopathologists bill for these services utilizing the CPT code 88172 which, in the past has been billed per “pass” rather than per FNA. The reimbursement provided by multiple units of this code is compensation for cytopathologists' time and clinical expertise. 88172 is also a measure of cytopathologists' work productivity being tied to 0.60 units of wRVUs as per the CAP guidelines. Following CMS transmittal, Version 15.3 of the NCCI Manual, effective Oct. 1, 2009 cytologists may only bill for one unit of 88172 per FNA regardless of the number of passes or time spent while performing immediate evaluation.

Materials and Methods: The current study was undertaken to calculate the probable monetary loss, as well as loss of work RVUs our cytology practice would incur in 2010, based on the data available from 2009. Ours is a university based cytology practice with over 1800 FNAs per year, almost all of which are triaged by an onsite cytologist. We were routinely billing for multiple units of 88172 for all radiologist or clinician performed FNA. Retrospective FNA data was reviewed for the year 2009 and the results are detailed below.

Results: There were a total of 1839 FNAs in 2009. 1466 FNAs were performed by radiologists or clinicians. A total of 4014 passes were documented for 1466 FNAs, hence CPT code 88172 was billed 4014 times by our practice. The compensation for one unit of 88172 will differ for different practices depending on geographic locations and payer mix. Taking an average reimbursement rate of $38.00 per unit of 88172, our physician group received $152532.00 for 4014 units of 88172 in 2009. The work RVUs provided by this code were 2408. Extrapolating this data to 2010 when we only bill for one unit of 88172 per FNA, our practice will incur a monetary loss of over $100000.00. More so, our cytopathologists will lose 1530 units of wRVU.

Conclusions: The above data is quite disturbing and brings into perspective what the cytology community is facing regarding optimal reimbursement for their services. While premier pathology organizations are working with CMS to remedy the situation, we as cytologists should be aware and offer our continual support for this cause. In the event that CMS fails to revert this decision, the cytology community has to take a unified stand whether we want to provide our expert services despite loss of compensation or not.

PP #10

Clinical Impact of Second Opinion in Thyroid Fine Needle Aspiration Cytology (FNAC): A Study of 922 Inter-Institutional Consultations

Jaya Bajaj, MD, MHA¹, Chiara Sugrue, MBA, MS, SCT(ASCP)¹, Nora Morgenstern, MD¹, Patricia Wasserman, MD, FCAP¹, Jason Wasserman, MD²

¹Pathology, Long Island Jewish Medical Center, New Hyde Park, New York; ²Jefferson Medical College, Philadelphia, Pennsylvania

Introduction: Inter-institutional consultation in pathology has shown to improve patient safety by detecting interpretive errors that may significantly impact patient care. A review of literature has shown that thyroid FNAC specimens are especially prone to a change in patient management based on second opinion. The aim of our study was to assess the magnitude of discrepancies and determine the impact of second opinion in thyroid FNAC on cases referred to our hospital over a 2-year period.

Materials and Methods: All thyroid FNAC cases referred to our center from January 2008 to December 2009 were reviewed. Original diagnoses and second opinion diagnoses were categorized as either no disagreement, minor disagreement, or major disagreement, the latter two defined as a 1-or 2-step deviation, respectively, on a scale of “non-diagnostic, benign, atypical, suspicious for neoplasm, suspicious for malignancy, and malignant.” The clinical impact of each disagreement was determined by pathologic and clinical follow-up via review of medical records.

Results: Out of 922 FNAC cases from outside institutions, there were 122 disagreements (13%), including 44 major and 78 minor. Follow-up revealed a change in management in 39/44 (89%) cases of major discrepancies, and in 38/78 (49%) of minor ones. From the perspective of fiscal analysis, which is influenced by the type of change in management, a change from surgical to medical management was seen in 33 cases, that from medical to surgical management in 29 cases, and 13 cases with a change in type of surgery (partial versus total thyroidectomy). The second opinion was supported on follow-up in 57% of major discrepancies, and the initial diagnosis was supported better in 7% cases. The remainder (36%) of major discrepancy cases did not undergo surgery, precluding tissue confirmation.

Conclusions: Discrepancies in diagnoses of FNAC specimens, particularly thyroid, may be attributable to inherent challenges in their interpretation. Mandatory review of outside pathology material prior to definitive treatment of referred patients is an important aspect of patient safety and patient care. Critics have alleged increased costs from this practice. However, placing a dollar value on a changed diagnosis is a formidable challenge. The cost avoidance from lost wages, potential surgical complications, life-long thyroid replacement therapy and its morbidity, and potential litigation is not easily quantified. Using a simplified model in which changed diagnoses and associated costs can be objectively measured, we estimate that second opinion of these 922 cases resulted in potential cost saving of $1,002,566 based on current Medicare codes. The details of these costs will be discussed in the presentation. A study involving mandatory re-review of prostate biopsies before radical prostatectomy also showed that second opinion pathology resulted in actual cost savings from reduced surgery. This excluded the indirect costs, which would vary notably from case to case. Our study indicates the need for a mandatory quality-control program of outside cytology slides, especially in thyroid where a cytology diagnosis is only confirmed by a major surgical excision. The role and additional cost of molecular diagnostic testing as a promising ancillary study will be addressed, along with a comprehensive literature review on studies on second opinion.

PP #11

A National Survey of Quality Assurance Practices in Gynecologic Cytology: Preliminary Results from a Pilot Study

Joseph Tworek, MD¹, Bruce Jones, MD², Michael Henry, MD³, Rhona Souers, MS⁴, Barbara Blond, MBA⁴

¹Pathology, St. Joseph Mercy Hospital, Ann Arbor, Michigan; ²Pathology, Henry Ford Health System, Detroit, Michigan; ³Pathology, Mayo Clinic, Rochester, Minnesota; ⁴College of American Pathologists, Northfield, Illinois

Introduction: The College of American Pathologist (CAP) was awarded a grant from the Center for Disease Control and Prevention (CDC) to develop an inventory of current practices in gynecologic cytology (GC) laboratories to standardize procedures for quality improvement. The CAP proposal includes a nation wide survey of GC laboratories to innumerate quality assurance (QA) practices to establish consensus guidelines for QA in gynecologic cytology. A web based discussion format and a consensus conference, both open to the public, will be utilized to establish consensus best practice guidelines for GC QA programs.

Materials and Methods: A survey of GC QA practices was developed and consists of 112 questions from 8 QA categories: 1) diagnostic rates (DR), 2) real time re-screen of initially negative Pap tests (RTR), 3) retrospective look back of NILM Pap tests after a HSIL diagnosis (RLBNH), 4) proficiency testing (PT), 5) Pap test and cervical biopsy correlation (PTCBC), 6) concurrence of diagnoses between cytotechnologist and pathologist (CBCP), 7) HPV rates (HPVR), and 8) turn around time (TAT). Across all 8 categories, labs were surveyed on how performance is tracked, how often, and what is done when a variance is identified. Nine labs were surveyed as a pilot program before send out nationally.

Results: Nine labs had volumes ranging from 1,284 to 59,000 Pap tests in 2009, mean (22,273). (1, DR): 8 of 9 labs followed DR. Labs ranked the following DR as most useful to monitor in descending order: HSIL, ASC-US:SIL ratio, and ASC-US. (2, RTR): 5 of 9 labs actively tracked upgrades from NILM to ASC-US and above on RTR by cytotechnologist. ASC-US was the most frequent upgrade diagnosis (106, total); then LSIL (8, total), ASC-H (3, total), and HSIL (3, total). (3, RLBNH): 7 of 9 labs actively track upgrades from NILM Pap tests on RLBNH. The most frequent upgrade diagnoses were ASC-H, ASC-US and LSIL. (4, PT): All labs actively monitor PT results. The most frequent remedial action for 1st time failure was to reenroll for retest. Almost all labs do not have a policy of remedial action for personnel who pass the exam but do not score 100%. (5, PTCBC): 5 of 9 labs actively summarize results from PTCBC, and 4 out 5 follow the percentage of positive Pap tests that correlate with biopsies. 5 out 9 labs have a written policy dealing with a cervical biopsy containing dysplasia and a current NILM Pap test. All 5 review previous NILM Pap tests when a cervical biopsy shows CIN2 and greater, and 3 review NILM Pap tests for CIN1 and greater. 3 labs also review UNSAT Pap tests, 2 labs ASC-US Pap tests and 1 lab reviews both ASC-H and LSIL Pap tests. (6, CCP): 3 of 7 labs that responded do not actively follow this monitor. (7, HPVR): Labs ranked the following as most useful to monitor by HPVR ASC-US (9), HSIL(6), and LSIL(4), but 5 out 7 respondents did not use HPVR to monitor trends in accuracy of diagnoses. (8, TAT): Lab expected TAT ranged from 2 to 14 days (median, 5 days). Most monitored TAT by percentile distribution within a certain TAT.

Conclusions: Preliminary results suggest a wide spectrum in tracking quality assurance measures. This may in part be a reflection of laboratory size and the number of personnel. Further participation is required to establish consensus best practice QA guidelines.

PP #12

Focused Rescreening of NILM Liquid Based Pap Slides from Women 30 Years Old or Greater with Positive Screening High Risk HPV DNA HCII Results: An Enhanced Quality Control Measure

Charles Sturgis, MD, Michael Schaaf, CT(ASCP), Ronald Tickman, MD

Pathology, CellNetix Pathology and Laboratories, Seattle and Everett, Washington

Introduction: CLIA 88 regulations specify that at least 10% of NILM Pap tests be re-screened by a pathologist or qualified senior cytotechnologist as a means of quality control (QC) with a portion of these cases being randomly selected. With the incorporation of high risk HPV DNA testing into ASCCP screening guidelines for women 30 years of age and older, a population of NILM patients with positive high risk HPV DNA results exists. In this study, slides from these patients were re-screened to judge the value of QC focused on this population.

Materials and Methods: A series of 386 consecutive NILM liquid based (mixture of ThinPrep and SurePath) Paps (09/09 to 12/09) from women 30 years of age and older was found to be high risk HPV DNA positive by HCII and was retrieved from the files of CellNetix Pathology and Laboratories, Washington State. The 386 slides were re-interpreted by cytotechnologists. All slides judged ASC or higher by cytotechnologists on re-interpretation were than viewed by one or more cytopathologists (CDS/RJT) who assigned a final re-screen interpretation.

Results: Of the 386 re-screen cases, 50 (13% of total) were placed in diagnostic categories of ASC or higher, and 11 (2.8% of total) were interpreted as LSIL or above (10 LSIL and 1 HSIL being discovered). By comparison, for the calendar year 2009, routine combined QC re-screens (encompassing random, Focal Point enriched, and historical high risk cases) were performed on a total of 20580 Paps (21% of total 99501 CellNetix annual cases). Historical combined QC revealed 2.1% (427 / 20580) cases that were upgraded to ASC or higher and 0.25% (52 / 20580) that were upgraded to LSIL or higher in 2009.

Conclusions: Routine re-screen QC from the 2009 calendar year resulted in a 0.25% detection rate of LSIL or higher cases. Focused rescreening of a four month cohort of NILM patients who were 30 years of age or older and who were positive for high risk HPV DNA by HCII resulted in a 2.8% detection rate of cases judged LSIL or higher and a 13% detection rate of those judges ASC or higher. Focused re-screening of NILM cases with positive high risk HPV DNA resulted in the detection of greater than ten times more SIL cases than did historical combined routine QC Pap slide review at CellNetix. As ASCCP management guidelines for NILM with positive high risk HPV DNA testing differ from those for ASC with positive high risk HPV DNA testing in women over 30, the increased ASC detection rate in this cohort could be important in clinical management. Focused re-screening of this patient set may enhance QC in cytopathology laboratories performing liquid based Paps. An inherent potential bias in study design is recognized, as results of DNA testing were by definition known at the time of re-screen.

Poster Presentations

Breast

1

Comparison of Breast Prognostic/Predictive Markers on Cell Blocks Obtained by Various Methods: Cellient™, Thrombin, and Formalin

Blythe Gorman, MD, Ognjen Kosarac, MD, Mary Schwartz, MD, Dina Mody, MD

Pathology, The Methodist Hospital, Houston, Texas

Introduction: Fine needle aspiration (FNA) is a rapid and minimally invasive method to obtain diagnostic material from lesions at many sites, including the breast. The relative ease of obtaining a sample makes FNA a desirable alternative to core needle or open biopsy. Currently, immunohistochemical staining for the prognostic/predictive markers estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor 2 (Her2-neu), and MIB1 (Ki-67) is performed on tissue from core biopsy or surgical specimen. The purpose of this study was to evaluate if immunohistochemical staining performed on cell blocks made via three different technologies; Cellient™, thrombin and formalin; can provide results comparable to those obtained using tissue sections. This type of study can serve as a model for internal validation of these methods for clinical use. Currently, there is limited data available on the Cellient™ method which has only recently become available on the market, and has the potential advantage of more rapid processing and more cellular preparations.

Materials and Methods: Cytologic preparations were obtained from 21 patients whose surgical resection specimen (mastectomy or lumpectomy) contained a grossly visible lesion. All patients had a biopsy-confirmed invasive mammary carcinoma. Scrapings of the tumor were placed in CytoLyt®, RPMI, and formalin and cell blocks were prepared by the three methods. All specimens collected in CytoLyt® were processed using the Cellient™ method and adequate Cellient™ cell blocks were obtained for all 21 cases. Fourteen of the 21 cases had adequate cell blocks using all three methods. Adequate formalin-fixed and thrombin cell blocks were obtained in 17 and 16 of 21 cases, respectively. Results of immunohistochemical staining for ER, PR, Her2-neu and MIB1 on the cell blocks were compared to results of stains performed on either the patient's previous core needle biopsy or the concurrent surgical resection specimen. ER and PR were scored using the Allred system and Her2-neu was scored using the ASCO/CAP criteria.

Results: Twelve (57%) of the patients had poorly differentiated carcinoma (ESBR score 8-9) carcinomas. Six patients (29%) had moderately differentiated carcinomas (ESBR score 6-7) and three (14%) had well differentiated carcinomas (ESBR score 3-5). In a comparison of the four biomarkers performed on tissue sections to those performed on the three types of cell blocks, the Cellient™ cell blocks showed the highest agreement for ER and Her2-neu (see Table 1). The formalin and thrombin cell blocks also showed a high rate of concordance.

Conclusions: Cell blocks prepared with the newly available Cellient™ method perform exceptionally well with regard to immunohistochemical testing for the two most clinically significant breast cancer biomarkers, ER and Her2-neu.

Table 1. Correlation between results of biomarkers on tissue sections and cell blocks (%)

	Cellient™	Formalin	Thrombin
ER	100%	87.5%	84.9%
PR	55%	82%	80%
Her2	100%	87.5%	85.7%
MIB1	64%	80%	63.6%

2

Immunohistochemical Detection of Estrogen Receptor, Progesterone Receptor and Human Epidermal Growth Factor Receptor 2 in Formalin-Fixed Breast Carcinoma Cell Block Preparations: Correlation of Results to Corresponding Tissue Block (Needle Core and Excision) Samples

Mary Darrow, MD, Momin Siddiqui, MD, Cynthia Cohen, MD, Krisztina Hanley, MD

Pathology, Emory University Hospital, Atlanta, Georgia

Introduction: Evaluation of estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor 2 (Her2) is routinely performed on breast carcinoma tissue samples. For accurate detection of these prognostic markers, compliance with the American Society of Clinical Oncology and College of American Pathologists (ASCO/CAP) guidelines is recommended, which were established on formalin fixed, paraffin embedded tissue sections. Fine needle aspiration (FNA) is widely used in the diagnosis of primary and metastatic breast carcinomas. Our previous study showed that alcohol fixation did not affect ER results when alcohol fixed cell block (CB) sections were compared to formalin-fixed needle core and resection samples. PR and Her2 IHC results on the other hand showed less concordance with formalin fixed tissue samples. The aim of this study was to evaluate if ER, PR and Her2 IHC results on formalin-fixed CB section correlate better to those observed on surgical (needle core or resection) specimens (SS).

Materials and Methods: Fifty cases of formalin fixed CB samples obtained from primary (18%) and metastatic (82%) breast carcinomas were studied, all of which had corresponding SS (31 needle cores and 19 surgical resections) available. ER, PR and Her2 IHC studies were done on all samples and the results were compared. Similar to our previous study, patients who had received neoadjuvant or adjuvant therapy between obtaining the samples were excluded.

Results: ER results on formalin-fixed CB samples showed excellent correlation with SS (92% positive agreement, ê= 0.82). PR results showed 78% agreement (ê= 0.54), while Her2 results by IHC had 74% agreement (ê= 0.57).

Conclusions: The use of formalin as a fixative for CB preparations does not significantly improve the detection of ER positive breast tumors. However, the concordance rate in PR and Her2 IHC results between formalin-fixed CB and SS samples showed improvement as compared to the alcohol fixed CB results. The number of equivocal Her2 IHC results on CB is comparable to those on SS samples. This improved correlation between the two preparation methods is also observed when FISH results are compared in Her2 IHC equivocal cases.

3

The Role of Axillary Lymph Node FNA in Breast Cancer Staging Prior to Chemotherapy with Histologic Correlation

Amal Kanbour-Shakir, MD, PhD, Amy Colaizzi, BS, ASCP, Kanbour Anisa, MD, Ahmad Ibrahim, MD, Sara Monaco, MD

Pathology, Magee-Womens Hospital of UPMC, Pittsburgh, Pennsylvania

Introduction: Axillary lymph node sampling in breast cancer staging is critical prognostic and predictive parameter to select patient for treatment and molecular targeted therapy. Thus, there has been recent interest in preoperative evaluation of the axillary lymph node using minimally-invasive techniques, such as fine needle aspiration (FNA). The aim of this retrospective study is to review the axillary lymph node FNAs performed at our institution prior to starting chemotherapy and/or surgery and to correlate the results with the available lymph node biopsy in order to evaluate the diagnostic utility of FNA in this setting.

Materials and Methods: We searched the pathology database at our institution for all axillary lymph node FNAs performed over a one year period from January 2009 to December 2009. We retrospectively reviewed these cases and correlated the findings with the available lymph node biopsy (core, sentinel or axillary content dissection) to look at the diagnostic utility of FNA in this setting. All restaging lymph node biopsies performed after chemotherapy were excluded.

Results: We identified 166 axillary lymph node FNAs with their biopsies, including 68 (41%) with concurrent core needle biopsies, 83 (50%) with sentinel lymph node biopsies, and 15 (9%) with axillary content dissection. Of these cases, there were 79 cases (48%) positive for malignant cells, 83 cases (50%) negative for malignant cells, and 4 unsatisfactory (2%). A total of 146 cases (88%) correlated with the lymph node biopsy diagnosis. Of the 20 cases (12%) that did not correlate. Of those, 1 was false positive, 3 were unsatisfactory and 16 were false negative FNAs, which were found to have metastases in the lymph node biopsy ranging from isolated tumor cells to 5 mm in greatest dimension. The overall sensitivity was 83%, specificity was 99%, positive predictive value was 99%, and negative predictive value was 81% (See Table 1).

Conclusions: Our findings show that pretherapy staging of breast cancer by minimally-invasive FNA is a useful diagnostic modality with a high specificity and positive predictive value. Due to the risk for false negative diagnoses in FNAs of lymph nodes with small metastases, a negative FNA should be followed up with further lymph node sampling. However, the use of FNA in axillary lymph node assessment can provide patients with a minimally invasive way to obtain staging information without the use of more invasive diagnostic modalities and allow for rapid treatment and management decisions.

Table 1. Correlation of axillary lymph node FNA cytology with biopsy findings in breast cancer patients

	Histology Positive	Histology Negative	Total
Cytology Positive	78	1	79
Cytology Negative	16	67	83
Total	94	68	162

4

The Emerging Role of Fine Needle Aspiration of Satellite Lesions Around Index Lesions Newly Diagnosed as Breast Cancer: A Study of 203 Patients

Manali Bendre, M.D¹, Wei Yang, M.D², Savitri Krishnamurthy, MD¹

¹Cytopathology, MD Anderson Cancer Center, Houston, Texas; ²Breast imaging, MD Anderson Cancer Center, Houston, Texas

Introduction: Core needle biopsy (CNB) is currently the procedure of choice for the investigation of breast lesions that are categorized as highly suspicious for malignancy by imaging. However, the role of fine needle aspiration biopsy (FNA) in the investigation of satellite lesions around the index lesions that are detected by imaging is not well recognized. The primary objective of our study was to evaluate the role of FNA of satellite lesions of the breast in patients with newly diagnosed breast cancer of the index lesion and their impact on subsequent management of the patient.

Materials and Methods: We searched for cases of ultrasound (US) guided FNA of nonpalpable breast lesions that were performed in MDACC (Jan 2007-Dec 2007) in patients with a breast cancer diagnosis of an index lesion. FNA of the satellite lesions was performed under US guidance by the radiologist using a 23-25 gauge needle; direct smears were made, fixed in Carnoy's Solution for Papanicolaou staining and air dried for Diff-quik staining. Immediate assessment was performed on all the cases by cytopathologists. The FNA specimens of the satellite lesions were classified into either: Non-diagnostic, benign atypical, suspicious and malignant. Subsequent follow up of the patient was recorded with respect to type of surgery including segmental mastectomy with or without inclusion of the satellite lesion in the field of surgery, total mastectomy and preoperative chemo/hormonal therapy. The impact of the FNA diagnosis of the satellite lesion on subsequent management of the patients with breast cancer was ascertained.

Results: We evaluated 203 FNAs performed on satellite lesions in patients with an index mass that was diagnosed as breast cancer. We categorized the satellite lesions obtained from 4 (2%) patients as nondiagnostic, 98 (48%) benign, 8 (3.9%) atypical, 8 (3.9%) suspicious and 85 (41.8%) as malignant. A benign categorization allowed the selection of segmental mastectomy in 58/98 (59%) patients. Five of 8 (62.5%) and 3/8 (37.5%) patients with an atypical diagnosis underwent total mastectomy and segmental mastectomy with inclusion of the satellite lesion in the field of surgery. The 8 patients with a suspicious diagnosis underwent total mastectomy (5,62.5%) and segmental mastectomy (3,32.5%). The 85 patients with a FNA diagnosis of malignant underwent total mastectomy in 30 (35%), segmental mastectomy in 8 (9.4%) and neoadjuvant chemo/hormonal therapy in 47 (55%). A malignant diagnosis allowed the radiologist to mark the satellite lesion with a marker clip for subsequent monitoring of the area for response to chemotherapy.

Conclusions:1. US guided FNA of satellite lesions around index lesions is a useful procedure that can guide subsequent management. 2. The FNA diagnosis of satellite lesions influences the decision making in the selection of the patients for segmental mastectomy or total mastectomy. 3. In patients selected for segmental mastectomy, inclusion or exclusion of the satellite lesion in the field of surgery is based on the categorization of the satellite lesions on FNA.4.In patients selected for neoadjuvant chemo/hormonal therapy, a malignant diagnosis on FNA allows the radiologist to put marker clip in the satellite lesion for future monitoring of the response to preoperative therapy.

5

Automated Cellular Imaging System (ACIS III) for HER2 Analysis-Reproducibility and Concordance with Fluorescence in Situ Hybridization (FISH) of a Laboratory Validated Scoring Technique

Douglas Minot, CT (ASCP), Jesse Voss, CT, MB (ASCP), Susan Rademacher, SCT, (ASCP), Toe Lwin, CT, (ASCP), Jessica Orsulak, CT, (ASCP), Aziza Nassar, MD, Beiyun Chen, MD, PhD, Amy Clayton, MD

Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota

Introduction: Quantitative analysis of HER2 immunohistochemical (IHC) stain is used in part to determine which patients with breast cancer may benefit from Herceptin® (trastuzumab) therapy. The goals of this study are to examine both the reproducibility of cytotechnologists (CT) and pathologists (PA) by manual and automated scoring means (using our lab validated, CT scored ACIS III methodology [Dako, Carpinteria, CA]) and concordance of each method with FISH (as the gold standard).

Materials and Methods: This retrospective study was performed using 50 cases that were originally used as part of a HER2 IHC validation study of the ACIS instrument. All specimens had concurrent fluorescence in situ hybridization (FISH) testing performed per validation study protocol. Three pathologists and three cytotechnologists, trained to evaluate HER2 stains, performed a manual assessment based upon current CAP/ASCO guidelines. In addition, the three cytotechnologists performed an ACIS-assisted analysis. The validated ACIS-assisted HER2 analysis consists of a mean score from 2 areas of tumor with intense stain, 2 areas of moderate intensity stain and 2 areas of low intensity stain. Concordance with FISH was determined and interobserver variability (kappa) statistics were calculated based on negative (0 and 1+), equivocal (2+) and positive (3+) HER2 assessments.

Results: Perfect concordances with FISH (100%) were noted for all analyses for IHC negative (0, 1+) and IHC positive (3+) categories (Table 1). The ACIS method had fewer 2+ results overall (10.0%) compared to either CT manual (14.7%) or PA manual methods (19.3%) and had a higher concordance with FISH (40.0% vs 22.7% for CT and 6.9% for PA), while identifying all FISH+ cases. There was also more variability in the rate of 2+ assessments for CT manual interpretation (range, 5-11) and PA manual interpretations (range, 6-14), but less variability by the CT-scored ACIS method (range, 4-6). ACIS-assisted CT interpretations had higher average kappa's than manual CT and manual PA interpretations, (.763 vs. .710 and .673), respectively (Table 2). Using the ACIS, the kappa statistic also improved for 2/3 CT comparisons.

Conclusions: Implementation of ACIS-assisted automated analysis of HER2 IHC can potentially lower the number of HER2 (2+) cases reflexed to FISH, while maintaining acceptable negative (IHC 0, 1+/FISH-) and positive (IHC 3+/FISH+) concordances with the gold standard (FISH). Interobserver statistics demonstrate that observers can, by utilizing the ACIS scoring technique, improve their reproducibility over the manual method. We find that our practice of having CTs pre-score the HER2 IHC cases with our automated ACIS method (with pathologist final review) can provide reproducible results with excellent concordance with FISH. Future prospective studies are needed to determine if using the ACIS III with a validated technique can lower overall costs to patients undergoing HER2 testing while maintaining accurate and reproducible results.

Table 1. Percentage of Patients with a Positive FISH Result by Observer (Pathologist or Cytotechnologist) and Assessment Type (Manual or ACIS-assisted)

Observer	Score (0)	Score (1+)	Score (2+)	Score (3+)	Total
Percentages of cases that are FISH positive are shown in parentheses.
Pathologist 1 (manual)	13 (0)	24 (0)	6 (0)	7 (100)	50
Pathologist 2 (manual)	7 (0)	28 (0)	9 (11)	6 (100)	50
Pathologist 3 (manual)	5 (0)	25 (0)	14 (7)	6 (100)	50
Cytotechnologist 1 (manual)	5 (0)	29 (0)	11 (18)	5 (100)	50
Cytotechnologist 1 (ACIS)	10 (0)	29 (0)	6 (33)	5 (100)	50
Cytotechnologist 2 (manual)	11 (0)	28 (0)	5 (20)	6 (100)	50
Cytotechnologist 2 (ACIS)	14 (0)	27 (0)	4 (50)	5 (100)	50
Cytotechnologist 3 (manual)	5 (0)	34 (0)	6 (33)	5 (100)	50
Cytotechnologist 3 (ACIS)	18 (0)	22 (0)	5 (40)	5 (100)	50

Table 2. Kappa Statistic by Observer (Pathologist or Cytotechnologist) and Assessment Type (Manual or ACIS-assisted)

Comparison	Kappa	95% CI	Rating
Negative categories (0 and 1+) are combined.
Pathologist 1 and Pathologist 2	0.684	0.468-0.901	Good
Pathologist 1 and Pathologist 3	0.684	0.483-0.885	Good
Pathologist 2 and Pathologist 3	0.651	0.444-0.857	Good
Tech 1 (manual) and Tech 2 (manual)	0.633	0.399-0.866	Good
Tech 1 (ACIS) and Tech 2 (ACIS)	0.765	0.545-0.986	Good
Tech 1 (manual) and Tech 3 (manual)	0.769	0.577-0.961	Good
Tech 1 (ACIS) and Tech 3 (ACIS)	0.831	0.645-1.016	Very Good
Tech 2 (manual) and Tech 3 (manual)	0.728	0.502-0.954	Good
Tech 2 (ACIS) and Tech 3 (ACIS)	0.693	0.438-0.948	Good

6

Cytology of Spontaneous Nipple Discharge: Is it worth it? Performance of Nipple Discharge Preparations in the College of American Pathologists Inter-Laboratory Comparison Program in Non-Gynecologic Cytology

Mary Schwartz, MD¹, Rodolfo Laucirica, MD², Christine Booth, MD³, Manon Auger, MD⁴, Rhona Thomas, MPH, CT(ASCP)⁵, Nicole Thomas, MPH, CT(ASCP)⁶, Ann Moriarty, MD⁷

¹Pathology, The Methodist Hospital, Houston, Texas; ²Pathology, Baylor College of Medicine, Houston, Texas; ³Pathology, Cleveland Clinic, Cleveland, Ohio; ⁴Pathology, McGill University, Montreal, Quebec, Canada; ⁵Biostatistics, College of American Pathologists, Northfield, Illinois; ⁶Surveys, College of American Pathologists, Northfield, Illinois; ⁷Pathology, AmeriPath Indiana, Indianapolis, Indiana

Introduction: The clinical utility of cytologic examination of spontaneous nipple discharge is controversial. Additionally, most cytologists do not have the opportunity to evaluate a significant number of nipple discharge preparations (NDP), and rarely see malignant cases. The College of American Pathologists (CAP) Interlaboratory Comparison Program in Non-Gynecological Cytology (NGC) has included nipple discharge specimens since its inception. Do nipple discharge preparations perform predictably in an educational interlaboratory comparison program? Can we use the experience from the program to predict the performance of nipple discharge preparations in actual practice? The objectives of this study were to evaluate participant responses for the general diagnostic categories of Benign, Suspicious and Malignant, comparing differences in reference diagnosis, participant type, stain, and survey year.

Materials and Methods: We evaluated 2506 nipple discharge preparation participant responses between 2005 and 2009. General diagnostic category (Benign, Suspicious, Malignant), participant type (Pathologist, Cytotechnologist), stain (Papanicolaou, Modified Giemsa) and year in the program (2005-2009) were analyzed using chi-square and a nonlinear mixed model for slide factor correlation structure. By design, all papillary lesions were categorized as suspicious, and comprised the only entity in the suspicious general category.

Results: Of the 2506 responses, the reference diagnosis was malignant in 1280 (51%), suspicious/papillary in 171 (7%) and benign in 1055 (42%). 222 responses were discordant to the general diagnostic category with a false positive/suspicious rate of 12.8% and a false negative rate of 3.4%. The most common false negative diagnosis of adenocarcinoma was mastitis/abscess (125 of 1272 responses; 9.8%). The most common false positive response for a benign case was papillary lesion (41 of 1049 responses, 3.9%). There were no differences between stain or program year. Cytotechnologists performed better than pathologists, with a higher false negative rate for pathologists (15.3 vs. 7.9%, P<.001.) Discordance was least for negative reference diagnoses and comparable for papillary lesions and adenocarcinoma.

Conclusions: Nipple discharge preparations show significant participant discordance to the general diagnostic category. Adenocarcinoma is often not recognized and interpreted as mastitis/abscess or papillary lesion. Benign lesions are sometimes overcalled as papillary lesions. If the findings in the CAP Interlaboratory Comparison Program parallel clinical practice, they illustrate potential diagnostic pitfalls which may impact patient care. A benign cytologic diagnosis does not exclude malignancy. Conversely, given the significantly high false positive/suspicious rate, these cytologic diagnoses should be corroborated before being used for patient management. 1

Table 1. Nonlinear Mixed Model Results

	Number of responses	Concordance to General Diagnostic Category	P-value
Reference Diagnosis			.01
Adenocarcinoma	1280	87.5%
Papillary lesion	171	84.8%
Benign	1055	96.6%
Participant Type			.03
Cytotechnologist	718	93.9%
Pathologist	1293	89.9%
Stain			.53
Papanicolaou	1842	89.6%
Modified Giemsa	664	95.3%

Education/Training/Current Trends

7

The Current and Future State of Cytology in New York State (NYS): A Needs Assessment Study

Indra Balachandran, PhD, SCT(ASCP),CFIAC¹, Maria Friedlander, MPA, CT(ASCP) CMIAC²

¹Cytotechnology Program, Albany College of Pharmacy and Health Sciences, Albany, New York; ²School of Cytotechnology, Memorial Sloan-Kettering Cancer Center, New York, New York

Introduction: In New York State, 60% (3/5) of cytotechnology training programs have closed since 2008 and only 2 schools remain. Contributing factors include financial reasons as well as perceptions of decreased demand for cytology screening services and impact of recent NYS licensure laws. A survey was conducted in an effort to get more accurate data concerning current and future direction of cytology practice in NYS.

Materials and Methods: The study utilized an on-line survey tool of 22-questions for a period of 7 weeks. One-hundred thirty (130) NYS laboratories licensed to perform GYN, Non-GYN and FNA cytology testing received e-mail communication to participate.

Results: Sixty-seven laboratories responded representing a 52% response rate. A summary of respondents and number of cytotechnologists (CT) employed by facility and job type are outlined in Table I. While the majority of labs reported stable test volumes over the past 2 years, hospital/medical center based labs reported an increase in both non-GYN (48% of labs) and FNA (62% of labs) at an incremental volume of 1-10% with no further increase expected to occur over the next 2 years. Eighty-six percent of responding labs utilize Thin Prep (TP) preparations to evaluate GYN specimens; a quarter of these labs (24%) are using the TP Imaging system (TPIS). Interestingly, screening productivity of labs using TPIS ranged from <50 slides/day to 90-100 slides/day. Nine labs report use of BD Surepath with only 3 using automated BD Focal Point systems. In hospital-based laboratories, additional CT responsibilities ranged from smear preparation assistance (54.8%), provision of specimen adequacy during FNAs (51.6%), use of laboratory information systems (71%), cytopreparation (56.5%), histology activities (21.4%), immunochemistry (14.3%) and molecular testing (17.9%). Of the 5 laboratories with CTs involved in molecular testing, CTs are involved with HPV testing (Digene and Cervista) and FISH. While majority of labs (45%) reported that they are not considering expanding scope of practice for CTs in the near future, 16% report that CT responsibilities will expand to include molecular testing. One-third of laboratories reported that they plan to hire 1-2 CTs within the next 3 years due to retirement of existing staff. Of the 37 labs who hired CTs within the past 3 years, 70% of labs hired new graduates (in-state trained-73%/out-of-state trained-27%). Direct contact with NYS training programs (56%) and word of mouth amongst local CT professionals (50%) were most useful recruiting strategies. Vacancies were filled within 3 months in the majority of hiring labs (64.6%).

Conclusions: While molecular pathology continues to evolve and involve cytology practice, survey data suggests that current CT scope of practice of is not likely to significantly change in the immediate future in the majority of NYS labs. Impending retirements of an aging workforce is anticipated and necessitates continued training of CTs to continue cytology practice. The continued viability of the two remaining NYS cytotechnology programs is essential to provide necessary training and staffing of NYS labs.

Table 1. Summary of Responding Labs and Number of CTs Employed by Facility and Job Type

	Staff CT	Lead/Sr/ Chief CT	Mgr/Supvr CT	Non-Traditional CT roles
Hospital/Large Medical Center (29)	149	23	53	10
Community Hospital (22)	79	25	26	11
Commercial Reference Lab (2)	2	2	1	0
Private Lab (11)	114	15	24	8
College/University Lab (2)	3	0	2	0
Public Health Lab (1)	16	0	2	0
TOTAL	363	65	108	29

8

Assessment of Technology Implementation among Cytotechnology Programs

Joe Walker, Jr., BS¹, Indra Balachandran, PhD²

¹Pathology, Rutland Regional Medical Center, Rutland, Vermont; ²Health Sciences, Albany College of Pharmacy and Health Sciences, Albany, New York

Introduction: The use of technology in the classroom has changed over the past century to reflect prevailing educational theories. Through the use of technology, students are provided freedom to explore information, manipulate the information and data, which enhances the development of higher order thinking skills that potentially affect the transfer of knowledge. Traditionally, cytotechnology training programs are slow to adopt technology due to several factors, which include: small accreditation and enrollment numbers, one on one instruction and slide screening components. There is no data that examines cytotechnology training programs and their adoption of technology. This study attempted to gauge the extent of implementation of different types of teaching technologies that have been integrated into cytotechnology curriculums in the US, explored the training of cytotechnology educators in the use of teaching technologies and pedagogy and attempted to correlate technology integration with successful completion of the Board of Registry(BOR) certification examination.

Materials and Methods: The authors created a survey instrument, which was distributed electronically to all 35 active program directors of Cytotechnology training programs in the United States. The survey instrument collected demographic information on the program director and education coordinator of each program in addition to specific questions related to technology implementation in their program. Survey results were analyzed for trends.

Results: Of the 35 active programs, 25 program directors responded (71.4%) but only 22 (62.6%) completed the entire survey. 81.8% of respondents do not require their students to have a laptop to attend their program but 54.5% do require students to have computer access. 40.9% of respondents have required their students to have computer access for 5 or more years. The three most common uses of computers within the curriculum are to access information on the web, review tutorials and to access online periodicals. 59.1% report students using computers to take exams. The majority of respondents (95.5%) teach their courses in a face to face setting with only 27.3% utilizing asynchronous distance learning courses. 63.6% of respondents utilize a course management system with 8 (50%) of those program utilizing it for 5 or more years. The most common use of the course management system is related to posting content online. The three most common technologies used in cytotechnology curricula are electronic tutorials, educational games and electronic books. 54.5% of respondents state performance on the BOR exam remained the same while 22.7% report an increase in performance. The primary reason for not adopting technology into the curriculum was due to lack of funds.

Conclusions: Of the respondents who utilize technologies within their curriculum very few report using advanced technologies like streaming content or online courses to enhance teaching. Only 22.7% of respondents report increased performance on the BOR exam and this corresponds to longer length of technology usage, utilization of examination software and advance technologies. While the vast majority of responses (81%) stated that technology has a role in cytotechnology education only half of the respondents felt that technology would improve on the delivery of content and assessment of their students.

Fluids/CSF

9

Cerebrospinal Fluid Cytology: A Four Year Study

Margarida Almeida, MD, Ines Pestana, CT(ASCP)

Cytology, Hospital Santa Maria, LIsboa, Portugal

Introduction: Cerebrospinal Fluid (CSF) is submitted for cytologic examination to evaluate for infections or inflammatory conditions, or a neoplastic process. The aim of this study is to review our CSF cytology results over a four year period.

Materials and Methods: We reviewed 981 CSF samples from 864 patients, ranging from 1 month to 90 years. CSF samples were processed by cytospin. Two glass slides were obtained for each case and stained with PAP stain. The cases were classified as: unsatisfactory (only blood or not well preserved material), benign, suspicious and malignant. Either histologic diagnosis or clinical definitive diagnosis was used as the final or gold standard diagnosis.

Results: Most of the cases were diagnosed as benign (n=753, 77,2%). These included: inflammation (n=170) and negative (n=583). Specific infectious organisms were identified in 5 cases (criptococos; n=4 and cocos; n=1). Eleven cases (1,2%) were reported as suspicious. On further investigations, four of these were from a primary tumour (two germinomas, one glioma and one medulloblastoma) while the other 7 cases were metastasis (three B-cell lymphoma, two breast carcinoma and one prostate adenocarcinoma). 96 samples from 58 patients were malignant. The primary tumors (n=15) included: three medulloblastomas, seven gliomas, three germinomas, one choroid plexus papilloma and one rhabdoid tumor. The metastatic tumours (n=81) were 36 breast carcinomas, 14 carcinoma of lung, 4 gastric adenocarcinoma, one melanoma, two rhabdomyosarcoma, 2 leukemia and 22 B-cell non Hodgkin lymphoma. In 122 cases the material was insufficient for diagnosis.

Conclusions: This study shows: 1- a high number of insufficient cases (12%) , 2- no false positive diagnosis, 3- the majority of malignant cases were metastatic neoplasm's, and the breast is the more common, 4- all the suspicious cases, proved to be malignant. In summary, CSF cytology is an accurate and reproducible approach in a well established setting, for the patients with neurological disorders. Immediate assessment optimises specimen adequacy.

10

Malignant Pleural Effusion in Pediatric and Adolescent Patients, Our Experience in a Cancer Center

Shaoxiong Chen, MD, Qiusheng Si, MD, Ming Guo, MD

Pathology, The University of Texas MD Anderson Cancer Center, Houston, Texas

Introduction: Malignant pleural effusion in pediatric/adolescent patients is usually caused by non-Hodgkin's lymphoma, and only occasionally caused by other small blue cell tumors or sarcomas. Reported studies of malignant effusion in pediatric/adolescent patients are scant, especially on the distribution of the malignancies other than non-Hodgkin's lymphoma. In this study, we reviewed pleural fluid in pediatric/adolescent patients with history of malignant neoplasms, analyzed the frequencies of the pediatric/adolescent malignancies in pleural fluid and showed how the cytology diagnoses were made.

Materials and Methods: We retrieved and reviewed 80 cases of pleural effusion from patients with history of malignant neoplasms at MD Anderson Cancer Center from 2002 to 2009. The age of the patients ranged from 1 to 20 years old with a mean age of 14.5 years. The patient's past medical history included non-Hodgkin's lymphomas (19 cases), osteosarcomas (17 cases), Hodgkin's lymphomas (14 cases), rhabdomyosarcomas (12 cases), Ewing's sarcoma/PNET (7 cases), Wilm's tumors (4 cases), neuroblastomas (3 cases), medulloblastomas (2 cases), and unclassified round cell sarcomas (2 cases). Cytology diagnoses with cytomorphology, immunocytochemical study and flow cytometry analysis were reviewed.

Results: Of the 80 cases of pleural effusion, 25 (31%) specimens contained malignant cells. Of the 25 malignant pleural effusion cases, there were 8 cases of non-Hodgkin's lymphomas, 6 cases of osteosarcomas, 4 cases of Hodgkin's lymphomas, 3 cases of Ewing's sarcoma/PNET, 2 cases of rhabdomyosarcomas and 2 cases of Wilm's tumors (Table 1). Ancillary studies, such as immunostains and flow cytometry analysis, were required for the diagnoses of non-Hodgkin's lymphoma. Diagnoses of malignant pleural effusion for the other malignancies were mainly based on patient's history and cytomorphology.

Conclusions: Non-Hodgkin's lymphoma is the most frequent malignant pleural effusion in our group of pediatric/adolescent patients. Osteosarcomas, Hodgkin's lymphoma and Ewing's sarcoma/PNET are also commonly encountered in malignant pleural effusion in our study. Except for non-Hodgkin's lymphoma, cytomorphology alone is sufficient for correct diagnoses of malignant pleural effusion in pediatric/adolescent patients with a known history.

Table 1. Malignant pleural effusion in pediatric and adolescent patients with history of malignant neoplasms

Diagnosis	No. of positive pleural effusion cases	%
Non-Hodgkin's lymphoma	8	32
Osteosarcoma	6	24
Hodgkin's lymphoma	4	16
Ewing's sarcoma/PNET	3	12
Rhabdomyosarcoma	2	8
Wilm's tumor	2	8
Neuroblastoma	0	0
Medulloblastoma	0	0
Unclassified round cell sarcoma	0	0

11

The Cytomorphological Spectrum of Small Cell Carcinomas and Large Cell Neuroendocrine Carcinomas in Serous Effusion. Our Experience with 68 Cases

Huaitao Yang, MD², Sara Monaco, MD¹, Qian Lian, MD, PhD², Walid Khalbuss, MD PhD¹

¹Pathology, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania; ²Pathology, Conemaugh Memorial Medical Center, Johnstown, Pennsylvania

Introduction: Small cell carcinoma and neuroendocrine carcinoma are uncommon in serous body effusion cavity. The purpose of this study is to examine the cytomorphological spectrum of small cell carcinoma and large cell neuroendocrine carcinoma in body cavity serous fluid, in addition to the role of ancillary studies was also examined.

Materials and Methods: We have examined 5171 serous body effusion specimens during five years in which 723 (14%) were positive for malignancy. There were 68 cases (1.4%) cases from 53 patients that were diagnosed as metastatic small cell carcinomas or metastatic neuroendocrine carcinomas. All cytology slides as well as the available clinical data, histological followup confirmation, and ancillary studies were reviewed.

Results: A total of 68 cases (60 pleural, 5 peritoneal, and 3 pericardial effusions) from 50 patients with an average age of 73 years were reported as diagnostic or suggestive of small carcinoma (52 cases) or large cell neuroendocrine carcinoma (16 cases). The primary site was lung in 56 cases, pancreatic in 6 cases, and cervical, colonic, and head and neck region (2 cases each). Of the 68 cases, 48 (71%) had no previous history of malignancy of the same type. Of the 20 cases with previous history of malignancy, the average period interval from primary to positive effusion was 36 months. Ancillary studies were utilized in 46 cases (68%) including flow cytometric studies in 5 cases. Three cytomorphological predominance patterns seen: predominance of small clusters (33 cases); predominance of large clusters mimicking non-small cell carcinoma (18 cases), and predominance of single cell pattern mimicking lymphoma (17 cases). Significant apoptosis was seen in 22 cases, and marked cannibalism was seen in 11 cases. Prominent nucleoli were noted in 16 cases.

Conclusions: The most common cytomorphological pattern was predominance of small clustering with nuclear molding. However, in 51% of the cases the predominant cytomorphological pattern was either single cell pattern mimicking lymphoma or large clustering pattern mimicking non-small carcinoma. Knowing the cytomorphological spectrum of small and neuroendocrine carcinoma in fluid cytology may help in establishing timely and accurate diagnoses.

12

Vitreous Cytology Diagnosis: An Analysis of 4683 Cases

Ana Solky, MD¹, Jay Duker, MD², Caroline Baumal, MD², Elias Reichel, MD², Nora Laver, MD¹

¹Pathology, Tufts Medical Center, Boston, Massachusetts; ²Ophthalmology, New England Eye Center, Boston, Massachusetts

Introduction: The vitreous humor is an important component of the visual system of the eye. It is a transparent, semisolid substance in close contact with the retina, ciliary body and lens. The cytologic analysis of vitreous fluid can be useful in the diagnosis of intraocular diseases. We employed a fluid-based, thin-layer system in the analysis of 4231 vitreous fluid samples over the course of eleven years. The most important cytological findings of vitrectomy specimens are discussed.

Materials and Methods: A total of 4683 vitrectomy specimens over an 11 year period (August 1999 to March 2010) were analyzed. A fluid based thin-layer system was employed until 2007 (ThinPrep) with two thin preps prepared, H&E and PAS stained. From 2007 until present, the Surepath method was employed for analysis and two slides were prepared, PAP and PAS stained. The most important cytological findings of vitrectomy specimens are discussed. The cases are divided into three major diagnostic groups: negative for malignancy, atypical/suspicious and positive for malignancy.

Results:Table 1 – The majority of the cases were negative for malignancy (99.4%). We divided this group into 3 subcategories: common findings, inflammatory/infectious finding, and rare findings. Most cases showed benign cytological findings. Fourteen cases (0.30%) were atypical/suspicious for malignancy and eleven (0.24%) were positive for malignancy.

Conclusions: Our data show that fluid based thin layer cytopathologic analysis is a reliable and convenient method for the diagnosis of vitreous sample pathology. Moreover, utilizing this method coupled with a cellblock preparation and the utilization of ancillary procedures can increase the diagnostic yield. The results show that both malignant and non-malignant pathology of the eye can be evaluated with victrectomy cytology.

Table 1. 

	Number of cases	‰
MOST COMMON FINDINGS	4217	90.06
vitreous fluid	1883	40.2
vitreous hemorrhage	1057	22.59
lens fragments/lens capsule	371	7.92
Retinal fragments	344	7.36
fibrovascular membranes	252	5.38
Degenerated retinal Frag.	242	5.18
Intern. limiting memb frag	48	1.02
subretinal membrane	18	0.38
Subretinal hemorrhage	1	0.02
Retinal atrophy	1	0.02
INFLAMMATORY/INFECTIOUS	417	8.92
Chronic inflammatory cells	344	7.37
Acute inflammatory cells	51	1.09
Endophthalmitis	10	0.21
Chronic uveitis	3	0.06
Candida	3	0.06
Chronic vitritis	2	0.04
Viral cytopathic effect	1	0.02
Eosinophils	1	0.02
CMV	1	0.02
Hydatic cyst wall	1	0.02
OTHER DIAGNOSIS	25	0.48
Asteroid hyalosis	12	0.26
vitreous amyloid	4	0.08
Uveal frag.	4	0.08
foreign bodies	1	0.02
Silicone oil	1	0.02
PHPV	1	0.02
ATYPICAL/SUSPICIOUS FOR MALIGNANCY	14	0.30
lymphocytes	13	0.28
Round blue cells	1	0.02
POSITIVE FOR MALIGNANCY	11	0.24
B-cell lymphoma	6	0.14
Melanoma	3	0.06
retinoblastoma	1	0.02
Metastatic carcinoma	1	0.02
TOTAL NUMBER OF CASES	4683	100

13

The Diagnostic Utility of D2-40, Calretinin, CK5/6, Desmin, and MOC-31 in the Differentiation of Mesothelioma from Adenocarcinoma in Pleural Effusion Cytology

Michael Barnes, MD, Laura Tabatabai, MD

Department of Pathology, University of California-San Francisco, San Francisco, California

Introduction: Differentiation between mesothelioma and adenocarcinoma remains one of the most problematic areas in effusion cytology. Recent studies have suggested that the novel lymphatic endothelial marker D2-40 may be useful in the diagnostic workup of effusions. We tested D2-40 along with Calretinin, CK5/6, Desmin, and MOC-31 to help establish the utility of these antibodies in pleural effusion cytology specimens.

Materials and Methods: Forty-five archival, formalin-fixed, paraffin embedded cell blocks of pleural effusions representing 13 reactive effusions, 11 mesotheliomas, and 21 metastatic adenocarcinomas were retrieved. All cases were clinically and/or histologically confirmed. All cases were immunostained with anti-D2-40, Calretinin, CK5/6, Desmin, and MOC-31 and the immunoreactions were evaluated in a blinded fashion by two pathologists.

Results: D2-40 showed membranous staining in 82% of mesotheliomas and 77% of benign reactive effusions. Calretinin and CK5/6 were positive in 100% and 64% of mesotheliomas and 92% and 64% of benign reactive effusions, respectively. Desmin was negative in all malignant cases and positive in 85% of benign reactive effusions. All metastatic adenocarcinomas were positive for MOC-31 and negative for D2-40, Calretinin, Desmin and CK5/6.

Conclusions: Among the markers studied, Calretinin was the most sensitive in detecting mesothelial differentiation, followed by D2-40 and CK5/6. Although D2-40 can serve as a useful adjunct to a panel of markers in the evaluation of pleural effusion samples, due to its sometimes focal and weak positivity, interpretation must be made cautiously, particularly in limited cellularity samples. All of these stains were equally and highly specific in detecting mesothelial differentiation. The muscle marker, Desmin, may be useful in differentiating benign effusions from malignant ones but is not useful in distinguishing mesotheliomas from adenocarcinomas. MOC-31 was both highly sensitive and specific for detecting adenocarcinoma and is very useful as part of a panel of stains in differentiating cells of mesothelial origin from metastatic adenocarcinoma.

Gastrointestinal Tract

14

Endoscopic Retrograde Cholangiopancreatography (ERCP) and Endoscopic Ultrasound-Guided Fine Needle Aspiration (EUS-FNA) in the Cytologic Diagnosis of Cholangiocarcinoma-A Comparative and Morphologic Study

Sima Fargahi, MD¹, John Lee, MD², Fritz Lin, MD¹, Mai Gu, MD, PhD¹

¹Department of Pathology and Laboratory Medicine, University of California Irvine Medical Center, Orange, California; ²Division of Gastroenterology of Department of Internal Medicine, University of California Irvine Medical Center, Orange, California

Introduction: EUS-FNA is sensitive and specific in the cytologic diagnosis and clinical staging of malignant neoplasms of the gastrointestinal tract, pancreas, liver, and lymph nodes. In some institutions, it has mostly replaced ERCP for the diagnosis of cholangiocarcinoma, while in others where EUS is not readily accessible ERCP still remains a main endoscopic procedure. We conducted a morphologic study comparing ERCP and EUS-FNA in the cytologic diagnosis of bile duct lesions.

Materials and Methods: All ERCP and EUS-FNA cases of suspected primary cholangiocarcinoma performed at the University of California Irvine Medical Center from January 1998 to December 2009 were retrieved. The cytologic features of those that had both procedures were compared and analyzed with available histology reviewed.

Results: Three hundred and eleven cases of ERCP and 223 cases of EUS-FNA were found, including 43 positive, 231 negative, and 37 atypical for ERCP and 73 positive, 136 negative, and 14 atypical for EUS-FNA. One hundred and nine cases of ERCP and 66 cases of EUS-FNA had histologic follow up. The sensitivity, specificity, positive predictive value, negative predictive value, and the diagnostic accuracy was 67%, 95%, 86%, 87%, and 86% for ERCP and 66%, 86%, 82%, 71%, and 76% for EUS-FNA, respectively. Seventy-five patients had both ERCP and EUS-FNA with follow up histology available in 28 (Table 1). The cytologic features of ERCP comparing with those of EUS-FNA were similar but with lower cellular yields in most cases of ERCP. Small 3-dimensional cell clusters, single cells, prominent nucleoli, smooth regular nuclear membrane, hypochromatic and evenly distributed chromatin, moderate nuclear pleomorphism, and moderate amount of vacuolated cytoplasm were features for adenocarcinoma for both procedures with necrosis and inflammatory cells rarely seen. Due to higher cellularity associated with EUS-FNA, false positive rate was higher. Due to lower cellularity for ERCP, more inconclusive cytologic diagnoses (eg atypical) were observed, particularly for well-differentiated cholangiocarcinoma.

Conclusions: Although EUS-FNA offers direct visualization of the bile duct lesions with targeted sampling, ERCP is comparable to and more specific than EUS-FNA for the cytologic diagnosis of cholangiocarcinoma and has a higher negative predictive value. The nature of the lesion, the technical skills of the endoscopists, and the experience and the confidence of the cytopathologists in the interpretation of samples of variable cellularity play an important role as well.

Table 1. Comparison of the Cytologic Diagnosis of ERCP with that of EUS-FNA in 75 Cases

ERCP	Negative	Positive	Negative	Positive	Negative	Atypical
EUS-FNA	Negative	Positive	Positive	Negative	Atypical	Negative
Number of Cases	43	8	10	8	2	4
Histology	11 Negative 4 Positive	8 Positive	2 Positive 1 Negative	2 Positive	None	None

15

Is It Time for the Interventional Anal Pap? A Follow Up of Gynecological and Anal Dysplasia Patterns in Female Patients

John Plavnicky, CT (ASCP ), Sharlin Varghese, MD, Donna Russell, CT (ASCP ), Ellen Giampoli, MD

Cytopathology, University of Rochester Medical Center, Rochester, New York

Introduction: Anal pap testing for women is fairly new and incorporating it into routine clinical care is unchartered. This retrospective study is designed to follow gynecologic and anal cytology specimens of HIV infected as well as non-infected female patients in our institution. The information of interest is whether concurrent or preceding pathological abnormalities of the cervix, vagina or vulva will alert the advent of anal dysplasia. The purpose is to study the correlation between the significant risk factors and the patterns of anal dysplasia occurrence.

Materials and Methods: A computerized search for women with anal cytology was performed from 2007 to 2009. The cytologic diagnosis was recorded, as well as patient age, HIV status, and any previous or concurrent dysplasia (anal, cervical, vaginal, or vulvar). All research and data collection was performed in accordance with the University IRB. After sorting the collected data, trends were identified and analyzed.

Results: The search of our database resulted in 150 female patients with anal Pap smear cytology. Of these women, 109 patients were HIV positive, 99 patients had a history of GYN dysplasia or atypia, and 9 patients had an unknown GYN history. (See Table 1)

High grade anal dysplasia occurred in 20 of the 150 patients. Out of those 20 patients, 13 patients (65%) were HIV positive, and 14 (70%) had a history of GYN dysplasia or atypia. 5 (25%) were both HIV positive and had a history of high grade GYN dysplasia while 2 (10%) of those 20 patients had a history of high grade GYN dysplasia and their HIV status was unknown.

The 7 high grade anal dysplasia cases that had associated high grade GYN dysplasia ranged in age from 19-60 years. The anal dysplasia was identified 3-5 years after the diagnosis of the high grade GYN dysplasia in three cases. In the remaining four cases the dysplasias were diagnosed in the same year.

Only 3 patients (15%) with a high grade anal dysplasia were found to be both HIV negative and have a negative or unknown GYN history.

Conclusions: Our results show that in the female population, high grade anal dysplasia is associated with a history of GYN dysplasia or atypia and that HIV positivity is a strong risk factor. Although there are guidelines for HIV positive women to obtain anal Pap testing, there are no universal guidelines on how or when to initiate anal Pap testing in HIV negative women.

Our data notes a small population of women without significant risk factors who had high grade dysplasia on the anal Pap test. This raises the concern that clinicians may need to take initiative to create public awareness that this testing is available to women who engage in anal receptive intercourse, regardless of their HIV/ GYN dysplasia status.

Though this study is limited by sample size, the data does enforce the need for early rigorous anal pathology screening not only in HIV positive women but also in those with high grade dysplasia in the proximal areas and any woman engaging in anal receptive intercourse.

Table 1. 

	HIV status	Anal Pap diagnosis	GYN history
LSIL or ASCUS diagnosis includes ASC-H and LSIL-H.
150 Patients	HIV + 109	HSIL 13	HSIL (5)
			LSIL or ASCUS (5)
			Negative (1)
			Unknown (2)
		LSIL or ASCUS 48	HSIL (13)
			LSIL or ASCUS (28)
			Negative (3)
			Unknown (4)
		Negative 48	HSIL (8)
			LSIL or ASCUS (20)
			Negative (16)
			Unknown (4)
	HIV - 20	HSIL 5	HSIL (0)
			LSIL or ASCUS (2)
			Negative (1)
			Unknown (2)
		LSIL or ASCUS 6	HSIL (2)
			LSIL or ASCUS (0)
			Negative (3)
			Unknown (1)
		Negative 9	HSIL (3)
			LSIL or ASCUS (2)
			Negative (0)
			Unknown (4)
	HIV unknown 21	HSIL 2	HSIL (2)
			LSIL or ASCUS (0)
			Negative (0)
			Unknown (0)
		LSIL or ASCUS 4	HSIL (1)
			LSIL or ASCUS (0)
			Negative (0)
			Unknown (3)
		Negative 15	HSIL (3)
			LSIL or ASCUS (5)
			Negative (4)
			Unknown (3)

16

The Use of P16/Ki67 Dual Immunostaining as an Adjunct to Routine Anal Cytology in Screening for Anal Squamous Intraepithelial Lesions

Brenda Sweeney, MS SCT (ASCP), Jessica Wang, MD, Rosemary Tambouret, MD, David Wilbur, MD

Cytopathology, Massachusetts General Hospital, Boston, Massachusetts

Introduction: The anal Pap cytology specimen is a useful screening tool for anal cancer. p16 staining has been previously shown to be a useful adjunct in the interpretation of anal cytology specimens. In histology anal samples, p16 together with Ki67 staining, and the hematoxylin and eosin appearance is useful in the identification and grading of anal squamous lesions. CINtec®Plus (mTm), is an immunocytochemical dual stain for p16/Ki67, became available in the United States market in February 2010, and has been shown to work well in cytologic specimens. The objective of this study is to assess the value of immunocytochemical staining of anal cytology specimens with p16 and Ki67 simultaneously as a marker of rare occult HSIL cells that may be otherwise missed or misinterpreted in routinely-stained Papanicolaou-stained specimens.

Materials and Methods: Residual material was obtained from 63 specimens submitted to the laboratory for anal Pap cytology. An additional SurePath slide was made and stained with the CINtec®Cytology Dual Stain kit. The stained slides were evaluated for the presence of individual cells that were dual stained (DS) for p16 and Ki67. Cases found to be DS-positive were rescreened for putative HSIL cells.

Results:(See Table 1)

Retrospective review of DS positive cases in categories lower than HSIL showed that 10 of 15 contained rare atypical putative high-grade cells on the original Papanicolaou stained slide (67% of DS positive cases).

Conclusions: The p16/Ki67 dual stain shows promise as a cytology adjunct to improve sensitivity and, and potentially specificity, in the interpretation of anal Pap tests. DS positivity aided in the identification of rare atypical cells not identified on primary manual screening. In addition, for cytologically abnormal samples, positive staining may indicate patients at greater risk of progression/coexistence of high grade disease. Follow up of this DS positive cohort is underway to determine the outcome of patients in all diagnostic categories.

Genitourinary (Kidney & Bladder)

Table 1. 

	Original morphologicPap result	DS Negative	DS Positive	Rescreen positive forputative HSIL
UNSAT	6	6	0	Not applicable
NILM	35	32	3	2
ASC-US	8	4	4	3
LSIL	13	6	7	5
HSIL	1	0	1	Not applicable
Total	63	48	15	10

17

The Clinicopathologic Correlates of Cytologic Atypia in Urinary Cytology of Ileal Neobladders

Ashley Cimino-Mathews, MD, Syed Ali, MD

Pathology, The Johns Hopkins Hospital, Baltimore, Maryland

Introduction: Ileal neobladders (INB) are routinely created in patients who are post-cystectomy for urothelial carcinoma (UC). These patients are periodically screened by urinary cytology for tumor recurrence, and the sensitivity of urine cytology for detecting high-grade UC exceeds 90%. However, the diagnosis of urothelial atypia remains a non-standardized category in INB specimens and has never been studied before.

Materials and Methods: The cytopathology archives from a large tertiary care center were searched for a period of 19 years (1991-2010) for all urine cytology specimens from patients with an INB and atypical urine cytology. Slides were prepared either by cytospins or SurePath methodology. The cytologic criteria used to define urothelial atypia included the presence of nuclear abnormalities such as hyperchromasia and increased nuclear to cytoplasmic ratio which lacked the quantitative or qualitative evidence of definitive carcinoma. All follow-up surgical pathology specimen results and clinical data were reviewed.

Results: A total of 624 urine specimens were identified from patients with INBs, 86 (13.8%) of which were diagnosed as “atypical.” Of these 86 atypical specimens, there were 60 patients with a known history of urothelial carcinoma and available follow-up. Of the 60 patients with atypical urine cytology, only 9 developed a concordant local recurrence (15%, positive predictive value), defined as a biopsy-proven recurrence within one year of atypical urine cytology. The first atypical urine after cystectomy occurred an average of 5.3 months sooner in patients who developed local recurrence (average 33.6 months) than in patients who did not (average 38.9 months, p = 0.65, t-test). The patients with local recurrence had more diagnoses of “atypical cells suspicious for carcinoma” than patients without local recurrence (33% versus 10%, p = 0.19, z-test). Furthermore, the recurrent patients had no diagnoses of “atypical cells, favor reactive” while that diagnosis was made in 10% of non-recurrent patients (p = 0.44, z-test).

Conclusions: While a diagnosis of urothelial atypia is not infrequently made in urine specimens from patients with INB (13.8%), a diagnosis of atypia has a low positive predictive value (15%) for local recurrence in patients with a history of urothelial carcinoma. In contrast, the reported positive predictive value for a diagnosis of urothelial atypia in urine specimens from predominantly native bladders is 77-79% (Am J Clin Pathol 2009; 132:785-793). These findings necessitate a much higher threshold for the diagnostic category of “urothelial atypia” in this patient population with INB.

18

UroVysion™ FISH Testing of the Urinary Tract: Comparison of Results Utilizing Manual Examination and Image Guided System

Nikhil Sangle, MD, Denice Smith, PhD, Evelyn Gopez, MD, Barbara Chadwick, MD, Brian Collins, MD

Department of Pathology, University of Utah, Salt Lake City, Utah

Introduction: UroVysion FISH testing on urinary tract specimens utilizes a spectrum of labeled probes to identify chromosomal abnormalities present in urothelial neoplasia. The multi-color, 4-probe mixture can detect amplifications of chromosomes 3, 7, and 17, or the deletion of 9p21. FISH findings are interpreted by fluorescent microscopy and a variety of methods are available for evaluation of results. Manual review and cell scoring with a fluorescent microscope is a commonly used technique. Image analysis systems are available to assist in FISH interpretation. Since introducing UroVysion FISH testing on urinary tract specimens in the cytopathology laboratory, both manual and an image guided system method have been employed for interpretation. The results and features of the two methodologies are compared.

Materials and Methods: UroVysion FISH testing on urinary tract specimens was introduced in the cytopathology laboratory and results were initially reported by a manual review method. All four probe signals were evaluated in at least 25 cells, with preference given to those cells having the most abnormal nuclei revealed by DAPI staining. As the testing volumes increased, an imaged guided system was introduced and utilized for interpretation. The system was the BioView Duet™ which consists of a fluorescent microscope, automated image capture hardware and software. The BioView Duet imaging system scans the slide, and finds, classifies, and captures digital images of cellular FISH targets, which are then presented in a gallery on a screen for review. The result for both methods classifies the findings in a quantitative manner which is reported as negative, positive, equivocal or unsatisfactory. An informational database search was performed and segregated into either the manual review or image capture method, and then the results of each were recorded.

Results: The manual cases occurred over a three and a half year time span after the initial introduction of the testing to the laboratory. There were 1509 cases by manual examination. The results were reported as negative (77.6%), positive (17.1%), unsatisfactory (5.1%), and equivocal (<1%). The image guided system cases occurred over a two year time span. There were 2701 cases by image guided system examination. The results were reported as negative (68.9%), positive (15.2%), unsatisfactory (13.1%) and equivocal (1.5%). A change in the methodology of specimen screening for adequacy prehybridization contributed to an increase in unsatisfactory results for the image guided system. Interpretation time commitment varies between methodologies. In general, pathologist manual review requires twenty to thirty minutes and an image guided review requires five to ten minutes.

Conclusions: The percentage of positive results showed a good correlation between methods. The difference between negative reporting rates could be attributed to differences in the classification of unsatisfactory specimens. A variety of pre-analytic and pre-hybridization factors can significantly impact reported unsatisfactory rates. Considering the differences in time commitment between the methods, a high volume service could benefit from an image guided system. In our experience, there is good correlation between the manual method and image guided method for the interpretation of UroVysion FISH testing of the urinary tract.

19

Cytologic Features of t(6;11)(p21;q12) Renal Cell Carcinoma

Jason Carvalho, MD, Stewart Knoepp, MD, PhD

Pathology, University of Michigan, Ann Arbor, Michigan

Introduction: Renal cell carcinomas with translocations resulting in TFEB gene fusion, t(6;11)(p21;q12), are recently discovered translocation-associated renal cell carcinomas of uncertain histogenesis with distinctive clinical and histopathologic features. The predominant cell type in these tumors are clear to oxyphilic, large epithelioid cells that show focal immunohistochemical staining for HMB45 and Melan A and a lack of staining for epithelial markers. Although these tumors share some histopathologic overlap with Xp11 translocation tumors, unlike Xp11 tumors they occur primarily in pediatric populations, and recent evidence suggests they behave less aggressively. To date, the cytomorphology of these tumors has not been fully characterized.

Materials and Methods: We report the case of a 16-year-old female with unremarkable past medical history who presented with hematuria, dysuria, urinary frequency and flank pain. A CT scan for suspected renal calculi (renal stone protocol) demonstrated a 12.5 cm heterogeneously-enhancing renal mass with peripheral calcifications. The mass was displacing the renal hilum but was confined to the kidney. The patient underwent a radical nephrectomy and an intraoperative fine-needle aspiration was performed. Fine-needle aspiration was performed with a 22-guage needle and air-dried (DiffQuik) and alcohol smears (Hematoxylin & Eosin and Papanicolaou) were prepared.

Results: Cytomorphologic analysis demonstrates cellular smears consisting of large tissue fragments and sheets of cells composed of three distinct cell types. The predominant cell type is characterized by abundant clear to granular cytoplasm and indistinct cellular borders; these cells show nuclei with vesicular chromatin and inconspicuous nucleoli. These cells corresponded to epithelioid cells seen on the surgical pathology resection specimen that immunohistochemically show focal HMB45 and Melan A staining. A second population of cells consists of smaller cells with very high nuclear to cytoplasmic ratios and hyperchromatic nuclei. The smaller cells occasionally form groups that surround distinctive hyaline globules. The final distinct cell type is defined by a very large, translucent, wrinkled, naked nucleus with a minute nucleolus that abuts the nuclear membrane. None of the smears showed necrosis or cells with mitoses. Psammoma bodies were not present.

Conclusions: Renal cell carcinoma with associated t(6;11)(p21;q12) shows unique cytomorphologic characteristics exemplified by three distinctive cell types that are highly specific for this entity. Unlike Xp11 translocation-associated renal cell carcinomas, t(6;11)-associated renal cell carcinomas do not exhibit psammomatous calcification, frequent mitoses or necrosis. The frequent presence of clear cells may also cause confusion with clear-cell renal cell carcinoma, but the highly distinctive cytologic features allow differentiation and definitive diagnosis of this important entity.

20

Fine Needle Aspiration and Core Biopsy of Small Renal Masses: Need to Revisit Their Roles in Renal Cytology

Anupma Nayak, MBBS, MD¹, Niti Manglik, MBBS¹, Chiara Sugrue, MBA, MS, SCT¹, Patricia Wasserman, MD¹, Igor Lobko, MD³, David Siegel, MD³, Gary Weiss, MD², Nora Morgenstern, MD¹

¹Pathology, NSLIJHS-Albert Einstein College of Medicine at Long Island Jewish Medical Center, New Hyde Park, New York; ²Urology, NSLIJHS-Albert Einstein College of Medicine at Long Island Jewish Medical Center, New Hyde Park, New York; ³Radiology, NSLIJHS-Albert Einstein College of Medicine at Long Island Jewish Medical Center, New Hyde Park, New York

Introduction: Recent advances in imaging techniques and treatment modalities enforce the need of revisiting the role of fine needle aspiration (FNA) cytology and core biopsy in diagnosing small renal masses. The objective of this study is to evaluate the efficacy of renal FNA and core biopsy in diagnosing small renal masses and to propose diagnostic categories for renal FNA cytology.

Materials and Methods: Retrospective review of renal FNAs and core biopsies performed at our center between 2001 and 2010 in patients with renal masses <4 cm in size. The smears were reviewed by 3 cytologists (AN, NJM, and CS) and correlated with tissue sections available from cell blocks, core biopsies and follow up surgical resections. Lesions were classified into 6 major categories: unsatisfactory, benign non-neoplastic, benign neoplastic, suspicious for malignancy, suspicious for oncocytic neoplasm and diagnostic for malignancy. Diagnostic efficacy of FNA and core biopsy was calculated in a subset of patients who underwent both FNA and a core biopsy.

Results: A total of 161 renal FNA and/or a core biopsy were accessioned during this period. Indications are shown in Figure-1. Of the 161 cases, 135 had only FNA, 24 had both FNA and core biopsy and 2 had only core biopsy. Touch imprints were prepared on-site from the latter 2 cases. On-site adequacy on Diff-Quik stained smears was performed in all the cases by a cytotechnologist. Immunohistochemistry was performed either on cell block or a core biopsy whenever needed. Out of 161 cases, 37 (22.98%) were deemed unsatisfactory consisting of blood, fibro adipose tissue, normal glomeruli or tubules. Of the remaining 124 cases, 18 were classified as benign non-neoplastic, 5 as benign neoplasm, 18 suspicious for malignancy due to scant material, 69 diagnostic for malignancy and 14 suspicious for oncocytic neoplasm (Table-1). Of the 24 cases with both FNA and core biopsy, a diagnosis was rendered in 23 (95.83%) cases. Only 1 (4.17%) case was called unsatisfactory for evaluation. In 4/23 cases, core biopsy was non diagnostic and the diagnosis was based on the cellular material present on FNA smears. Conversely, there was only 1 case where core biopsy contributed to the diagnosis and FNA was non diagnostic (Table-2). The diagnostic efficacy of isolated FNA procedure and core biopsy was 76.73% and 80.76%, respectively. Seemingly higher efficacy of core biopsy in this study may be due to lesser number of biopsies as compared to FNAs. Only 18 cases had a surgical follow up available in this study (Table-3). Two unsatisfactory cases were diagnosed as clear cell RCC on histology and 3 suspicious cases were confirmed as clear cell RCC (2) and papillary RCC (1), respectively. One of the clear cell RCC case was diagnosed as Papillary RCC on histology. And, a diagnosis of oncocytic neoplasm was equally split between oncocytoma and oncocytic variant of papillary RCC on histopathology. We are presently working on formulating a strict on-site adequacy criterion for renal FNAs at our center based on this study.

Conclusions: FNA biopsy with or without a core biopsy is an excellent tool for diagnosing small renal masses. Both the techniques can complement each other in difficult cases. Performing an on-site cytologic adequacy and classifying the lesions under 6 major categories is encouraged as it guides the clinician to alter the management.

Figure 1 . Indications for FNA or core biopsy.

Table 1. Cytologic Diagnosis in 161 cases under 6 major categories

Diagnostic Category	No. of Cases
Unsatisfactory	37
Benign (n=18)
Cyst contents	11
Inflammatory (Abscess, chronic inflammation, etc.)	7
Benign neoplasm (n=5)
Angiomyolipoma	3
Oncocytoma	2
Other rare tumors	0
Suspicious (n=18)
Suspicious for malignant renal neoplasm	14
Suspicious for oncocytic neoplasm	3
Suspicious for malignant urothelial neoplasm	1
Diagnostic for malignant neoplasm with subtyping (n=69)
Lymphoma	6
PNET	1
TCC	8
Papillary RCC	7
MTSCC	1
Oncycytic variant of chromophobe carcinoma	4
Clear cell RCC	41
Adenocarcinoma with papillary features	1
Oncocytic neoplasm	14
Total	161

Table 2. Comparison of diagnostic efficacy of FNA and Core biopsy in 161 Cases

Biopsy Technique	Diagnostic	Non Diagnostic	Diagnostic (%)
*NA=Not applicable.
FNA only (n=135)	100	35	74.07
Core Biopsy only (n=2)	2	0	100
FNA plus Core Biopsy (n=24)	23	1	95.83
FNA non diagnostic and core  diagnostic (n=1)	1	0	NA
FNA diagnostic and core  non diagnostic (n=4)	4	0	NA
Both FNA and Core  non diagnostic (n=1)	0	1	NA

Table 3. Cytologic-histologic correlation in 18 cases with surgical followup

Cytological diagnosis (No. of cases)	Surgical diagnosis (No. of cases)
TCC (n=3)	Urothelial carcinoma (n=3)
RCC clear cell (n=6)	RCC clear cell (n=5)Papillary RCC (n=1)
Oncocytic neoplasm (n=2)	Oncocytoma (n=1)Oncocytic variant  of papillary RCC (n=1)
PNET (n=1)	PNET (n=1)
Lymphoma (n=1)	Lymphoma (n=1)
Non diagnostic (n=2)	Clear cell RCC (n=2)
Suspicious for RCC (n=3)	Clear cell RCC (n=2)Papillary RCC (n=1)

21

Atypical Squamous Cells in the Urine: A Cause of False Positive Urine Cytology

Jill Ono, MD¹, David Kindelberger, MD², Vikram Deshpande, MD¹

¹Pathology, Massachusetts General Hospital, Boston, Massachusetts; ²Pathology, Brigham and Women's Hospital, Boston, Massachusetts

Introduction: The sensitivity and specificity of urine cytology for high grade urothelial neoplasms is excellent. High grade urothelial lesions occasionally demonstrate squamous differentiation, which can readily be detected on cytologic analysis and false positive diagnoses are rare. Herein, we report ten cases with a cytologic diagnosis suggesting transitional cell carcinoma (TCC) with squamous differentiation that on prolonged follow-up were negative for malignancy.

Materials and Methods:53 cases with cytologic diagnosis of “suspicious” or “positive” for TCC with squamous differentiation were identified in the archival records from the Massachusetts General Hospital and Brigham and Women's Hospital between 1993 and 2008. A diagnosis of TCC could not be confirmed in 10 cases. Ten cases with positive cytology and histologic confirmation served as a control group. The cytomorphological features were evaluated by all 3 observers. Clinical records were evaluated in all cases.

Results: All ten cases of TCC with negative follow-up were from male patients, ranging in age from 34 to 86 years (median age = 58). The control group was statistically older (p=0.04), with a median age of 66.5 years. Multiple diagnostic procedures including cystoscopy, CT imaging, and bladder biopsies failed to identify malignancy in the urothelial tract. The duration of clinical follow-up ranged from 22 to 94 months (median = 37 months). One patient was found to have nephrolithiasis and the urine of one patient was detected to be positive for high risk HPV. Cytomorphologically the squamous component of the study group lacked bizarre keratinized 'tadpole' and 'fibre' cells, while these bizarred keratinized cells were noted in seven (70%) of the biopsy proven TCCs. Atypical epithelial cells were seen in both the study (n=9) and control groups (n=10). Moderate and severely atypical non-keratinized cells were seen in both the study group (n=5) and the control group (n=8).

Conclusions: Atypical keratinized cells are a cause of false positive urine cytology, predominantly in middle aged men. Potential etiological factors include urothelial condyloma and urinary stones. Although these specimens show overlapping cytological features with TCCs, the absence of bizarre keratinized cells may help prevent false positive diagnoses.

22

The Hopkins Template for Urinary Cytology Interpretations

Dorothy Rosenthal, MD, Frances Burroughs, SCT(ASCP), CT(IAC), Douglas Clark, MD

Pathology, Johns Hopkins Hospital, Baltimore, Maryland

Introduction:Purpose: Determine the effectiveness for patient management of a novel template of diagnostic categories developed and used at Johns Hopkins Hospital by the John K. Frost Cytopathology Laboratory, for interpreting samples from the urinary bladder, ureters and urethra.

Background: In most cytopathology laboratories that process a variety of samples, those from the urinary tract are second only to Pap tests in annual volume. With such familiarity, one would expect that performance parameters of urine cytology would be exceptionally high. Quite the contrary! As a result, faith by some in the results of urinary samples may be low. In discussions with our urologists, we learned that the most important indicator for cystoscopy was a cytologic diagnosis of High Grade Urothelial Carcinoma (HG-UC). We designed our template in order to standardize our diagnostic categories to enable our clinicians to uniformly manage their patients based on the report generated by the sample they sent us.

Materials and Methods:Construction of the Template: The diagnostic template is based in part on The Bethesda System for Gynecologic Cytology. We included reactive/inflammatory changes in the negative group (NUAM); considered Atypical Urothelial Cells of Undetermined Significance (AUC-US) akin to Atypical Squamous Cells of Undetermined Significance (ASC-US); created the category of AUC-H, Atypical Urothelial Cells, favor High Grade carcinoma; and added the category of Urothelial Carcinoma, with the choices of Low and High Grade.

Data Collection: The Pathology Data System at the Johns Hopkins Hospital was searched for cases that met the following criteria over a period from July 1, 2007 to June 30, 2009: All surgical specimens with the diagnosis of high-grade urothelial carcinoma, regardless of invasion status; cytologic specimens matched with the biopsies by medical records number (MRN) during the same time period; all surgical specimens following a cytologic diagnosis of AUC-US and AUC-H.

Results: Over two thirds of patients with biopsy confirmation of high grade urothelial carcinoma had a preceding cytologic diagnosis of AUC-H or HGUC (Table 1). Of those patients with the diagnosis of AUC-H, 38% had a urothelial cancer discovered at biopsy compared with only 10% of those with an AUCUS diagnosis (Table 2).

Conclusions: Our template is effective in targeting those patients who need to be cystoscoped. Further analysis is in process to determine what other factors were present that directed clinicians to biopsy patients in the AUC-US or NUAM categories.

Table 1. Interpretations of GU cytologic samples for FY2008-2009 with biopsy data

Cyto Dx	#cases	%total	BX	%BX	%CA
%BX = percent of each category that was biopsied %CA = percent of HG-UC on biopsy
NUAM	2894	60	180	6	42
AUC-US	1246	26	183	14	42
AUC-H	237	5	82	34	69
CA	283	6	143	50	74
Unsat	193	4	13	6	11
TOTAL	4853	100	601	12	52

Table 2. Patients with AUC-US and AUC-H with resultant biopsy diagnoses, FY 2008-2009

BIOPSY	AUC-US 607 pts	AUC-H 118 pts	TOTAL 725 pts
% = biopsy diagnosis/# patients in each cytology category.
Benign	52pts 8%	17pts 14%	69pts 10%
Lo grade	14pts 2%	6pts 5%	20pts 3%
Hi grade	49pts 8%	39pts 33%	88pts 12%
TOTAL	114pts 18%	62pts 53%	176pts 24%

23

Utility of Fine Needle Aspiration and Core Biopsy with Touch Prep in the Diagnosis of Renal Lesions

Lori Scanga, MD, PhD, Susan Maygarden, MD

Cytopathology, The University of North Carolina, Chapel Hill, North Carolina

Introduction: Cytologic evaluation by fine needle aspiration (FNA) and core biopsy (CB) with touch prep (TP) is used in the diagnosis of renal lesions.

Materials and Methods: We reviewed the FNAs and CB with TP of renal lesions performed at our hospital since 2004. The cytology diagnoses were correlated with the radiological features, surgical specimens, and clinical course.

Results:98 procedures (63 FNA, 15 FNA+CB, 15 CB+TP, 5 FNA+CB+TP) were performed for renal lesions with radiological features of malignancy (73), indeterminate for malignancy (5), benign cysts (19), or diffuse renal parenchymal abnormality (1). The procedures were satisfactory in 88% of FNAs (73/83), 90% of TPs (18/20), and 89% of CBs (31/35). The specimen types were diagnostic of neoplasm in lesions with malignant or indeterminate radiological features (78) in 71% of FNAs (45/63), 90% of TPs (18/20), and 89% of CBs (31/35). The lesions with benign radiological features (simple cysts and diffuse renal parenchymal abnormality) were evaluated by FNA only, and all were satisfactory and negative for malignancy. Cytology diagnoses of the lesions with concerning radiological features (78) were: malignant primary epithelial neoplasm (50), malignant primary non-epithelial neoplasm (3), metastatic neoplasm to kidney (6), atypical (3), primary epithelial neoplasm indeterminate for malignancy (2), oncocytoma (3), negative (7), and unsatisfactory (4). 26 cases had corresponding surgical specimens, 25 were concordant and one was false negative. The lesions evaluated by cytology only without corresponding surgical specimens all correlated with clinical course. Treatment in the patients with radiologically concerning lesions (78) was surgical resection (18), radiofrequency ablation (RFA) (24), systemic chemotherapy and/or radiation (30), or no treatment (6). The most common treatment of these 78 lesions was resection from 2004-7 and RFA from 2008-9. The average size of the excised lesions (9.2 cm +/- 5.9 cm) was greater than the ablated lesions (3.3 +/- 1.9 cm) (p < 0.0001).

Conclusions: Common indications for cytologic evaluation of renal lesions are confirming primary renal malignancy before resection, ablation, or systemic therapy, examining cyst fluid to exclude cystic renal cell carcinoma, and distinguishing primary from secondary neoplasm in the kidney. For lesions with malignant or indeterminate radiological features, core biopsy had higher yield than FNA alone (p=0.04). Cytology diagnoses were highly accurate when correlated to surgical specimens. At our hospital treatment of renal lesions with malignant or indeterminate radiological features following cytologic evaluation shifted from resection of large lesions to ablation of smaller lesions over the past 6 years.

24

Evaluation of Urinary Diversion Specimens by Routine Cytology and Fluorescence In-Situ Hybridization

Jesse Voss, CT MB(ASCP)¹, Amy Clayton, MD¹, Michael Henry, MD¹, Jun Zhang, MD¹, R. Jeffery Karnes, MD², Kevin Halling, MD, PhD¹

¹Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota; ²Department of Urology, Mayo Clinic, Rochester, Minnesota

Introduction: Cytologic evaluation of urinary diversion specimens (stoma, ileal conduit, neobladder, etc.) can be diagnostically challenging, but plays an important role in the surveillance of patients following cystectomy. The goal of this study was to evaluate the performance of routine cytology and fluorescence in situ hybridization (FISH) on urine obtained from patients with a urinary diversion.

Materials and Methods: A retrospective medical record review identified 38 urine specimens collected from 28 patients with a urinary diversion that were analyzed by routine cytology and FISH. Specimens were equally split for routine cytology and FISH testing. Routine cytology was performed on each pap-stained slide prepared either by the filter or ThinPrep® methods (Hologic, Inc., Bedford, MA). Each specimen was screened by a cytotechnologist then interpreted by a cytopathologist. FISH was performed using the UroVysion® probe set (Abbott Molecular Inc., Des Plaines, IL) and specimens were considered abnormal when a cytotechnologist identified four or more cells with gains (>2 signals) in at least two of the four probes. The percentage of 100 consecutive urothelial cells that showed FISH abnormality was determined for each FISH positive case. Patients were considered to have recurrent or metastatic urothelial carcinoma when urine cytology, fine needle aspiration of a distant site, or urinary biopsy specimen was diagnosed as positive during routine clinical follow-up.

Results: Twenty-one of the 38 specimens (55%) had a corresponding follow-up specimen diagnosed as urothelial carcinoma. A majority of follow-up cancers (n=12; 57%) were diagnosed by FNA of a distant metastasis, 5 (24%) were positive on subsequent urine cytology and 4 (19%) patients had a positive upper tract biopsy. FISH detected significantly more urothelial carcinomas than routine cytology (sensitivity 47.6% versus 9.5%; P=0.005), while maintaining high specificity (94.1 versus 100% for FISH and cytology, respectively). Three specimens were diagnosed as suspicious by cytology and all were found to have urothelial carcinoma. If suspicious cytology diagnoses were considered as positive for malignancy, the sensitivity of routine cytology increased to 23.8%, but was still significantly lower than FISH (P=0.025). FISH was positive in 6 of the 12 specimens from patients found to have distant metastatic disease, while cytology detected one. FISH identified very few (<2%) abnormal cells in the one false positive specimen, but was highly abnormal in the 10 specimens that were found to have recurrent or metastatic disease (mean = 22%, range 4-51%).

Conclusions: These preliminary data suggest that FISH has higher sensitivity on urinary diversion specimens compared to standard routine cytology. Prospective studies are needed to determine whether FISH can be useful as part of routine clinical practice for the diagnosis of carcinoma from patients with urinary diversions.

Gynecologic

25

Atypical Glandular Cells; A Diagnostic Dilemma: A Three Year Retrospective Study of Liquid Based Pap Tests with Cytohistologic Correlation

Israh Akhtar, MD, Maud Major-Nicolas, MD, Ric Bowlin, SCT, Rhonda Alexander, SCT, Anwer Siddiqi, MD, Rhyne Flowers, MD, Mithra Baliga, MD

Cytopathology, University of Mississippi Medical Center, Jackson, Mississippi

Introduction: The category of Atypical Glandular Cells (AGC) in the cervicovaginal Pap tests still remains a diagnostic challenge to both clinicians and cytopathologists because of the lack of uniform cytologic criteria, lack of interobserver agreement in the diagnosis and lack of standardized patient management guideline.

Materials and Methods: We undertook a retrospective study collecting and analyzing all AGC diagnoses occurring in a 39 month period, from January 2007 to April 2010 in our institution. During this time period, there were 35,918 Pap tests performed by liquid based cytology. 36 cases were diagnosed as AGC and were subcategorized into different categories according to The Bethesda 2001.

Results:36 AGC diagnoses were retrieved from the archives during this 39 month period amongst over 35,000 Pap tests. These females ranged in age from 29 to 82 years with a mean age of 55.5 years and median age being 57, 58. The most common category into which the AGC was subclassified was “not otherwise specified” (n=13). AGC, favor endometrial pathology, was the next common subcategory (n=11). The remainder 12 cases were subcategorized into, AGC, favor endometrial adenocarcinoma (n=3), AGC, favor neoplasia (n=7), and AGC, favor metastatic malignancy (n=2). 30 out of 36 cases had a histologic follow up (83.3%) while 6 patients had no follow up (16.7%).

Conclusions: Majority of these patients were followed up by colposcopy, endocervical curettage or more commonly endometrial biopsy. 8 of 36 (22.2%) patients had a benign pathology (for example endometrial breakdown, endometrial polyp etc.) while the remainder (78.8%) had a more serious pathologic diagnosis (such as endometrial adenocarcinoma). Thus, the category of AGC represents a cytologic diagnosis associated with a high probability of underlying pathologic conditions especially in the perimenopausal age group. Even though it constitutes a minor percentage of Pap test diagnosis, averaging 0.1-2.5 % nationwide, and 0.1% in our study, yet it is of utmost importance for a timely intervention of any underlying serious glandular abnormalities related to female genital tract.

26

Effect of TIS on Laboratory Rate of ASC-H and HGSIL Interpretations

Jenna Benson, CT(ASCP), Donna Russell, CT(ASCP), Thomas Bonfiglio, MD, Sharlin Varghese, MD

Cytopathology, University of Rochester Medical Center, Rochester, New York

Introduction: Automated screening of Thin Prep® Papanicolaou Tests has become increasingly common in clinical practice. Increased productivity has initiated laboratory use of the Thin Prep® Imaging System (TIS). Increased sensitivity is a potential added benefit of TIS. Published studies have shown an increase in discovery of dysplastic cells. This study evaluates the effect of TIS on the incidence of atypical squamous cells, high grade squamous intraepithelial lesion not excluded (ASC-H) and high grade squamous intraepithelial lesion (HGSIL) results on Thin Prep® Pap Tests by comparing TIS-assisted and manual screening findings and the diagnoses on subsequent follow-up in a screening population over a one-year time period.

Materials and Methods: A compilation of all ASC-H and HGSIL cases was prepared by conducting a computerized search over a one-year period (7/06-6/07). The accumulated cases include Thin Prep Pap tests that were both TIS and manually screened. Follow-up results of cytologic and histologic cervical specimens were obtained for a time period extending to 2010. Interpretation utilizing TIS was in place ten months prior to the study's initiation.

Results: During the study period 68,892 Pap tests were performed in our laboratory. One third (33.5%) of Pap tests were screened with assistance of TIS. Manual screening was performed on 45,789 Pap tests of which 174 (.4%) were interpreted as ASC-H and 193 (.4%) were interpreted as HGSIL. During the same time period automated screening (TIS) was performed on 23,103 Pap tests. Interpretation of 46 (.2%) cases provided an ASC-H result, while 60 (.3%) were HGSIL. Follow-up cervical dysplasia by colposcopic biopsy and cone biopsy was distributed proportionally between TIS and manual screening for both ASC-H and HGSIL categories. Cervical intraepithelial neoplasia (CIN II/III) was identified on follow-up biopsy of 35% TIS cases (57% all CIN) and 30% manually screened cases (56% all CIN) for ASC-H. In the HGSIL subset 58% of TIS cases (73% all CIN) and 59% manually screened cases (72% all CIN) showed CIN II/III on follow-up. The differences were statistically insignificant (Tables 1 and 2).

Conclusions: Our laboratory was not significantly affected by the introduction of the Thin Prep Imaging System pertaining to rates of ASC-H and HGSIL. Use of the imager facilitates efficient screening, thus providing the cytotechnologist a reliable tool for Pap test assessment.

Table 1. 

Histology	ASC-H IM		ASC-H Without IM
	# Cases	%	# Cases	%
* Histologic follow-up unavailable, Cytologic follow-up available.
HGSIL	16	35	53	30
LGSIL	10	22	45	26
NEG	8	17	34	20
N/A*	7	15	30	17
Unknown	5	11	11	6
EM Adeno	0	0	1	1
Total	46	100	174	100

Table 2. 

Histology	HGSIL IM		HGSIL Without IM
	# Cases	%	# Cases	%
* Histologic follow-up unavailable, Cytologic follow-up available.
HGSIL	35	58	113	59
LGSIL	9	15	26	13
NEG	6	10	21	11
N/A*	10	17	18	9
Unknown	0	0	15	8
Total	60	100	193	100

27

Abnormal Pap Tests in Women 70 Years and Older With Surgical Follow-Up: A 5 Year Retrospective Review at William Beaumont Hospital, Royal Oak, Michigan

Melissa Bowman, MD, Donna Cary, BS, CT (ASCP), Jeanne Jax, BS, CT (ASCP), Tomi Kuntzman, DO

Anatomic Pathology, William Beaumont Hospital, Royal Oak, Michigan

Introduction: The use of the Papanicolaou (Pap) test in screening for squamous cell carcinoma of the cervix has significantly impacted the morbidity and mortality associated with the disease. The American Cancer Society (ACS) guidelines for the discontinuation of screening in the geriatric population state that women 70 years and older, who have had 3 consecutive normal Pap tests and no abnormal Pap test in the last 10 years can discontinue screening. We reviewed a series of Pap test results in women 70 years and older and correlated presenting symptoms, types of gynecologic malignancy, and prior cervical screening history in those with clinically significant lesions on surgical follow-up.

Materials and Methods: The electronic archive was searched for Pap test results and surgical follow-up in women 70 years and older from 7/1/2004-6/30/2009. A single Pap result on each patient during the 5 year time period was included (either one of the patient's multiple normal results or the patient's first abnormal result). The symptoms and ages in women with significant findings on surgical follow-up were assessed. Pap tests collected prior to the patient's first abnormal result were compared with ACS criteria for screening cessation.

Results: In a coinciding 5 year interval (1/1/2004-12/31/2008) the laboratory at William Beaumont Hospital, Royal Oak performed 325,580 Pap tests, of which 17,623 were interpreted as abnormal (5.41%). In 9,879 women age 70 years and older, 14,354 Pap tests were performed during our 5 year review. Of the 14,354 cases, 302 were abnormal (2.1%). 241 women had at least one abnormal Pap test: ASC-US (59), LGSIL (62), ASC-H (26), HGSIL (13), AGC (62), and malignant (19). Of these women, 123 had surgical follow-up. 41 patients (mean age 77.68) had clinically significant lesions: primary uterine carcinoma (27; mean age 78), primary cervical carcinoma (6; mean age 79), primary ovarian carcinoma (2; mean age 71.5), and cervical/vaginal high grade dysplasia (6; mean age 76). The primary uterine carcinomas included endometrioid (10), serous papillary (8), mucinous (1), endometrial not otherwise specified (1), malignant mixed Mullerian tumor (4), and undifferentiated carcinoma (3). 28 of the 41 women had recorded genitourinary symptoms: abnormal bleeding (23) or abdominal pain (5) while 8 were asymptomatic. 20 of the 41 women had no prior Pap test on record. 5 women had prior abnormal Pap tests in our records; 16 had only normal results but lacked at least 3 consecutive normals (mean 2.44) or did not have 10 years of history to review (mean 5.94). None of the 41 women met all of the ACS criteria for Pap test cessation.

Conclusions: The most common malignancy identified in this study was primary uterine carcinoma. Of patients with significant findings the majority presented with genitourinary symptoms and almost half had no prior Pap test on record, including 50% of those with cervical carcinoma. The ACS recommendations on cervical screening played no role in the discovery of neoplastic lesions in our patient population. While the purpose of the Pap test is to screen for primary cervical lesions, extracervical neoplasms can often be detected. Therefore, the Pap test can serve as a useful diagnostic tool in this older population. High clinical value should be placed on presenting symptoms and a lack of prior medical evaluation.

28

HPV Testing and the ASCCP Guidelines: Are They Followed? An Academic Experience

Leigh Ann Cahill, BS, CT(ASCP)CMIAC, William Frable, MD, Celeste Powers, MD, PhD

Department of Cytopathology, Virginia Commonwealth University Health System, Richmond, Virginia

Introduction: Human Papilloma Virus (HPV) testing on cervical specimens diagnosed as Atypical Squamous Cells of Undetermined Significance (ASCUS) has become standard practice based on data from the 2000 ALTS Trial. In 2006, the American Society of Colposcopy and Cervical Pathology (ASCCP) developed practice guidelines for the management of women with abnormal Papanicolaou (PAP) tests. This current study reviews our HPV test results and how the ASCCP guidelines are followed.

Materials and Methods: The cervical specimens were SurePath liquid base preparations. The HPV tests were performed by a reference laboratory utilizing the HPV DetxTM High Risk Assessment by PCR method. The HPV tests were ordered reflexively with an ASCUS diagnosis or by clinician request. We reviewed the available prior and follow biopsy and cytology reports with relation to the index cytology and HPV tested specimens from 2004 to 2006. This study concentrated on patients with positive HPV results.

Results: Of the 563 patients with HPV tests, 271 (48%) had positive HPV tests. 199 were positive for high risk virus, 35 low risk and 37, risk unknown. A preceding negative cytology or no prior cytology was found in 146 cases (54%). 23 patients had a prior cytology of ASCUS, 86 of LGSIL, 3 of ASC-H and 13 HGSIL. The initiating smear for ordering the HPV test was negative in 42 patients, ASCUS in 140 cases, LGSIL in 81 cases, ASC-H in 5 cases and HGSIL in 3 cases. 161 patients (60%) had concurrent follow-up biopsies. These biopsies were negative in 66 cases, LGSIL in 65 patients, HGSIL in 32, VIN in 4 and VAIN in 1 patient. The diagnosis of a precancerous lesion prior to the index cytology and HPV testing was CIN I in 34 cases, CIN II-III in 15 patients, VAIN I in 3 patients and VIN III in 2 cases. Table 1 depicts the results for the number of patients with ASCUS, HPV+ high risk and HPV+ low risk or risk unknown who were followed according to the ASCCP guidelines. 53 patients (20%) had no follow-up cytology or biopsies. This occurred in 11 patients whose index smear was negative, 31 cases of ASCUS, 9 cases of LGSIL, and 2 cases ASC-H.

Conclusions: Observing the ASCCP guidelines after an index smear of ASCUS and HPV test indicates that the appropriate follow-up occurred in 52.5% of the HPV+ high risk cases and 58.5% of the HPV+ low risk (or risk unknown) cases. Combining these two groups of HPV+ patients, 55.5% had the appropriate follow-up.

Table 1. ASCUS Cases Positive for HPV High Risk and Low Risk Followed According to ASCCP Guidelines

ASCUS HPV+ Risk Factor	ASCCP GuidelinesFollowed	Method of Follow-up	Number of cases (%)
ASCUS HPV+ High Risk (N=99)	Yes	Colposcopy	52/99 (52.5%)
	No	No Colposcopy/Cytology  Follow-up	47/99 (47.5%)
ASCUS HPV+ Low Risk/Risk  Unknown (N=41)	Yes	Follow-up Cytology	24/41 (58.5%)
	No	Colposcopy or No  Follow-up	17/41 (41.5%)

29

Use of Dual-Stain for P16 and Ki-67 in the Interpretation of Abnormal Pap Cytology Results: A Prospective Study

Mamatha Chivukula, MD¹, Marshall Austin, MD, PhD¹, Jonee Matsko, CT(ASCP)¹, Alexander Duwe, MS², Teresa Freidman, CT (ASCP)¹, Nancy Mauser, CT (ASCP)¹, Rudiger Ridder, PhD²

¹Pathology, Magee Women's Hospital of UPMC, Pittsburgh, Pennsylvania; ²mTm Laboratories, Heidelberg, Germany;

Introduction: p16 has been found to be strongly over-expressed in nearly all high-grade pre-cancerous and cancerous cervical lesions and may serve as a surrogate marker for the transforming activity of high-risk HPV. As over-expression of the cell-cycle regulatory protein p16 in cells with intact cell cycle regulation should prevent those cells from proliferating, we further tested the hypothesis that the detection of individual cells simultaneously co-expressing p16 protein and proliferation marker Ki-67 can be used as an indicator for the presence of cervical dysplasia. To accomplish this goal we initiated a large prospective study on cervical cytology specimens showing Pap abnormalities.

Materials and Methods: Residual materials from liquid-based cytology specimens of women attending cervical cancer screening at a major tertiary hospital and reference center over one year period were used for the analysis. Specimens with any abnormal Pap cytology result (ASC-US+) were included. For each case, an additional ThinPrep™ slide was prepared and immuno-stained using a prototypic dual staining reagent kit (CINtec®Cytology, Dual stain) for the simultaneous detection of p16 and Ki-67 expression on the same slide. The presence of one or more individual cells co-expressing both p16 and Ki-67 were interpreted as “positive” test result. Follow-up biopsy and HPV results were obtained.

Results: Sensitivity of the Dual stain was 87.0% for CIN2+ and 95.7% for CIN3+, with specificity of 89.8% for non high-grade CIN. In ASC-US/ASC-H/LSIL categories with positive HPV results, a positive Dual stain result identified all CIN3+ cases at high levels of specificity (up to 80%).

Conclusions: Initial results from a first set of cytology specimens subjected to simultaneous p16/Ki-67 dual staining and with biopsy follow-up (n=661) indicate both high sensitivity and specificity of this novel screening approach to detection of CIN2/3+ on biopsy follow-up. Results showing high specificity rates for the Dual stain support this approach as an enhancement for detecting histopathological CIN2/3+.

30

Digital Slide Imaging in Cervicovaginal Cytology, a Pilot Study

Debora Smith, CT, Bakula Dhurandhar, CT, Todd Fairley, CT, Stacie Churchwell, CT, Subhendu Chakraborty, MS, Dina Mody, MD, Donna Coffey, MD

Pathology, The Methodist Hospital, Houston, Texas

Introduction: Full digital slide imaging is considered by some to be the future of anatomic pathology where glass slides will be replaced by scanned images. Although this has been very successful for small anatomic pathology specimens, its usefulness and clinical application have been limited in cytology for several reasons. These include unavailability of the Z axis and the rather large unmanageable size of the digitized files. More recently at least three systems are available in the United States for full glass slide imaging and the Z axis is finally available for scanning of cytology slides.

Materials and Methods: We evaluated 11 cervicovaginal cytology cases (ThinPrep and SurePath) scanned at 20x, 40x, 40x with a Z stack and manually using the Bioimagene IsCAN Coreo 3.0 scanner. These cases were unidentified and each scan was considered one case. Four cytotechnologists were blinded to this information and they evaluated each scan as a separate case, unaware that the same case was being repeated at different magnifications in a randomized sequence. The diagnoses as well as the time it took to reach a diagnosis were recorded. After completing all digital slides, the cytotechnologist also evaluated the corresponding glass slides, recording diagnosis and time it took to screen the slides.

Results: The results of the scans were tabulated and are as depicted in Table 1.

Conclusions: Full slide imaging is an available option especially for the purposes of teaching, diagnoses, consultations and as a means of archiving cases. However, considering the file size and the time it took to get to the diagnosis on digital images, it is not yet ready for daily diagnostic cytology/screening use. Among the digitally imaged slides, the 40x or 40x with a Z stack images had an increased diagnostic accuracy. However, the best results and lowest average time of evaluation were still obtained with glass slides.

Table 1. 

ReferenceDiagnosis	Accuracy at 20x	Accuracy at 40X	Acc 40Xwith Z stack	Glass slide
*Two cases (herpes and AIS) were inadvertently not scanned on 40XZ which may have skewed the time data as these cases had longer evaluation times on scanned slides. (TP) ThinPrep (SP) SurePath.
Ca Endometrial (SP)	1/4 (25%)	1/4 (25%)	4/4 (100%)	4/4 (100%)
LSIL (SP)	3/4 (75%)	4/4 (100%)	3/4 (75%)	4/4 (100%)
HSIL (TP)	4/4 (100%)	4/4 (100%)	4/4 (100%)	4/4 (100%)
HSIL (SP)	2/4 (50%)	3/4 (75%)	4/4 (100%)	3/4 (75%)
ASC-HVS HSIL (SP)	0/4 (0%)	1/4 (25%)	0/4 (25%)	3/4 (75%)
Atrophy (SP)	3/4 (75%)	3/4 (75%)	4/4 (100%)	4/4 (100%)
Trich (TP)	4/4 (100%)	4/4 (100%)	3/4 (75%)	4/4 (100%)
Herpes (TP)	0/4 (0%)	2/4 (50%)	N/A*	4/4 (100%)
AIS (SP)	6/8 (75%)	3/4 (75%)	N/A*	4/4 (100%)
LSIL (TP)	4/4 (100%)	4/4 (100%)	3/4 (75%)	4/4 (100%)
NILM Shift	2/4 (50%)	4/4 (100%)	4/4 (100%)	4/4 (100%)
Average time/case	4.4	3.37	2*	2.2

31

Can We Improve the Positive Predictive Value of AGC?

Rachel Mariani, MD, Kathryn McGarrah, CT, Ruth Dietrich, CT, Odile David, MD

Pathology, University of Illinois, Chicago, Illinois

Introduction: Although there are discrete criteria that define the category of Atypical Glandular Cells (AGC) in gynecologic cytology, this interpretation presents well-documented challenges related to inter-observer variability, specificity, and optimal clinical management. Often there is no significant pathology identified on follow-up colposcopy, even when the threshold for “significant pathology” is set relatively low. The purpose of this study is to determine if the current Bethesda System 2001 criteria can be further refined to increase the positive predictive value of AGC-NOS.

Materials and Methods:140 cases categorized as AGC were identified in the files of the UIC Department of Pathology between January 1, 2007 to December 31, 2008. We excluded all cases of AGC, favor neoplasia, as this category is more specific by definition and more likely to be associated with a significant clinical outcome. We also censored all AGC-EC and AGC-EM cases. Follow-up information was obtained in the course of routine Cytology Laboratory quality assurance procedures. Cases with only Pap test as follow-up were excluded. The 53 remaining cases that had histologic follow-up were reviewed. A worksheet was prepared that listed selected Bethesda System 2001 criteria for AGC (Table 1). Three experienced reviewers independently reviewed these slides without knowledge of outcome and ranked each criterion for each case individually by the following scale: 3 - prominent, 2 - present, 1 - equivocal/not observed. For statistical purposes, two or more scores of 3 per criterion for each case was considered positive. A clinical, i.e. histologic, outcome of CIN 2 or 3 or adenocarcinoma was considered positive. Simple and multiple logistic regression analysis was performed using SAS. Sensitivity and specificity for each criterion were calculated. Positive predictive value of AGC-NOS itself was also calculated.

Results: AGC-NOS had a positive predictive value of 20.8%. Although hyperchromasia and increased N/C ratio were the most specific criteria at 95.2% and 83.3%, respectively, simple logistic regression analysis showed that only nuclear enlargement had even a marginally significant association with AGC cases that had significant pathology as an outcome (p = 0.0566). None of the other criteria independently were significantly associated with outcome. Multiple regression analysis showed that four of the criteria in combination—cells occurring in sheets and strips with some cell crowding and nuclear overlap, nuclear enlargement, nuclear pleomorphism, and increased N/C ratio—discriminated outcome slightly better than the nuclear enlargement criterion alone.

Conclusions: AGC-NOS is currently a nebulous, but necessary, category in gynecologic cytopathology. Although the Pap test is a screening test and as such is expected to be more sensitive than specific, AGC-NOS too often results in interventions that fail to reveal significant pathology. Developing ancillary criteria that may be applied reproducibly might be a worthwhile albeit formidable challenge for the cytopathology community.

Table 1. Selected Bethesda 2001 Criteria for AGC

Cells occuring in sheets and strips with some cell crowding and nuclear overlap*
*Criteria significantly associated with outcome in combination by multiple logistic regression analysis.
Nuclear enlargement*
Nuclear pleomorphism*
Hyperchromasia
Nucleoli
Mitotic figures
Increased N/C ratio*
Distinct cell borders

32

Predictive Value of Reflex HPV DNA Testing in Post-treatment Follow up for Women with ASC-US Pap Results

Andrea Diaz De Vivar, MD, Marilyn Dawlett, CT (ASCP), Jian-Ping Wang, CT (ASCP), Annie Jack, CT (ASCP), Ming Guo, MD

Pathology, The University of Texas MD Anderson Cancer Center, Houston, Texas

Introduction: In cervical cancer screening, reflex human papillomavirus (HPV) DNA testing is the current standard of care for women with Pap results of atypical cells of undetermined significance (ASC-US). However, published studies on the utility of HPV DNA testing for post-treatment follow up for women with cervical cancer or high-grade precancerous lesions are scant. The aim of this study is to assess the predictive value of reflex HPV DNA testing using Hybrid Capture 2 (HC2) on ASC-US Pap specimens from women in post-treatment follow up for cervical cancer or high-grade precancerous lesions.

Materials and Methods: We retrospectively reviewed data of 1308 patients with Pap results of ASCUS from 2006 to 2009 in our center. A total of 180 patients had history of cervical/vaginal carcinoma or high-grade cervical/vaginal intraepithelial lesions (CIN/VAIN2/3). Data of the reflex HPV DNA testing using HC2 assay and the follow up biopsies collected at the same day or after the Pap test (mean, 2.7 months) were reviewed. We compared the biopsy results with the HPV DNA testing results to determine the predictive value of HPV DNA testing for residual or recurrence of high-grade precancerous lesions in the patients with ASCUS Pap results during post-treatment follow up.

Results: Patient age ranged from 15-85 years with a mean of 45.7 years. Patient history includes cervical/vaginal squamous carcinoma (52 cases), cervical adenocarcinoma (16 cases, 12 invasive and 4 in situ), CIN/VAIN 2-3 (93 cases), and remote history of high-grade dysplasia (19 cases). The majority of patients had treatment (88.8 %). In these 180 ASC-US Pap specimens, 61 (34%) tested positive and 119 (66%) tested negative for high-risk HPV DNA. In patients with HPV negative results, 7.6% (9/119) had CIN/VAIN2/3 results. A significantly higher rate of CIN/VAIN 2/3, 28% (17/61), was observed in patients with HPV positive results (P= 0.002). The sensitivity, specificity, PPV and NPV of HC2 HPV DNA testing for predicting CIN2+ were 65%, 71%, 28% and 92%, respectively.

Conclusions:

• 1
  In patients with history of CIN2+ and subsequent ASC-US Pap results with HPV testing during post-treatment follow up, ASC-US Pap smears with positive results of HPV DNA testing using HC2 assay have a higher predictive value for residual or recurrent high-grade cervical/vaginal precancerous lesions than those with negative HC2 results.
• 2
  A relatively low sensitivity of HC2 HPV DNA testing for predicting CIN2+ was observed. The cause of false negative HC2 results (7.6%) in these patients is unknown. Further evaluation is required to determine the cause of the false negative HC2 results, such as small lesions located in the upper canal of the cervix.

33

HPV DNA Detection in ThinPrep® Pap Test Vials Is Independent of Cytologic Sampling of the Transformation Zone in HSIL Pap Tests and Subsequent Histological Results

Faye Gao, MD, Chengquan Zhao, MD

Pathology, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania

Introduction: The vast majority of HPV infection and high grade CIN lesions are identified in the transformation zone. Data were very limited with regard to the correlation between sampling of the TZ component, HSIL cytology, HPV results, and disease risk assessment.

Materials and Methods: Cases diagnosed as HSIL in ThinPrep tests (TPPT) that also had hrHPV testing results by HC2 over a 57 months span from July 2005 to April 2010 were retrieved from Magee-Womens Hospital Data base. The presence or absence of a TZ/EC and histological follow-up results were obtained.

Results: A total of 321 women with HSIL (307 cervical and 14 vaginal) TPPT also had hrHPV testing. The median age of women with cervical HSIL was 33 years (16-75) and 57 years (41-93) for vaginal HSIL. Overall positive hrHPV rate was 95.3%. There were no difference of HPV infection rates among women with and without TZ/EC and vaginal Pap (Table 1). The age-stratified hrHPV DNA prevalence analyzed in 10-year intervals showed no statistical significant differences. 250 women with cervical HSIL Pap and HPV test results had at least one follow-up cervical biopsy result. Average follow-up period was three months range from 0.2 to 40 months. 170 of 239 patients with HPV positive HSIL (71.1%) had a follow-up histologic diagnosis of ≥CIN2 including 8 squamous cell carcinoma. 4 of 11 (36.4%) patients with HPV negative HSIL had CIN 2/3. The difference of ≥CIN2 detection rates was significant (P=0.015). The presence or absence of an TZ/EC did not affect the subsequent rates of diagnoses of ≥CIN 2 lesions (Table 2). 5 women with vaginal HSIL Pap had histologic follow up. Among them one had squamous cell carcinoma, two had VAIN2 and two had VAIN1.

Conclusions:

• 1
  Cervical HSIL Paps with or without TZ/EC had the same rate of hrHPV DNA detection and the same subsequent surgical follow up results. hrHPV detection in TPPT vials is independent of cytologic sampling of the TZ/EC.
• 2
  5% cervical and 15% vaginal HSIL Paps showed negative HPV tests. These cases need further study to explore the reasons for the negative test.

Table 1. Comparison of hrHPV prevalence in women with HSIL in cervical Pap with and without TZ/ECS, and vaginal Pap

Types of Pap	Case No.	No. HPV Positive	% HPV Positive	95% CI	P value
TZ/EC Present	253	243	96.0	93.6-98.4	0.707
TZ/EC Absent	54	51	94.4	88.3-100
Vaginal	14	12	85.7	67.4-100	0.134
Total	321	306	95.3	93.0-97.6

Table 2. Histologic Findings in Women with Cervical HSIL Pap in the presence and absence of TZ/EC

PAPs	No. Case	SCC (%)	CIN3 (%)	CIN2(%)	CIN1(%)	Negative
Cervical TZ/EC+	213	6 (2.8)	59 (27.7)	81 (38.0)	47 (22.1)	20 (9.4)
Cervical TZ/EC-	37	2 (5.4)	11 (35.1)	15 (40.5)	5 (8.1)	4 (10.8)
Total	250	8 (3.2)	70 (28.0)	96 (38.4)	52 (20.4)	24 (9.4)

34

Multi-Center Validation of the Use of the Digene Hybrid Capture® 2 HPV DNA Test™ With BD SurePath™ Liquid-Based Cervical Cytology Samples

Allison Fenderson, BS, RN, BSN¹, Susan Sampson, BS¹, Betty Stephenson, BS¹, Jason Alarcon, BS¹, Barbara Preisel-Simmons, PhD¹, Dorothy Rosenthal, MD²

¹Women's Health and Cancer, BD Diagnostics, Durham, North Carolina; ²Division of Cytopathology, Johns Hopkins Medical Institutions, Baltimore, Maryland

Introduction: The Digene Hybrid Capture ® 2 (HC2) HPV DNA Test™ is currently approved by the FDA for detecting HPV DNA from cervical specimens collected and stored in Digene Specimen Transport Media™ (STM). A multi-center, in vitro diagnostic study was performed to validate an expansion of the indications for use of the Digene HC2 HPV DNA test to include the BD SurePath™ Liquid-Based Cytology (LBC) specimen as an acceptable sample type. The study was designed to establish the performance characteristics of the HC2 HPV DNA from cervical specimens stored in BD SurePath LBC preservative fluid.

Materials and Methods: Two cervical samples were collected from each consenting subject. The first sample was collected and stored in BD SurePath LBC preservative fluid. After SurePath Pap slides were prepared, the residual SurePath pellet was processed for HPV testing using the Digene HC2 assay. The second cervical specimen was collected using the Digene DNA Collection Device, stored in Digene STM, and processed for HC2 HPV testing. Among specimens with an abnormal cytology result, only those with an ASC-US diagnosis were included in the study. To provide clinical endpoints, these patients were required to undergo a colposcopy and/or biopsy procedure. Estimates were obtained for sensitivity, specificity, PPV, and NPV compared to the clinical outcome of presence or absence of cervical disease (defined as CIN2+). The HC2 HPV DNA test results from the BD SurePath pellet were compared to a composite comparator (Digene HC2 HPV DNA result from the STM vial combined with PCR/Sequencing result) to establish the ability to detect the HC2-targeted HPV types when the test is performed from the BD SurePath pellet.

Results: Results from the Digene HC2 HPV DNA test were available for 966 ASC-US subjects and adjudicated histology results were obtained for 906 of the 966. Overall sensitivity of the HC2 HPV DNA test from the BD SurePath pellet for the adjudicated diagnosis of CIN2+ was 95.8% (95% CI= 86.0, 98.8) and NPV was 99.6% (95% CI= 98.7, 99.9). Specificity was 55.7% (95% CI= 52.4, 59.0) and PPV was 10.8% (95% CI= 9.7, 11.7). When stratified by age, specificity increased from 16.2% for the youngest age group (18 to <21) to 84.4% for the subjects age 39 and older. The corresponding PPV estimates ranged from 6.1% to 13.5%. Positive, negative, and overall agreement between the HC2 HPV DNA test from the BD SurePath pellet and the composite comparator was 96% (for all 3 agreement categories) for ASC-US subjects.

Conclusions: The results of this large multi-center study confirm the ability to successfully perform the HC2 HPV DNA test from BD SurePath LBC specimens, with the clinical performance for detection of CIN2+ disease consistent with previously published studies from multiple sample types.

35

Pap Tests that are Negative on Computer-Imaged ThinPrep® Cytology but Positive for HPV on Hybrid-Capture Test: Study of 344 Cases

Rana Hoda, MD, Behzad Vakil, SCT(ASCP), Gui-Ying Ma, CT(ASCP), Edele Romain-Delva, CT(ASCP), Carrie Tolentino, CT(ASCP), Maria Erroll, CT(ASCP)

Pathology & Laboratory Medicine, New York Presbyterian Weill Cornell, New York, New York

Introduction: Data on Pap Tests that are negative on computer-imaged ThinPrep cytology but positive for HPV on Hybrid-Capture Test are limited. Herein, we study the rate of HPV-positivity (POS) by Hybrid-Capture System (Digene, Gaithersburg, MD) with NILM or NILM/BCC on ThinPrep Pap Test (TPPT).

Materials and Methods: All TPPT interpreted as “No evidence of squamous Intraepithelial Lesion or Malignancy” (NILM) or NILM with Benign Cellular Changes (NILM/BCC) and referred for HPV-testing by clinician request were studied as a Quality Assurance (QA) monitor over a 6-month-period (10/'09-3/'10). Originally, all TPPT were processed by the TP Imager System (TP-TIS) and subjected to a 22 field of view (FOV) screening by a cytotechnologist (CT). Upon prompting by morphology or clinical history, TPPT were manually screened (MS). All NILM/BCC cases were TP-TIS and MS for pathologist review. If HPV test was POS, TPPT were MS by supervisory-level CT (SVCT). Age groups of =/<29 and =/>30 were also considered.

Results: During the study period, 3,436 TPPT were interpreted as either NILM (2,835) or NILM/BCC (601), and were then tested for HPV. Age ranged from 18-85 years with =<29, 101 and =>30, 243 cases. Overall, HPV results were POS in 344/3436 (10%) cases: NILM, 253/2835 (8.9%) and NILM/BCC, 91/601 (15.1%). Upon re-screening and pathologist review, 5/344 (1.4%) cases were re-classified as abnormal: ASCUS, 4; LSIL 1. 4/5 re-classified abnormals (ASCUS: 3 and LSIL: 1) were in age =>30 and originally signed-out by CT. Of these 4 cases, 3 were routine cases screened on TP-TIS only; while 1/4 with NILM had high risk history and was screened by TP-TIS, then MS by SVCT. Only 1/5 re-classified abnormal (ASCUS) was in a =>29-year-old, originally reported as NILM/BCC by a pathologist. All 5 reclassified cases were revised and results communicated to clinician. Colposcopic and follow-up data are being collected.

Conclusions: HPV was POS in 10% (344/3436) of NILM and NILM/BCC computer-imaged TPPT. Of these cases, 1.4% (5/344) were reclassified as abnormal on QA review. QA review of HPV-POS NILM and NILM/BCC computer-imaged TPPT is of value for optimal patient management.

36

Utility of WT1 Immunoreactivity in Pelvic Washings to Determine the Origin of Serous Carcinomas of the Peritoneum and Gynecologic Tract

Ricardo Lastra, MD¹, Kathleen Montone, MD², John Brooks, MD¹, Julieta Barroeta, MD¹

¹Department of Pathology and Laboratory Medicine, Pennsylvania Hospital of the University of Pennsylvania Health System, Philadelphia, Pennsylvania; ²Department of Pathology and Laboratory Medicine, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania

Introduction: Serous carcinoma originating from the gynecologic tract and those presenting as primary peritoneal tumors have similar histologic morphologies, and determining the site of the primary lesion can be challenging. Positive WT1 staining has been reported in serous carcinomas of the ovary, but is absent in those of endometrial origin. We evaluate the usefulness of WT1 in intraoperative pelvic washings to determine the site of origin of the neoplasm, allowing for accurate classification, staging and treatment of these neoplasms.

Materials and Methods: Twenty-four cases of serous carcinoma from endometrium (n=6), ovary (n=12) and primary peritoneal (n=6) origin, with positive pelvic washings and available cell-block, received between 1991 and 2010, were retrieved from our records. WT1 staining (clone 6F-H2, Cell Marque Corp., CA; Ventana Benchmark XT System) utilizing the labeled strepavidin-biotin method was performed on sections of both the primary tumor and the cell block. Reactivity was scored based on nuclear staining in neoplastic cells as follows: 0 (<5%), 1 (6-25%), 2 (26-50%), 3 (51-75%) and 4 (>75%). The percentage of tumor cells staining, independent from the intensity, was also determined (see Figures 1, 2 and 3). Similarly, the WT1 staining pattern of the primary tumor was recorded and compared to the staining in the respective cell block.

Results: Positive WT1 staining of the tumor cells in the cell blocks was present in all ovarian (100%) and primary peritoneal cases (100%), but only in 1 endometrial case (16.7%) (p<0.0005) (see Table 1). Ten cases of ovarian (83.3%) and all cases of primary peritoneal tumors (100%) had moderate to high staining intensity. Furthermore, 8 cases of ovarian (66.7%) and 5 cases of primary peritoneal (83.3%) origin had staining in more than 50% of the tumor cells compared to the single endometrial case with WT1 staining, in which <50% of the tumor cells had staining with moderate intensity. The staining patterns in the original tumor and in the corresponding pelvic wash cell block were similar in all cases.

Conclusions: We have demonstrated that WT1 staining of cell blocks obtained from intraoperative pelvic washings is useful in determining the origin of serous carcinoma of the gynecologic tract and the peritoneum, which could serve as an aid in the accurate classification of these neoplasms and subsequent management and assignment to treatment protocols. Furthermore, this staining pattern may be useful in evaluating the origin of the neoplasm in cases not amenable to surgery, such as in positive ascites in widely metastatic tumors. Limitations raised by our study included blocks with limited cellularity and potential focal positivity for WT1 in reactive mesothelial cells. Careful histologic evaluation of the cells with positive staining is necessary to avoid this pitfall.

Figure 1. Ovarian serous carcinoma in cell block of peritoneal fluid (H&E 200x magnification).

Figure 2. Positive WT1 nuclear stain in cell block of peritoneal washings in ovarian serous carcinoma case (200x magnification).

Figure 3. Negative WT1 stain in cell block of peritoneal washings in endometrial serous carcinoma case (200x magnification).

Table 1. WT1 Immunoreactivity in Serous Carcinomas of Ovarian, Endometrial and Peritoneal Origin

	Staining of Neoplastic Cells (%)			Total
	0 (Negative)	1-2 (Weak)	3-4 (Strong)
Ovary	0 (0)	4 (33.3)	8 (66.7)	12
Peritoneal	0 (0)	1 (16.7)	5 (83.3)	6
Endometrium	5 (83.3)	1 (16.7)	0 (0)	6
Total	5	6	13	24

37

CINtec® Dual Immunostain: A New Test for Triage of Women with Abnormal Pap Smears

Sanam Loghavi, MD, Ann Walts, MD, Shikha Bose, MD

Pathology and Laboratory Medicine, Cedars Sinai Medical Center, Los Angeles, California

Introduction: Atypical squamous cells of undetermined significance (ASC) and low grade squamous intraepithelial lesion (LSIL) comprise the vast majority of all abnormal Pap smears. Currently, high-risk human papillomavirus (HR HPV) testing is utilized to triage these women for colposcopy although a high grade lesion (CIN II/III; HG CIN) is detected in only 5-20% of HR HPV+ cases. P16 and Ki-67 have each been shown to be good biomarkers for HG CIN. This study was designed to assess the role of a p16/Ki-67 dual immunostain as a marker for HG CIN and to compare this stain with HR HPV screening in ASC and LSIL Pap smears.

Materials and Methods: Eighty-seven cervical SurePath Pap smears (37 ASC and 50 LSIL) with histological and/or cytological follow up (range 1 to 83 mos; median 9 mos) were retrieved from our departmental files. The Pap stained slides were destained and then immunostained utilizing the CINtec dual staining reagent kit (CINtec® Cytology, mtm laboratories, Inc Westborough, MA) that detects over-expression of p16 and Ki-67 as brown/cytoplasmic and red/nuclear reaction products, respectively. Dual staining in one or more squamous cells was interpreted as a positive result. Staining for either p16 or Ki-67 alone was recorded as negative. 60 (35 ASC and 25 LSIL) of the 87 cases had also been tested for high risk (HR) HPV (Third Wave Technologies, Hologic Inc, Madison, WI). Results of the dual stain and HR HPV screen were correlated with follow up diagnoses. Sensitivity, specificity, and positive and negative predictive values for HG CIN were compared.

Results: Please refer to Tables 1–3.

Conclusions: CINtec dual immunostain:

• Has a higher specificity and a similar negative predictive value for detection of an underlying or subsequent HG CIN compared to the HR HPV screen.

• Would be more efficacious than HR HPV screening for triage of women with abnormal Pap smears and would decrease the need for immediate colposcopy in a large number of women.

• Is less costly, less labor intensive, and provides a shorter turn around time and potential for assessment by image analysis.

• Can be applied to routinely prepared and previously screened liquid based Pap smears without loss of cytological detail.

Prospective studies are needed to confirm these findings.

Table 1. ASC: CINtec, HR HPV test, and follow-up

	CINtec (N=37)		HR HPV (N=35)
Follow-up pathology	Negative	Positive	Negative	Positive
Negative	5/6 (83%)	1/6 (17%)	5/5 (100%)	0/5 (0%)
HPV/CIN I/LSIL	14/22 (64%)	8/22 (36%)	1/21 (5%)	20/21 (95%)
CIN II/CIN III/HSIL	1/9 (11%)	8/9 (89%)	0/9 (0%)	9/9 (100%)

Table 2. LSIL: CINtec, HR HPV test, and follow-up

	CINtec (N=50)		HR HPV (N=25)
Follow-up pathology	Negative	Positive	Negative	Positive
Negative	7/13 (54%)	6/13 (46%)	3/7 (43%)	4/7 (57%)
HPV/CIN/LSIL	22/30 (73%)	8/30 (27%)	6/15 (40%)	9/15 (60%)
CIN II/CIN III/HSIL	2/7 (29%)	5/7 (71%)	0/3 (0%)	3/3 (100%)

Table 3. Comparison of CINtec with HR HPV test for HG CIN

	ASC		LSIL
	CINtec	HR HPV	CINtec	HR HPV
Sensitivity	89%	100%	71%	100%
Specificity	68%	23%	67%	41%
Positive predictive value	47%	31%	26%	19%
Negative predictive value	95%	100%	94%	100%

38

Spectral Cytopathology of Cervical Cells Infected with Human Papillomavirus

Jennifer Schubert, BA¹, Kristi Bedrossian, BS, CT², Nora Laver, MD², Benjamin Bird, PhD¹, Miloš Miljković, PhD¹, Max Diem, PhD¹

¹Chemistry and Chemical Biology, Northeastern University, Boston, Massachusetts; ²Cytopathology Division, Department of Pathology, Tufts Medical Center, Boston, Massachusetts

Introduction: Viruses are carcinogenic agents in humans, and have been linked to approximately 15-20% of cancers worldwide. Over 100 genotypes of the human papillomavirus (HPV) have been identified, and specific strains have been implicated as important risk factors in cancers of the anogenital tract and some skin cancers. Approximately 97% of cervical cancers contain oncogenic high-risk strains of HPV, i.e. types 16 and 18 that cause 80% of cervical cancers. Clinical studies have shown that persistent infection with high-risk strains of HPV is more likely to lead to cervical cancer, predominately in women over 30 years old.

Materials and Methods:Figure 1 - Spectral Cytopathology (SCP), a combined technique employing infrared micro-spectral imaging for data collection and analysis by multivariate statistics (Principal Component Analysis), was used to investigate single cells infected with HPV in cervical specimens. This study includes several dozen clinical samples that were diagnosed by traditional cytological protocol, assayed for high-risk strains of HPV DNA (Digene Hybrid Capture II), and analyzed by SCP.

Results: The samples were successfully differentiated by SCP based on their HPV status, positive versus negative, regardless of their cytological diagnosis. We also observed a small number of HPV negative clinical samples that co-clustered with the HPV positive samples (false positive). However, patient history for these samples revealed a history of cervical dysplasia that we believe could be attributed to a latent viral effect. Further investigations are required to verify our assumptions, but we presently postulate the spectral changes observed are due to the viral proteome or an immune response opposed to the actual viral DNA, since the dominant spectral changes are found in the protein region of the spectrum.

Conclusions: This study manifests SCP's potential as an enhancing screening tool for cytopathologists by aiding in the accurate and reproducible diagnosis of cells and identifying earlier stages of disease.

39

Dual Localization of p16INK4a and Ki-67 Improves Specificity for HSIL and AGUS

Meenakshi Singh, MD¹, Stephanie Burke, BS², Ali Akalin, MD, PhD³, Kenneth Shroyer, MD, PhD²

¹Pathology, State University of New York at Stony Brook; Stony Brook University Medical Center, Stony Brook, New York; ²State University of New York at Stony Brook; Stony Brook University Medical Center, Stony Brook, New York; ³Pathology, University of Massachusetts, Worcester, Massachusetts

Introduction: Previous studies have shown that the co-expression of p16INK4a and Ki-67 is sensitive and specific for high grade lesions (≥ CIN2) on cervical biopsy. Limited information is available, however, regarding the relative clinical performance of a dual staining approach for p16INK4a plus Ki-67 as a diagnostic adjunct in cervical cytology. This study was designed to evaluate the performance of a dual p16INK4a and Ki-67 ICC assay to determine test performance including sensitivity and specificity of dual staining versus detection of p16INK4a alone in a range of cytology specimens from negative to HSIL and AGUS.

Materials and Methods: The study included 87 cervical cytology specimens, including 21 NILM (negative), 19 ASCUS, 11 LSIL, 8 ACH, 6 AGC, and 22 HSIL cases. The Cytology diagnosis was rendered on Thin Prep smears. Multiplexed immunocytochemistry (ICC assay) was performed using the CINtec(R) Plus Kit (mtm laboratories, Westborough, MA) according to the manufacturer's guidelines. A Thin Prep 2000 was used to prepare the slides. Evaluation was performed by two individuals who were blinded to the cytology diagnosis. Negative controls included substitution of the primary antibody with subclass-matched monoclonal immunoglobulins and HeLa cell spiked cytology specimens served as positive controls. Positive cytology test results were defined as staining in epithelial cells with p16INK4a as brown (DAB chromogen), Ki-67 as bright red (alkaline phosphatase chromogen) and maroon for dual staining, without regard to cellular morphology.

Results: At least scattered p16INK4a immunoreactive cells were detected in a high proportion of cases from all diagnostic groups. The co-localization of p16INK4a plus Ki-67, however, was frequently detected in HSIL and AGC but was rarely detected in NILM cases (Table 1). When cases were scored positive based on the detection of any cells with p16INK4a, the assay showed a sensitivity of 91.7% and a specificity of 39.2% for the detection of HSIL versus combined NILM/ASCUS/LSIL categories. By contrast, the co-localization of p16INK4a plus Ki-67 improved specificity for HSIL to 79.7 % (p <0.0001) but failed to detect one HSIL case that was scored positive for p16INK4a but not Ki-67.

Conclusions: The p16INK4a ICC testing considered positive based on immunoreactivity of cells without regard to cytology diagnosis had a high sensitivity but a low specificity in separating high-grade lesions from ASCUS/LSIL/negative smears. The dual staining dramatically increased the specificity for diagnosis of high-grade lesions with minimal impact on sensitivity.

Table 1. 

Diagnosis	Number of cases	p16 positive	p16/Ki-67 positive
NILM	21	14	3
ASCUS	19	10	5
ASCH	8	7	4
LSIL	11	7	5
HSIL	21	20	19
AGC	6	9	4

40

The Utility of BioView Duet™- Aided Nuclear Scoring to Evaluate p16INK4a Protein Over-expression Using Direct-from-Vial Pap Test Cytology Specimens

G. Denice Smith, PhD¹, Elke Jarboe, MD², Sheryl Tripp, MT(ASCP)¹, Sherrie Perkins, MD², Brian Collins, MD²

¹Institute for Clinical and Experimental Pathology®, ARUP Laboratories, Salt Lake City, Utah; ²Department of Pathology, University of Utah, Salt Lake City, Utah

Introduction: Although p16INK4a protein over-expression is of value for identifying patients with HSIL+ lesions in biopsy specimens, its use for direct-from-vial Pap test cytology specimens can be more problematic because p16INK4a expression is sometimes observed in normal endocervical, metaplastic, or atrophic cells. Our goal was to develop a consistent method to score p16INK4a-expressing cells using BioView Duet-generated data following an automated scan and segregation of p16INK4a-expressing from non-expressing cells.

Materials and Methods: Residual cervical ThinPrep or SurePath Pap test cytology samples were used. Slides were prepared and immunolabeled using a p16INK4a-specific antibody from MTM Laboratories with 3,3'-diaminobenzidine substrate and hematoxylin counterstain. The slides were then scanned automatically using the Duet instrument, followed by cytotechnologist evaluation and re-classification of target cell classes. Data for target cells in the abnormal category were then collected in a spreadsheet, including nuclear and cytoplasmic areas, as well as nuclear:cytoplasmic (N:C) ratios. A modification of a nuclear scoring system developed by Wentzensen, et al. (Cancer Cytopathology 2005; 105:461-467), was then used by pathologists to score select target cells in the abnormal category. The individual slide score corresponded to the highest score given to individual target cells.

Results: Depending on the area of interest, automated scans required 1-2+ hours, with target cells captured and sorted into one of four categories. Because the brightfield software application used was not designed to segregate p16INK4a-expressing from non-expressing cells, considerable post-scan target cell reclassification was required. The software attempts to determine nuclear and cytoplasmic areas, as well as N:C ratios for target cells, and a ruler tool is available to determine diameters. This works best for single cells with clearly defined membranes with no cell overlap. An imaging study (Maeda, et al., Acta Cytologica 1997; 41:744-748) described an average N:C ratio of 0.225 for CIN2 and 0.385 for CIN3, which we used as a guide for nuclear scoring.

Eighteen specimens (14 SurePath; 4 ThinPrep) that had been interpreted as HSIL by cytology were used for p16INK4a-labeling to determine if the Duet imaging system could be used to scan the slides by brightfield microscopy, and to aid in the evaluation of p16 INK4a expression. Biopsy confirmation of HSIL was available for 10 samples. In 7 cases the positive expression of p16INK4a correlated with the cytology and biopsy HSIL interpretations. For 1 case that was scored as negative for p16INK4a expression, HSIL was biopsy confirmed. For 2 cases, cell overlap and p16INK4a localization in both the nucleus and cytoplasm, made it difficult to determine if the p16INK4a expressing cells were abnormal, as would be the case for a manual screen as well.

Conclusions: The BioView Duet automated scanned images are of generally high quality, and the ability to relocate target cells interactively allows for live review of cells of concern. Refinements of the software and additional experience would be required to utilize it in the clinical laboratory. The use of the Duet imaging system to aid in the interpretation of p16INK4a expression in cytology samples demonstrated diagnostic and interpretative potential.

41

A Pilot Study of Rapid Prescreening in Liquid-Based Papanicolaou Tests

Debora Smith, CT(ASCP), Todd Fairley, CT(ASCP), Michael Thrall, MD

Department of Pathology, The Methodist Hospital, Houston, Texas

Introduction: Rapid prescreening (RPS) is an alternative method of quality assurance based on rapid (<1 minute) interpretations of slides by cytotechnologists (CT) prior to full screening. To our knowledge the utility of this method in laboratories using liquid-based Papanicolaou tests (LBPT) has not been previously evaluated. Intuitively, LBPT should make rapid prescreening easier and, perhaps, more effective because of the smaller screening area and cleaner background.

Materials and Methods: A senior CT and a junior CT performed RPS on 439 and 389 cases, respectively, without the knowledge of the CT who performed full screening. ThinPrep (35%) and SurePath (65%) preparations were used. Each RPS was timed so as to not go over 45 seconds for SurePath or 60 seconds for ThinPrep. A RPS diagnosis was given within the categories of: Negative for Intraepithelial Lesion or Malignancy (NILM), Atypical Squamous Cells of Undetermined Significance (ASCUS), Low grade Squamous Intraepithelial Lesion (LSIL), High grade Squamous Intraepithelial Lesion (HSIL), or Atypical Glandular Cells (AGC) and higher. The results were compared with the final interpretation rendered by the screening CT, or the pathologist for the cases sent for review. ThinPrep slides were screened using the ThinPrep Imaging System.

Results: A comparison of the RPS interpretation with the final interpretation for the senior CT is presented in Table 1, for the junior CT in table 2, and for both in Table 3. The RPS interpretation correlated exactly with the final interpretation for 89.1% of slides. Using the final interpretation as the gold standard, the sensitivity of RPS was 66.7% and the specificity was 94.8%. Most disagreements between RPS and the final interpretation were minor, with only 3.9% of cases having a two-step discrepancy.

Conclusions: In this pilot study of RPS in LBPT, we have found results comparable to published studies performed on conventional smears. RPS flagged all HSIL cases in this study as abnormal. RPS found few cases where re-review before sign-out might lead to significant changes in patient follow-up.

Table 1. RPS Interpretation vs. Final Interpretation for Senior CT

	RPS
Final		NILM	ASC-US	LSIL	HSIL	AGC+
Bold = exact agreement, Underlined = two-step discrepancy.
	NILM	354	9	13	0	0
	ASC-US	9	3	5	1	0
	LSIL	4	0	25	1	0
	HSIL	0	2	5	3	0
	AGC+	2	1	1	0	1

Table 2. RPS Interpretation vs. Final Interpretation for Junior CT

	RPS
Final		NILM	ASC-US	LSIL	HSIL	AGC+
Bold = exact agreement, Underlined = two-step discrepancy.
	NILM	327	13	5	0	0
	ASC-US	7	6	5	0	0
	LSIL	2	1	18	1	0
	HSIL	0	0	0	1	0
	AGC+	3	0	0	0	0

Table 3. RPS Interpretation vs. Final Interpretation - Total

	RPS
Final		NILM	ASC-US	LSIL	HSIL	AGC
Bold = exact agreement, Underlined = two-step discrepancy.
	NILM	681	22	18	0	0
	ASC-US	16	9	10	1	0
	LSIL	6	1	43	2	0
	HSIL	0	2	5	4	0
	AGC	5	1	1	0	1

42

Combined Staining of Cervical Lesions for p16INK4a and Ki67 Distinguishes Cervical Dysplasia from Metaplasia

Miriam Reuschenbach, MD¹, Mirjam Seiz¹, Svetlana Vinokurova, PhD¹, Christina von Knebel Doeberitz, MD², Dietmar Schmidt, MD³, Magnus von Knebel Doeberitz, MD¹

¹Department of Applied Tumor Biology, Institute of Pathology, University of Heidelberg, Heidelberg, Germany; ²Cion Consulting GmbH, Heidelberg, Germany; ³Institute of Pathology Mannheim, Mannheim, Germany

Introduction: The immunohistochemical detection of p16INK4a has substantially improved the accuracy of histopathological diagnoses of cervical lesions. p16INK4a expression in non-transformed cells is occasionally found, most likely as a result of alterations of normal squamous epithelial differentiation and/or cellular senescence, in particular in metaplastic and atrophic lesions. In the present study we characterized the proliferative potential of p16INK4a positive cells in metaplastic lesions.

Materials and Methods: Formalin-fixed, paraffin-embedded tissues of each 22 CIN1 and CIN2 lesions, 25 CIN3/CIS, 46 carcinomas and 26 metaplasias were stained for the expression of p16INK4a and Ki67 (CINtec Plus, MTM Laboratories, Heidelberg).

Results: A diffuse expression of p16INK4a was found in 12/22 (54.5%) CIN1, 17/22 (77.3%) CIN2, 25/25 (100%) CIN3/CIS and 45/45 (100%) carcinomas. p16INK4a expression was found in all metaplasias in an inhomogeneous or focal pattern. The extent of Ki67 expression increased with the severity of the lesion. Double-positive cells were found in 11/22 (50.0%) CIN1, 17/22 (87.3%) CIN2, 25/25 (100%) CIN3/CIS and 46/46 (100%) carcinomas. In metaplasias only in 3/26 (11.5%) specimens single (1 to 5) double-positive cells were observed.

Conclusions: p16INK4a/Ki67 co-expression appears to be highly specific for HPV-transformed cells in cervical dysplasia while p16INK4a expressing cells in metaplastic lesions do not or only very rarely co-express Ki67. These data are particularly relevant for the application of p16/Ki67 double-staining methods in cytology specimens. Our data suggest that the combined application of antibodies against Ki67 and p16INK4a allows distinguishing HPV-transformed cells from cells that display p16INK4a overexpression for other reasons.

43

Differentiation Dependent Methylation of the Human Papilloma Virus Type (HPV) 16 Upstream Regulatory Region (URR) in Latent, Permissive, and Transforming Infection Modes

Magnus von Knebel Doeberitz, MD, Svetlana Vinokurova, PhD

Department of Applied Tumor Biology, Institute of Pathology, University of Heidelberg, Heidelberg, Germany

Introduction: High risk human papilloma viruses (HR-HPVs) infect squamous epithelial cells. They may cause cervical lesions that eventually undergo neoplastic transformation. Replication of HPVs is restricted to terminally differentiated squamous cells of the more superficial cell layers. The primary host cells in the basal cell layer, however, display a highly restricted viral gene expression pattern. Here we show that functionally relevant shifts of the methylation profile of distinct CpG islands in transcription factor binding sites in the viral upstream regulatory region (URR) are linked to epithelial differentiation. Moreover, neoplastic transformation is associated with the de novo methylation of distinct CpG dinucleotides within the E2 binding site 1 that results in strong transcriptional activation of the viral E6 and E7 genes.

Materials and Methods: Selected biopsies encompassing HPV 16 infected normal tissues adjacent to CIN1 lesions that were located in direct anatomical association to high grade (CIN3) lesions were micro-dissected: DNA was isolated from basal, intermediate and superficial squamous cells and the methylation pattern of the URR of HPV 16 was determined. Promoter activities were monitored using luciferase expression plasmids and purified HPV 16 DNA with or without methylated CpG dinucleotides and quantitative RT-PCR experiments.

Results: In normal squamous epithelium fully methylated HPV genomes were detected in all layers of the squamous epithelium. In CIN1 lesions representing the productive phase of the HPV life cycle in the squamous epithelium, a consistent but complex pattern of differential methylation was found in critical CpG dinucleotides suggesting that the differentiation dependent activity of HPV 16 gene expression is linked to differential methylation of distinct CpG-islands of the viral promoter and enhancer element. In transforming HPV-infections (CIN3 lesions) distinct shifts of the methylation pattern of the viral URR were linked to substantial over-expression of the viral oncogenes. All shifts of the methylation pattern of the URR were associated with substantial shifts of the activity of the URR as measured by luciferase or qRT-PCR assays.

Conclusions: These data strongly suggest that the HPV life cycle is regulated by differential methylation of the URR of the HPV genome. There are three distinct modes of HPV infections: i.) latent infections during which the viral genome is completely methylated blocking viral gene expression and thus preventing any pathological effects; ii) the permissive or productive phase during which differential methylation patterns cause the differentiation dependent viral gene expression profile in squamous epithelial cells; and iii.) the transforming mode of viral gene expression during which a distinct shift of the HPV 16 URR methylation pattern triggers over-expression of the viral oncogenes E6 and E7 that subsequently cause neoplastic transformation of the infected epithelium. These data suggest that epigenetic modification of the HPV genome is a major determinant of their biology and associated pathology.

44

BD Surepath Plus Test, a Combination of Immunocytochemistry for MCM2 and MCM7 with a Pap Counterstain, Correlates Biomarker Over-Expression with CIN2+ Disease

Clark Whitehead, PhD, Adriann Taylor, BS, Qin He, PhD, Lisa Allen, BS, Susan Sampson, PhD, Ramona Nelson, BS, Marquetta Gore, BS, Roxanne Hudson, BS, Nikki Prpic, HT (ASCP), Susan Wallace, BS, Hema Doobay, MS, Douglas Malinowski, PhD, Timothy Fischer, BS

Women's Health and Cancer, BD Diagnostics, Durham, North Carolina

Introduction: Overexpression of minichromosome maintenance proteins 2 and 7 have been shown to correlate with aberrant S-phase induction and persistent HPV infection. To further improve the accuracy of identifying abnormal cells and true disease from BD SurePath liquid-based cervical (LBC) cytology specimens, our goal was to develop an automated test composed of reagents and staining platform that combines the morphology of a standard SurePath Pap with protein biomarker immunostaining on a single slide (hereafter referred to as SurePath Plus). This study evaluated the performance of the new BD SurePath Plus test in identifying CIN2+ disease within a biopsy confirmed research cohort.

Materials and Methods: This study included 133 cytology specimens ranging from NILM to HSIL with follow-up biopsy results. For each sample two slides were produced, one slide prepared as a SurePath Pap and a second SurePath Plus slide prepared using the BD PrepStain Plus instrument that combines cell deposition with optimized immunocytochemical processing and Pap counterstaining. Both slides were scored using standard Bethesda 2001 criteria. The SurePath Plus slide was further evaluated for the presence of nuclear immunostaining in morphologically abnormal cells. The distribution of cases within the various morphologic categories and their biopsy status were compared.

Results: Using a cytology endpoint there was 45% more high grade disease identified for the SurePath Plus slides compared to the standard SurePath Pap slides. More specifically within this research cohort, there was a 45% increase in the HSIL+ detection rate (18 cases) and a corresponding 21% decrease in the LSIL (10 cases) and a 30% decrease in the ASCUS (8 cases) detection rates. Looking at biopsy endpoints within the same cohort there was a significant increase in the number of CIN2+ cases associated with HSIL+ cytology and a corresponding reduction in the amount of CIN2+ within the LSIL and ASCUS groups for the SurePath Plus test when compared to the SurePath test. Specifically the number of CIN2+ cases within the HSIL+ group increased 53% (16 cases) while the number of CIN2+ cases decreased by 46% (10 cases within the LSIL and 86% (6 cases) for the ASCUS population.

Conclusions: We have successfully developed reagents, assay and instrumentation that combine biomarker specific immunostaining with standard Pap counterstaining. The use of the SurePath Plus test leverages the advantages of both biomarker expression and morphologic assessment on a single LBC slide. Within this biopsy confirmed research cohort, the SurePath Plus test resulted in a more accurate detection of high grade disease. Potentially this improvement over the standard LBC tests will result in an increase in disease detection rates coupled with a corresponding relative reduction in unnecessary follow-up procedures. Additional prospective studies are warranted to further understand the utility and performance of this test.

45

Evaluation of a Real-Time Computer-Assisted Imaging System for Cervical Cytology in China

Bin Yang, MD, PhD.¹, Jun Wang, MD², Chun Wang, MD², Hui Du, MD², Susane Belinson, PhD³, Ruifang Wu, MD², Jerome Belinson, MD¹

¹Cleveland Clinic, Cleveland, Ohio; ²Peking University Shenzhen Hospital, Shenzhen, Guangdong Province, China; ³Northwestern University, Chicago, Illinois

Introduction: Success of cervical cytology in western countries relies largely on well-trained cytotechnologists who are rare or non-existent in vast developing countries. As most current computer-assisted cytology imager systems still require cytotechnologists' review and input after slides scanned, a real-time imager with integration of scanning, screening and diagnosing system is much needed. We have evaluated the performance of a real-time computer-assisted imager system in cervical cytology in a large cohort of community-based population in rural China.

Materials and Methods: A total of 10,000 non-screened Chinese women, between 25 and 59 years and from three counties in Guandong province, China, was recruited. A real-time Thin-Prep I2 imager system (Hologic Inc, USA) was installed and validated on site in China. After a 90 second whole-slide scanning on a microscope, 22 areas of likely abnormality were automatically selected and reviewed in sequence. Cytologic diagnosis was rendered by two cytopathologists. No cytotechnologist was involved in the project. HPV status was tested by both HCII (Qiagen, Germany) and Cervista (Hologic, US). Patients with abnormal cytology (>=ASC) or HPV positivity were recalled, colposcopied and biopsied with POI's micro-biopsy protocol. Cytologic diagnosis was correlated with histologic follow up and HPV test results.

Results: On average it took about 3 minutes for a cytopathologist to review a slide with automatically selected 22 areas using I2 imager. We have finished reviewing cytology of all 10,000 cases. Intermittent data analysis was done on 5190 cases with histologic follow up so far. Among them, 4626 (89.1%) were cytologically normal and 564 (10.9%) abnormal. Abnormal cytology includes 328 (6.3%) ASC, 92 (1.8%) LSIL, 45 (0.87%) ASC-H, and 95 (1.8%) HSIL. The sensitivity, specificity, positive and negative predictive values for cytologic detection of ≥CIN2 lesions was 89.0%, 91.5%, 24.1%, 99.6% respectively. In correlation with HPV tests, HPV positivity was 51.8% for ASC, 80% for ASC-H, 79.4% for LSIL, and 94.7% for HSIL. Approximately 5% of HPV-negative CIN2+ lesions were detected by the imager-assisted cytology. No single HSIL case was missed analytically by the imager with its selected 22 fields.

Conclusions: Our study demonstrates that ThinPrep I2 imager is an efficient and relatively accurate cytology screening system. This imager-assisted cytology has a high sensitivity (89%) and specificity (92%) in identifying CIN2+ cells with acceptable ASC and HPV positive rate. Our data indicate that the I2 imager can be a useful tool in assisting cervical cytology screening in low-resource areas lacking skilled cytotechnologists.

46

Web-Based Decision Support System for Guidance of HPV DNA Test Utilization

Xiaoyin Jiang, MD, Kenneth Youens, MD

Department of Pathology, Duke University Medical Center, Durham, North Carolina

Introduction: High-risk human papillomavirus (HPV) now plays an indispensible role in cervical cancer screening. However, given the relatively recent widespread acceptance of the technique, some confusion exists among clinicians and laboratorians alike about the correct indications for HPV DNA testing. To remedy this difficulty, a comprehensive, evidence-based set of guidelines for human papillomavirus (HPV) DNA test utilization was recently developed and published by the Cytology Education and Technology Consortium (CETC). The guidelines provide a convenient summary of the clinical indications and contraindications for HPV DNA testing.

Materials and Methods: Using ASP.NET and Microsoft SQL server, we created a web-based decision support system to guide HPV DNA test ordering and utilization based on the CETC guidelines. The site presents a web form that displays a set of drop-down menus into which patient characteristics and prior testing results are entered. The entered data is compared against a comprehensive database of patient characteristics and test indications created by pathologists familiar with the CETC testing guidelines. For each given set of characteristics the site returns a result indicating 1) whether HPV testing is indicated, 2) whether testing should be performed as reflex or standalone, and 3) the most relevant guideline(s) informing the result.

Results: A working prototype of the tool is available at http://www.thetrainedeye.org. The source code is freely available. A function to print customized, pre-completed requisition forms based on each query is currently under development, and a study to track the system's utilization and efficacy is underway.

Conclusions: The complex evidence upon which the HPV DNA testing guidelines are based makes it difficult for the busy laboratorian or clinician to commit them to memory for easy application. Web availability solves the problem of deployment to our institution's various geographically dispersed clinical and laboratory sites. Our tool provides readily available decision support to clinicians as they decide whether to order HPV testing, and to laboratorians as they review the indications for requested tests.

Head/Neck

47

Thyroid Cyst Aspirates: An Analysis of TBS Adequacy Guidelines

Stephen Baker, MD, Debbie Sabo, CT(ASCP), Maureen Croyle, CT(ASCP), David Deeds, CT(ASCP), Beverly Cash, CT(ASCP), Jennifer Brainard, MD, Charles Biscotti, MD

Anatomic Pathology, Cleveland Clinic, Cleveland, Ohio

Introduction: The 2010 Bethesda System for Reporting Thyroid Cytopathology (TBS) considers thyroid cyst fluid nondiagnostic/unsatisfactory if fewer than 6 groups of 10 benign follicular cells are present. This departs from the usual practice of many labs, including our own, that routinely treat these specimens as satisfactory and sign them out as “negative for malignancy, cyst contents.” We analyzed a series of thyroid cyst aspirates to evaluate TBS adequacy guidelines.

Materials and Methods: An electronic search of the pathology database identified all benign cystic thyroid fine needle aspirate (FNA) cases from 2007. A cytotechnologist (DS, MC, DD, BC) rescreened the slides. Follicular cell clusters were counted and the presence of hemosiderin, atypia, and oxyphils was recorded. Each specimen was classified as adequate or non-diagnostic using TBS guidelines. The electronic medical record and the pathology database provided follow-up.

Results: The study group included 93 specimens from 88 patients. Adequacy evaluation yielded 58 (62%) adequate and 35 (38%) non-diagnostic. Table 1 summarizes the cellular features. Notably, we interpreted atypical cells in a single specimen on rescreen. This patient has negative clinical follow-up at 24 months. Clinical follow-up was available for 29 of 30 (97%) patients from the non-diagnostic group and 52 of 58 (90%) patients from the adequate group. The follow-up periods ranged from 1 to 39 months (average 27 months) for the non-diagnostic group and 2 to 37 months (average 28 months) for the adequate group. Only one significant lesion occurred during follow-up. An anaplastic thyroid carcinoma was diagnosed 31 months after cyst aspirate with adequate cellularity. Another patient had an incidental papillary microcarcinoma discovered in a thyroidectomy specimen 10 months after a cyst aspirate that contained no follicular epithelium.

Conclusions: Our experience does not support the current TBS guidelines for reporting cyst aspirates with <6 follicular groups as non-diagnostic, a term that conveys an inadequate or insufficient sample. Rather, our data suggest that cyst fluid samples can be reported as benign. The only significant lesion occurred in one of the patients with an “adequate, benign” cyst aspirate. Since false negatives occur, a description of cellularity seems prudent with the understanding that correlation with clinical/sonographic features is implicit.

Table 1. Cellular Features by Adequacy

Features	Adequate (N=58)	Non-Diagnostic
		1-5 Clusters (N=20)	0 Clusters (N=15)
Hemosiderin	54 (93%)	18 (90%)	13 (87%)
Oxyphils	42 (72%)	7 (35%)	N/A
Atypia	1 (2%)	0	N/A

48

Cytopathologic Analysis of Salivary Gland Neoplasms with Epithelial/Myoepithelial Cells but no Stroma on Fine Needle Aspiration

Nancy Ray, BS, Aaron Hoschar, MD, Longwen Chen, MD

Anatomic Pathology, The Cleveland Clinic, Cleveland, Ohio

Introduction: Fine-needle aspiration cytology (FNAC) is a widely used, safe procedure that can quickly provide useful information for clinical management of salivary gland neoplasm. A great deal of salivary gland aspirates, particularly of the parotid glands, will turn out to be pleomorphic adenomas (PA). The diagnosis of PA on FNAC is usually obvious after identifying the following components: extra-cellular chondromyxoid stroma, myoepithelial cells, and ductal (epithelial) cells. However, stroma-deficient and/or cellular cases may be confused with other benign and malignant salivary gland neoplasms. Published cytology studies addressing the stroma-deficient/cellular salivary gland aspirates are very limited. The aim of this study was to study the cytomorphology of this cohort of cases and evaluate the clinical follow-up.

Materials and Methods: Twenty-eight cases of salivary gland aspirates with epithelial and myoepithelial cells, but no stroma were retrospectively reviewed. Follow-up histology or cytology on these cases was also studied.

Results: Twenty-three of the 28 cases were from the parotid gland and 5 were from the submandibular gland. The diagnosis of this group of aspirates was “atypical cells, favor/suspicious for a low-grade salivary gland epithelial/myoepithelial neoplasm”. The cytologic smears of these cases were mostly cellular and comprised of predominantly bland epithelial and myoepithelial appearing cells without typical chondromyxoid stroma seen in PA. Occasional basaloid appearing cells were also seen. Eighteen of the 28 cases had clinical follow-up, 17 patients had histology and 1 patient had cytology follow-up. Ten of the 18 cases were PA with 7 of them cellular PA. Two cases were basal cell adenoma, 1 case was myoepithelioma, and 1 case was benign adenoma, NOS. Two cases showed epithelial/myoepithelial carcinoma on surgical resection, 1 case was basal cell adenocarcinoma, and 1 case was adenoid cystic carcinoma.

Conclusions: Stroma-deficient cellular salivary gland aspirates imposed diagnostic difficulties. Although 14 of the 18 (77.8%) cases showed benign neoplasms with more than 70% cases of pleomorphic adenomas, a quite significant amount of cases (22.2%) were malignant salivary gland neoplasms. The diagnostic category of “atypical” and appropriate clinical surgical follow-up are recommended for this cohort of FNAC.

49

Cytopathologic Interpretative Variation of Nodular Goiter of the Thyroid on Fine Needle Aspiration Cytology: Synopsis of 1050 Responses from the ASCP NonGyn Assessment Program

Stan Eilers, MD¹, Michael Facik, MPA, CT(ASCP)², Perkins Mukunyadzi, MD³, Paul Wakely, Jr., MD⁴, Amy Wendel Spiczka, MS, SCT, MP, HTL(ASCP)CM⁵, Paula LaPolice, CT(ASCP)⁶, Ajay Shah, MD⁷, Husain Saleh, MD, MBA⁸, Umesh Kapur, MD⁹, Bryan Hunt, MD¹⁰, Jennifer Clark, SCT(ASCP)¹¹, Larry Lemon, CT(ASCP)¹²

¹Pathology, Mercy Medical Center, Cedar Rapids, Iowa; ²Pathology, University of Rochester Medical Center, Rochester, New York; ³Pathology, Arkansas Pathology Associates, Little Rock, Arkansas; ⁴Pathology, The Ohio State University, Columbus, Ohio; ⁵Pathology, Mayo Clinic, Scottsdale, Arizona; ⁶Pathology, Baystate Medical Center, Springfield, Massachusetts; ⁷Pathology, FNA Clinic, Toledo, Ohio; ⁸Pathology, Detroit Medical Center, Detroit, Michigan; ⁹Pathology, Loyola University Medical Center, Maywood, Illinois; ¹⁰Pathology, Medical College of Wisconsin, Milwaukee, Wisconsin; ¹¹Continuing Professional Development, American Society for Clinical Pathology, Indianapolis, Indiana; ¹²Continuing Professional Development, American Society for Clinical Pathology, Indianapolis, Indiana

Introduction: Non-neoplastic nodular goiter is a common disease seen in Thyroid cytopathology samples, and therefore Thyroid FNA specimens of this entity were incorporated into the American Society for Clinical Pathology (ASCP) NonGYN Assessment Program.

Materials and Methods: The ASCP NonGYN Assessment Program is a glass-slide program developed with oversight from the ASCP NonGYN Assessment Committee. Each annual program is composed of FNA and NonGYN samples for a total of 20 patient cases, divided into 4 quarterly shipments of five cases. Laboratory and individual Peer-Comparison statistics for each case reviewed are provided to all participants post-event participation, including those of cases from Thyroid FNA non-neoplastic goiter.

Results:Table 1 - Performance data from 1050 total responses from five cases of Thyroid FNA non-neoplastic nodular goiter circulating in the NonGYN Assessment program since 2007 were extracted and reviewed. 78.5% of participants chose the correct response of non-neoplastic nodular goiter; 6.4% classified the cases as follicular neoplasm; 5.4% Hashimoto's thyroiditis; 3.3% parathyroid cyst; 1.5% Hurthle cell neoplasm; 1.1% medullary carcinoma; and 1.0% papillary carcinoma. 2.8% of participants did not provide an interpretation response (Table 2).

Conclusions: Non-neoplastic goiters account for the majority of thyroid aspirates. Accurate recognition of this common disorder is important to avoid unnecessary surgical intervention. Although the overall performance is good, a significant number of these aspirates are misinterpreted. 10% of the aspirates were interpreted as neoplasms and/or malignancies which would result in surgical intervention. Follicular neoplasm is the most common differential diagnosis with non-neoplastic goiter. Aspirates from goiters may be cellular; however, the follicular cells tend to be arranged in sheets and macrofollicles as opposed to microfollicles. Hypercellularity with large sheets and single cells may account for the other neoplastic interpretations. However, the lack of neuroendocrine features in the case of medullary carcinoma, the lack of monotonous Hurthle cells and the lack of papillary nuclear features should rule these entities out.

8.7% of the aspirates were interpreted as either Hashimoto's thyroiditis or parathyroid cyst. Although these interpretations may or may not lead to surgery, the clinical management of the patient would be impacted. Stripped nuclei from follicular cells may resemble lymphocytes, however the presence of abundant colloid and lack of Hurthle cells make Hashimoto's thyroiditis less likely. Differentiating parathyroid from thyroid on aspirates can be difficult. Parathyroid aspirates typically have a dimorphic pattern with stripped nuclei and intact cells resembling follicular cells. Colloid may be mistaken for cystic debris.

The 2010 release of The Bethesda System for Reporting Thyroid Cytopathology, (Syed Z. Ali MD and Edmund S. Cibas MD, (Eds.), sets forth a uniform system for reporting the results of thyroid fine needle aspiration (FNA). Noted is the fact that Thyroid FNA is one of the most commonly performed cytologic procedures and is the standard diagnostic method for managing the patient with a thyroid nodule, supporting the need for enhanced cytopathology education in this area.

Table 1. Response Categories - Thyroid FNA Non-neoplastic Goiter

General Response Category	MD	%	CT	%	Not Indicated	%	Total	%
Negative/Reactive/  Hyperplasia/  Developmental	394	37.5%	296	28.2%	61	5.8%	751	71.5%
Infectious/Inflammatory  Process	20	1.9%	36	3.4%	5	0.5%	61	5.8%
Benign Neoplasm	32	3.0%	108	10.3%	13	1.2%	153	14.6%
Lesion of Uncertain  Biologic Potential	28	2.7%	25	2.4%	6	0.6%	59	5.6%
Positive for Malignancy	15	1.4%	9	0.9%	2	0.2%	26	2.5%
Totals	489	46.6%	474	45.1%	87	8.3%	1050	100.0%

Table 2. Interpretation Response - Thyroid FNA Non-neoplastic Goiter

ReferenceInterpretation	MD	%	CT	%	Not Indicated	%	Total	%
Non-neoplastic goiter	407	38.8%	346	33.0%	71	6.8%	824	78.5%
Follicular neoplasm	27	2.6%	34	3.2%	6	0.6%	67	6.4%
Hashimoto's thyroiditis	22	2.1%	27	2.6%	7	0.7%	56	5.4%
Parathyroid cyst	4	0.4%	31	3.0%	0	0.0%	35	3.3%
Hurthle cell neoplasm	8	0.8%	8	0.8%	1	0.1%	17	1.5%
Medullary carcinoma	10	1.0%	1	0.1%	1	0.1%	12	1.1%
Papillary carcinoma	4	0.4%	5	0.5%	1	0.1%	10	1.0%
No interpretation  selected	7	0.7%	22	2.1%	0	0.0%	29	2.8%
Totals	489	46.6%	474	45.1%	87	8.3%	1050	100.0%

50

The Bethesda System for Reporting Thyroid Cytopathology: A Community Hospital Experience

Mohiedean Ghofrani, MD, MBA, Keri Epperson, CT(ASCP), Elizabeth Gillihan, CT(ASCP), Kathleen Reising, CT(ASCP), Kevin Young, CT(ASCP)

Department of Pathology, Southwest Washington Medical Center, Vancouver, Washington

Introduction: The Bethesda System (TBS) for Reporting Thyroid Cytopathology has been recently proposed to potentially improve communication, enhance cytologic-histologic correlation, facilitate research, and enable data sharing. Although a few reports in the literature describe adoption of this new diagnostic terminology in academic settings, to the best of our knowledge there are no reports on its use in non-academic laboratories. In this article we describe our almost 1-year experience with implementation of TBS in a community hospital-based laboratory.

Materials and Methods: A total of 724 thyroid FNA cases accessioned after implementation of TBS were compared with 629 cases from a similar time period immediately before adoption of TBS. In the pre-TBS period, diagnoses would be categorized as insufficient, negative, atypical, suspicious, and positive, while in the TBS period the diagnostic categories were insufficient, benign, atypia of undetermined significance, suspicious for follicular (or Hürthle cell) neoplasm, suspicious for malignancy, and malignant. For each case, the cytotechnologist's screening diagnosis and pathologist's final diagnosis were recorded. Histopathologic follow-up, if available, was also documented. Distribution of various diagnostic categories, diagnostic agreement between cytotechnologist and pathologist, and cytologic-histologic correlation were tabulated and tested for statistically significant differences in the pre-TBS and TBS periods.

Results: Distribution of various diagnostic categories in the pre-TBS and TBS periods is listed in Tables 1 and 2. There was a significant drop in the number of indeterminate (atypical or suspicious) diagnoses from 12.08% in the pre-TBS period to 4.70% in the TBS period (p<0.0001). Also, despite an increase in the number of diagnostic categories after adoption of TBS, exact diagnostic agreement between cytotechnologist and pathologist significantly increased from 77.83% to 83.22% (p<0.02). Although the number of cases with histopathologic follow-up was relatively small in our study, the rate of histologic malignancy in the various diagnostic categories roughly correlated with the rates predicted by TBS.

Conclusions: Use of TBS terminology for thyroid FNA cytology appears to significantly decrease indeterminate diagnoses and to improve cytotechnologist-pathologist diagnostic agreement. As TBS is adopted by a wider range of laboratories it is likely that these and other benefits will become more evident.

Table 1. Distribution of Diagnostic Categories before Implementation of TBS

Diagnostic Category	Number of Cases (%)
Insufficient	39 (6.20%)
Negative	504 (80.13%)
Atypical	49 (7.79%)
Suspicious	27 (4.29%)
Positive	10 (1.59%)
Total	629 (100.00%)

Table 2. Distribution of Diagnostic Categories after Implementation of TBS

Diagnostic Category	Number of Cases (%)
Insufficient	56 (7.73%)
Benign	599 (82.73%)
Atypia of undetermined significance	17 (2.35%)
Suspicious for follicular neoplasm	27 (3.73%)
Suspicious for malignancy	17 (2.35%)
Malignant	8 (1.10%)
Total	724 (100.00%)

51

Distribution of the Bethesda System -Thyroid Diagnostic Categories in African-American Population: Initial Report of an Urban Institution's Experience

Raja Gidwani, MD¹, Opada Alzohaili, MD², Julie Samantray, MD², Sudeshna Bandyopadhyay, MD¹, Jining Feng, MD, PhD¹, Dongping Shi, MD¹, Paul Tranchida, MD¹, Josephine Akinfolarin, MD¹, Mujtaba Husain, MD¹, Tamar Giorgadze, MD¹

¹Pathology, Detroit Medical Center, Detroit, Michigan; ²Endocrinology, Detroit Medical Center, Detroit, Michigan

Introduction: The incidence of benign or malignant thyroid lesions occurring within the African-American (AA) population is largely unrecognized. A recent study by Morris et al (2008) showed that although most cancers are more common in AA than Caucasian Americans, the incidence of thyroid cancer is nearly half of that of Caucasians. It is unclear whether this gap can be attributed to inferior cancer screening in the AA population or may represent a true population difference in thyroid cancer prevalence. To the best of our knowledge, we present the first study that examines the thyroid fine needle aspiration (FNA) results within the AA population.

Materials and Methods: During the period from January 1, 2009 until December 31, 2009, consecutive patients seen in a single endocrine practice were prospectively enrolled into this study. Demographic data (race, gender, age), social history (smoking), pertinent clinical history, FNA and follow-up surgical pathology results were collected. The FNA was performed by an attending endocrinologist and/or rotating endocrinology fellow. Cytopathologists performed onsite evaluations and triage. Paired slides were prepared from each pass for Diff Quik and Papanicolaou stain and needle rinses were used for cell block preparation. An aliquot was used for ancillary studies when appropriate. Once the FNA was complete, the sample was routinely processed, evaluated and reported according to The Bethesda System - Thyroid (TBS-Thy) terminology. Repeat FNA or surgical follow-up was recommended as per TBS-Thy guidelines.

Results: A total of 207 patients were enrolled in this study. 142 (68.60%) of the patients were AA, 9 (4.35%) were Arabic, 16 (7.73%) were Caucasian, 40 (19.33%) were of other ethnicities. The average age of the patients was 52.7 years (age range 19-89) while the average age of the AA group was 55.23 years (age range 19-89). Within the entire cohort there were significantly more females, 190 (91.79%) vs. 17, (8.21%), males. The FNA diagnoses were categorized into Classes I through VI per TBS-Thy. The distribution of these categories within the entire cohort is summarized in Table 1. Figure 1 shows the incidences of TBS-Thy categories in the AA and non-AA subgroups.

Conclusions: This pilot study suggests that the distribution of TBS-thyroid diagnostic categories in the African American population is similar to that in the reported general population study (Yang et al. 2007, Figure 1). However, a larger cohort study with longer follow-up is needed to validate our preliminary observations. We intend to continue adding patients into this database with prospective follow up including clinical, repeat FNA, ultrasonographic and surgical data (where applicable) for a comprehensive study.

Table 1. 

Ethnicity	Age (Range)	I	II	III	IV	V	VI	Total
AA	55 (19-89)	10	110	18	3	0	1	142
Arabic	43 (38-58)	1	7	1	0	0	0	9
Caucasian	48 (20-76)	1	13	0	1	0	1	16
Other	47 (20-69)	6	28	3	1	0	2	40
Total	53 (19-89)	18	158	22	5	0	4	207

52

HPV Infection and Oral Cytological Changes in Patients with Oropharyngeal Squamous Carcinoma and Their Sexual Partners

Ming Guo, MD¹, Nour Sneige, MD¹, Shobha Patel, BS¹, Jie Feng, MD¹, Vali Papadimitrakopoulou, MD², Erich Sturgis, MD², Carol Etzel, MD², Waun-ki Hong, MD², Anne Tsao, MD²

¹Pathology, The University of Texas MD Anderson Cancer Center, Houston, Texas; ²Thoracic/Head and Neck Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas

Introduction: High-risk HPV infection, especially HPV16, has been recognized as one of the major etiologies for the development of oropharyngeal squamous carcinoma (OPC). HPV infection is sexually transmitted. However, studies to document the HPV infections and cytology changes in patients with oropharyngeal carcinoma and their sexual partners are scant. The aim of the study is to observe the distribution pattern of HPV between patients with OPC and their sexual partners as well as oral cytological changes associated with HPV infections.

Materials and Methods: From 2006 to 2010, we collected 225 paired oral-pharyngeal cytology specimens from patients with OPC and their sexual partners using swabs/brushings. Pap-stained cytology specimens were prepared using SurePath method. HPV DNA testing and genotyping was performed using EasyChip HPV assay. HPV positivity and genotypes were compared between patients and their sexual partners. Cytological changes were compared with HPV DNA testing results.

Results: A total of 190 paired specimens were qualified for HPV testing. In 380 patients and their sexual partners, a total of 75 individual had positive HPV DNA testing results. Of these, 95% (71/75) had high-risk HPV infections including HPV 16, 31, 45, 51, 52, 56 and 59. In 190 patients with oropharyngeal carcinoma, 35 (18.4%) had positive HPV DNA testing. Of these, 54% (19/35) were HPV 16 infection. In 35 patients with positive HPV DNA testing results, the identical HPV genotypes were detected in 26 sexual partners (74%, 26/35). HPV DNA was also detected in 14 sexual partners (7.4%) only. Cytological changes such as parakeratosis and dyskeratosis/LSIL were observed more frequently in individuals with positive HPV testing results. No significant difference of hyperkeratosis was observed between HPV positive and negative groups.

Conclusions:1. A high prevalence of high-risk HPV infection in the sexual partners of the patients with oropharyngeal carcinoma and the high frequencies of identical HPV genotypes in patients and their sexual partners indicate the risk of HPV transmission between patients with oropharyngeal carcinoma and their sexual partners. 2. Parakeratosis and dyskeratosis/LSIL are more frequently associated with HPV infections in oral-pharyngeal cytology specimens. Oral cytology with HPV DNA testing may be useful in the follow up for patients with OPC.

53

Retrospective Study of 71 Cases of Fine Needle Aspiration Biopsies of Cystic Lesions of the Head and Neck

Shaolei Lu, MD, PhD, Latha Pisharodi, MD, Gengsheng Yu, MD, PhD

Pathology, Rhode Island Hospital, Providence, Rhode Island

Introduction: Fine Needle Aspirations (FNA) of cystic lesions of the Head and Neck are commonly performed procedures. These lesions include branchial cleft cyst, epidermal inclusion cyst, sialoadenitis, Warthin's tumor and metastatic cystic squamous cell carcinoma. Many of these lesions are benign and an unequivocal diagnosis can be made on careful cytologic examination on most occasions. However, metastatic cystic squamous cell carcinoma not uncommonly poses a difficult diagnostic challenge in the evaluation of a cystic lesion in the head and neck area. In this study, we retrospectively reviewed 71 cases of FNA biopsied cystic lesions of the head and neck area of which ten cases were diagnosed as cystic squamous cell carcinoma or suspicious for cystic squamous cell carcinoma. The cytologic features on FNA are summarized and compared with the benign cystic lesions, such as branchial cleft cyst and keratinous cysts.

Materials and Methods: We reviewed 71 cases of FNA biopsied cystic lesion of the head and neck area. The cytologic features evaluated included cellularity (scantily, moderately, and richly cellular), nuclear grade (low, intermediate and high grade), presence/absence of inflammation (chronic and/or acute), presence or absence of anucleate squamous cells and necrosis.

Results: Of the 71 cystic lesions of the head and neck evaluated, there were 10 cases of cystic squamous cell carcinoma, 26 cases of Warthin's tumor, 20 cases of benign simple cyst; 5 cases of epidermal inclusion cyst; 3 cases of branchial cleft cyst; 1 case of thyroglossal duct cyst; 6 cases of others (abscesses, lymph node and low grade epithelial neoplasm). All of the 10 cases with squamous nuclear atypia had subsequent excisional biopsies and the surgical pathologic diagnoses included 2 cases of branchial cleft cyst, and 8 cases of squamous cell carcinoma.

Conclusions: Most cystic lesions in head and neck area are benign entities. The cellularity of these lesions is usually low but concurrent inflammation may increase the background cell density. Nuclear grade and necrosis are very useful indices to differentiate benign from malignant lesions, even in a hypocellular specimen. A benign diagnosis of cystic lesion can be made easily on cytologic examination, but well differentiated squamous cell carcinoma must be among the differential diagnoses if there is intermediate and above nuclear atypia.

54

Application of the New Bethesda Thyroid Classification on Previously Diagnosed Thyroid Cytology Specimens and Correlation with Surgical Follow-Up

Matthew Martin, MD, Maria Policarpio-Nicolas, MD, I-Tien Yeh, MD, Philip Valente, MD

Pathology, University of Texas Health Science Center, San Antonio, Texas

Introduction: The recently proposed Bethesda System for Reporting Thyroid Cytopathology provides guidance for clinical follow up with associated risk of malignancy for each diagnostic category. The diagnostic criteria and categories additionally provide guidance for more uniform reporting and relate to easier discernment for clinical colleagues. The objective of this study was to report our experience utilizing the Bethesda System for Reporting Thyroid Cytopathology criteria on previously diagnosed thyroid fine needle aspiration (FNA) specimens and correlation with the surgical excisions.

Materials and Methods: A computer search of thyroid lesions diagnosed by FNA with a corresponding surgical specimen from January 2008 to December 2009 was performed. These cases were then distributed to three pathologists for review and re-classification using the Bethesda System for Reporting Thyroid Cytopathology (unsatisfactory, benign, atypical cells of undetermined significance, suspicious for follicular neoplasm, suspicious for malignancy, and malignant). The pathologists were blinded to the clinical data, prior cytology diagnosis and corresponding surgical pathology findings. Comparison with prior cytology and final surgical excision diagnoses was performed.

Results: A total of 84 cases were re-classified using the Bethesda system with results described in Table 1. The original cytology diagnoses often used descriptive terms. When comparing the original cytology reports to the applied Bethesda system, the original diagnosis correlated with the surgical excision in 74 of 79 cases (94%), while the Bethesda criteria correlated in 73 of 80 cases (91%). The atypical “categorical term” was used in 17% of originally reported cases, while the atypical category was used in only 9.5% of cases using the Bethesda criteria. Additionally, 30 cases (36%) showed a categorical change when compared to the original diagnoses. Of these, 28 of 29 originally diagnosed cases (97%) correlated with the surgical excision, compared to 25 of 29 cases (86%) using the Bethesda system.

Conclusions: Prior thyroid cytology diagnoses had approximately twice the rate of “atypical” interpretation, when compared to applying the Bethesda system, and resulted in surgical excisions. These findings are most likely due to lack of prior definitive criteria and uniformity in the utilization of the atypical category. The atypical category in the Bethesda system has defined criteria with a recommendation for repeat FNA. Additionally, given a 91% surgical correlation with the utilization of the Bethesda classification, experience in applying the diagnostic criteria and categories is essential for accuracy and better patient management.

Table 1. 

Diagnostic Category	Original Diagnosis	Re-Classified Diagnosis
Unsatisfactory/Non-diagnostic	5 cases (6%)	4 cases (4.8%)
Benign	32 cases (38%)	47 cases (56%)
Atypical	14 cases (17%)	8 cases (9.5%)
Suspicious for Follicular Neoplasm	13 cases (15%)	9 cases (10.7%)
Suspicious for Malignancy	9 cases (11%)	6 cases (7.1%)
Malignancy	11 cases (13%)	10 cases (12%)

55

Assessing the Utility of Fine-Needle Aspiration in Male Patients with Thyroid Nodules

Kristen Mead, MBBS¹, Roberto Izquierdo, MD², Cindy Steele, SCT(ASCP)¹, Kamal Khurana, MD¹

¹Pathology, SUNY Upstate Medical University, Syracuse, New York; ²Endocrinology/Medicine/Pediatrics, SUNY Upstate Medical University, Syracuse, New York

Introduction: Fine-needle aspiration (FNA) is considered the first line of investigation for evaluation of thyroid nodules. The risk of malignancy in males with thyroid nodules is significantly higher. A recent study concludes that FNA biopsy is unnecessary in all males with thyroid nodules due to high false negative rates and recommends direct surgical management (Berri et al., Am J Surg, 195 (2008) 396-400). We retrospectively analyzed our database to determine the utility of FNA biopsy of thyroid nodules in male patients at our institution.

Materials and Methods: We retrospectively reviewed 149 male patients who underwent a total thyroidectomy from January 2004 to June 2009 at our institution. The pre-operative FNA reports were reviewed and correlated with the surgical specimen diagnosis. The accuracy of thyroid FNA in determining final pathology was analyzed.

Results: Our patients ranged in age from 6 to 89 years with an average age of 48.6 years. Fifty-one of 149 patients (34.2%) had a final histologic diagnosis of malignancy and of these patients 23 (45.1%) had a prior FNA with an accurate cytologic diagnosis in 22 of 23 (95.7%) cases. The incidence of malignancy was higher in patients who underwent preoperative thyroid FNA compared to patients who directly underwent thyroidectomy (47.9% versus 27.7%, p<0.02). Of the 48 patients that underwent preoperative FNA, the cytologic diagnoses were classified as follows: positive for malignancy 17 (35.4%), indeterminate 13 (27.1%), benign 15 (31.3%) and unsatisfactory 3 (6.3%). For calculation of sensitivity and specificity, lesions with a diagnosis of “indeterminate” or “positive for malignancy” were combined. Surgical follow-up revealed eight false positive cases and one false negative case. Follow-up on all unsatisfactory cases revealed benign thyroid lesions and these were excluded from calculations. The thyroid FNA procedure in males at our institution had a sensitivity, specificity, positive predictive value and negative predictive value of 95.7%, 63.6%, 73.3% and 93.3%, respectively. The false negative rate was 4.3%.

Conclusions: We conclude FNA biopsy of thyroid nodules in male patients has an acceptably low false negative rate. The incidence of cancer in male patients that undergo FNA and thyroidectomy is significantly higher than male patients that have thyroidectomy alone. Fine needle aspiration is a useful diagnostic tool for evaluation and triaging male patients with thyroid nodules to surgery.

56

Correlation of Fine Needle Aspiration Cytology with Lymph Node Thyroglobulin Levels in the Detection of Recurrent/Metastatic Papillary Thyroid Carcinoma: A Retrospective Analysis with Histological Follow-up

Summer Nugent, MD¹, Joely Straseski, PhD², Barbara Rosenthal, CT(ASCP)¹, Douglas Clark, MD¹, Qing Li, MD, PhD¹

¹Pathology, Division of Cytopathology, The Johns Hopkins Hospital, Baltimore, Maryland; ²Pathology, Division of Clinical Chemistry, The Johns Hopkins Hospital, Baltimore, Maryland

Introduction: Papillary thyroid carcinoma (PTC) is the most predominant thyroid malignancy, accounting for 75-90% of all thyroid malignancies. Although the majority behaves in an indolent fashion, approximately 5-20% of patients develop metastases, most commonly to regional lymph nodes. PTC patients may develop recurrences several years after their initial diagnosis and/or treatment, therefore, long-term follow-up is needed. Currently ultrasound-guided fine-needle aspiration (US-FNA) cytology is a commonly used method in the detection of clinically suspicious cervical lymph nodes, with 76-98% sensitivity and 92.9-100% specificity. However, it also shows a 5-10% non-diagnostic rate and a 6-8% false negative rate. Measuring thyroglobulin (Tg) levels in US-FNA specimens has been increasingly used as an additional tool to increase diagnostic yield. However, the usefulness of this assay has been neither well studied nor standardized. In this study, we investigated the utility of the US-FNA-Tg assay and correlated Tg levels with cytological and, when available, histological diagnoses. The limitations of this test are also discussed.

Materials and Methods: The archive of the Department of Pathology at The Johns Hopkins Hospital was searched for metastatic PTC of cervical lymph node over a period of 10 years. A total of 254 cases were identified. Among them, 36 cases had FNA cytology and corresponding US-FNA-Tg measurements. The cytological and histological diagnoses were correlated with US-FNA-Tg levels.

Results: Among the 36 cases, the M:F ratio was 1:2.3 with a median age of 40 years (range from 17-67 years). Within the 36 cases, 20 cases with cytological diagnoses of negative lymph nodes had a mean US-FNA-Tg level of <0.7 ng/mL (range from <0.1 to 0.7 ng/mL with one measuring 3.8 ng/mL). Of 16 cases with morphological diagnoses of metastatic PTC, 11 cases (68.8%) had elevated US-FNA-Tg levels and 5 cases (31.2%) had US-FNA-Tg levels below 50 ng/mL (Table 1). Furthermore, among cases of metastatic PTC, one case with cytology suspicious for PTC had US-FNA-Tg levels of 1645 ng/mL; and two cases with atypical cells on cytology had US-FNA-Tg levels of 40765 ng/mL and 2.4 ng/mL. All three cases had follow-up surgery that revealed metastatic PTC. The average size of lymph nodes was 1.3 cm in the negative lymph node group and 1.4 cm in the metastatic PTC lymph node group (P>0.05).

Conclusions: US-FNA-Tg is a semi-quantitative test. Our data demonstrated that US-FNA-Tg had a sensitivity of 62.5% and specificity of 100% when morphology is used as a gold standard with an arbitrary cutoff level of 50 ng/mL. The test may be useful in cases with focal lymph node involvement and rare tumor cells present on smears. However, caution must be exercised in the interpretation US-FNA-Tg results, since several factors may affect the accuracy of the test such as sampling error, the presence of autoantibodies, radioiodine ablation therapy, sample volume, and other laboratory variables. A large scale study is still needed to validate the usefulness of this test in daily cytological practice.

Table 1. Thyroglobulin levels in lymph nodes with metastatic PTC

Case	Age	Sex	S/p thyroid-ectomy (Mos)	FNA Tg (ng/mL)	LN size (mm)	Cytology
PTC: papillary thyroid carcinoma. S/p: status-post. Mos: months. FNA: fine needle aspiration. Tg: thyroglobulin. LN: lymph node. *Cytology diagnosis: atypical/suspicious for PTC.
1	49	M	0	245090	12	Metastatic PTC
2	48	F	168	40765*	16	Atypical cells
3	32	M	6	23143	30	Metastatic PTC
4	28	F	24	14171	13	Metastatic PTC
5	49	F	48	6631	12	Metastatic PTC
6	49	F	60	5217	6	Metastatic PTC
7	49	F	60	2579.7	20	Metastatic PTC
8	47	M	132	2470.7	11	Metastatic PTC
9	20	M	36	1645.1*	9	Suspicious for PTC
10	37	F	12	100.4	6	Metastatic PTC
11	60	F	36	59	18	Metastatic PTC
12	34	M	24	2.4*	13	Atypical cells
13	39	F	60	<0.1	22	Metastatic PTC
14	40	F	<1 (20 days)	0.9	11	Metastatic PTC
15	34	F	<1 (16 days)	0.2	11	Metastatic PTC
16	56	F	48	<0.2	6	Metastatic PTC
Total: 16					Mean=13.5

57

Fine Needle Aspiration of Metastatic Solid Neoplasms to the Thyroid: A Single Institution Experience

Jordan Reynolds, MD, Paul Lappinga, MD, Michael Henry, MD

Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota

Introduction: Fine needle aspiration of the thyroid has proven to be useful in the assessment of primary thyroid pathology. Metastatic solid neoplasms to the thyroid are rare and the usefulness of FNA remains unclear. Our goal is to report our FNA experience in detecting metastases in a series of patients over the last 30 years.

Materials and Methods: The Mayo Clinic archived cytologic material from 1980-2010 was searched for cases that were diagnosed as a metastatic solid neoplasm on cytology (excluding lymphomas). Thyroid resection specimens, when available, were also reviewed to support the cytologic diagnosis and to evaluate parameters of the tumor (size and whether it was unifocal or multifocal). All thyroid resection cases were then re-reviewed and those containing a metastatic solid neoplasm, along with the corresponding FNA, were included. Statistical analysis was performed to compare survival between patients who underwent thyroid resection versus those who did not have surgery.

Results:10,798 thyroid FNAs with concomitant thyroid resection specimens were performed from 1980-2010. 1,803 (16.7%) total malignancies were identified, and of these 72 (0.6%) were originally diagnosed as a metastatic neoplasm on FNA. An additional 5 cases were uncovered after reviewing all thyroid resection cases (4 were diagnosed cytologically as positive for papillary thyroid carcinoma and 1 was diagnosed positive for follicular neoplasm). None of the cases was called negative on cytology, and similarly, there were no false positive diagnoses for metastases. The sensitivity for detecting metastases was 94%, and the specificity was 100%.

Fifty-nine cases had sufficient data for further analysis, of which the results are shown in Tables 1 and 2. The average age was 63.8 years (range 40-89 years) and 32 patients were male and 27 were female. Notably, 19 patients underwent thyroid resection. The average metastatic tumor size was 4.5 cm. 16 cases had one distinct mass while 4 had multiple masses. Overall, survival in all patients with metastases was mean 38.5 months (median 9 months, range 2-167 months). Survival in patients who underwent thyroid resection for metastases was mean 38.5 months (median 34 months, range 8-89 months) while survival in those patients who did not proceed to thyroid surgery was mean 38.5 months (median 6 months, range 2-167 months, log-rank test p=0.04).

Conclusions: The most common primary origin of metastases was from the kidney, lung, and head and neck sites. The most common tumor types were carcinomas, with adenocarcinoma being the most common followed by squamous cell carcinoma. Metastatic mesenchymal, endocrine and melanocytic neoplasms were less common. The follow up thyroid resection specimens showed that most tumors were unifocal with an average size of 4.5 cm. Though they are exceedingly rare, in general the presence of thyroid metastases was a terminal event with a median survival of 9 months. Surgically treated patients demonstrated increased survival, and this demonstrated statistical significance; however, other underlying factors such as presence of metastases to other sites may have an effect. Fine needle aspiration remains a sensitive and specific way to detect metastatic solid neoplasms to the thyroid.

Table 1. Primary site of origin

Head and Neck	9 (15.2%)
Esophagus/GI	6 (10.1%)
Lung/Thoracic	12 (20.3%)
Breast	3 (5.0%)
Kidney	13 (22.0%)
Skin	3 (5.0%)
Ovary/uterus	3 (5.0%)
Other (colon, liver, prostate, spine)	4 (6.7%)
Unknown	6 (10.1%)
Total	59

Table 2. Tumor Type

Adenocarcinoma	32 (54.2%)
Mesenchymal tumors - (2 malignant peripheral nerve sheath tumors, 1 synovial sarcoma, 1 chordoma) Endocrine (6 metastatic neuroendocrine tumor, 1 metastatic paraganglioma).
Squamous cell carcinoma	13 (22.0%)
Mesenchymal	4 (6.7%)
Endocrine	7 (11.9%)
Melanoma	3 (5.0%)
Total	59

58

The Bethesda System for Reporting Thyroid Cytopathology: the UCLA Experience

Gregory Shaw, BS, Sharon Hirschowitz, MD, Lalita Kumar, BS, Peggy Sullivan, MD

Pathology, UCLA, Los Angeles, California

Introduction: The proposal to standardize thyroid fine needle aspiration (FNA) reporting, The Bethesda System for Reporting Thyroid Cytopathology (TBS), estimates both a distribution and risk of malignancy for six diagnostic categories including Non-diagnostic, Malignant, Suspicious for Malignancy, Follicular Neoplasm, Benign, and “Follicular lesion of undetermined significance” (FLUS). We sought to compare our institution's experience with the proposed diagnostic categories and their respective risks of malignancy.

Materials and Methods:3117 thyroid FNAs were performed at our institution between 2003 and 2007. Cases were re-classified using proposed TBS criteria. Patients with a history of thyroid carcinoma or thyroidectomy were excluded. The most abnormal cytologic diagnosis was retained for each patient. Of the remaining 2057 cases, the distribution into six diagnostic categories, surgical follow-up, and subsequent surgical diagnosis were recorded. The risk of malignancy percentage was calculated based on excised nodules only. Incidental papillary thyroid carcinomas (<1 cm) were not considered malignant. Follicular lesion of undetermined significance/atypia of undetermined significance (FLUS) and Follicular Neoplasm/Suspicious for Follicular Neoplasm (FN) categories were further stratified into Hurthle and Non-Hurthle categories.

Results: See Tables 1 and 2

Conclusions: Overall, the distribution and malignancy risks as reported from our institution are comparable to those reported by others (Faquin et al., 2010; Jo et al., 2009). The Non-Diagnostic category shows a slightly higher risk of malignancy than expected, highlighting the importance of further follow-up in these cases. Although numbers are small, Hurthle subclassification does not portend a higher risk of malignancy in either FLUS or FN categories.

Table 1. Outcome of Thyroid FNA cases Re-Classified Using TBS criteria

Diagnosis	Cases # (%)	Excision # (%)	Malignancy # (%)
Non-Diagnostic	286 (14)	38 (13)	4 (11)
Benign	1412 (69)	84 (6)	2 (2)
FLUS	87 (4)	35 (40)	7 (20)
FN	71 (3)	40 (56)	14 (35)
Suspicious for Malignancy	64 (3)	49 (77)	39 (80)
Malignant	137 (7)	100 (73)	99 (99)

Table 2. FLUS and FN Classified to Non-Hurthle and Hurthle Subcategories

Diagnosis	Cases # (%)	Excision # (%)	Malignancy # (%)
Non-Hurthle FLUS	79	33 (42)	7 (21)
Non-Hurthle FLUS	8	2 (25)	0 (0)
Non-Hurthle FN	44	24 (55)	10 (42)
Hurthle FN	27	16 (59)	4 (25)

59

The Dilemma of Cystic Thyroid Lesions in The Bethesda System for Reporting Thyroid Cytopathology: Is There a Need for Reappraisal?

Jeremy Spencer, MD, M. Katayoon Rezaei, MD, Sana Tabbara, MD

Pathology, George Washington University, Washington, District of Columbia

Introduction: According to the Bethesda System for Reporting Thyroid Cytopathology (TBS-TC) nomenclature, cystic lesions (CL) with or without histiocytes and not meeting the adequacy criteria, are now classified as “nondiagnostic or unsatisfactory” (UNSAT). While “cyst fluid only” may be included in the report, an UNSAT diagnosis may not acknowledge the presence of a lesion, including a possible cystic papillary carcinoma. It may also reflect negatively on the individual performing the fine needle aspiration (FNA). The aim of the study is to assess the value of including a CL diagnostic category.

Materials and Methods: All thyroid FNAs diagnosed as CL in our institutional database in the last ten years were retrieved and follow-up (F/U) information was obtained. The original diagnoses included CL not otherwise specified (CL NOS), CL with atypia (CL ATYP), and CL suspicious for papillary carcinoma (CL SUSP). All available slides from cases with F/U FNA or thyroidectomy were reviewed and reclassified according to the TBS-TC, while maintaining a CL NOS category. A CL NOS diagnosis required the presence of numerous histiocytes and less than six groups of ten follicular cells. Otherwise, the cases were classified based on the findings in the follicular cell component; follicular lesion of undetermined significance (FLUS) when cytologic atypia was qualitatively not sufficient for a definitive diagnosis, and suspicious for papillary carcinoma (SUSP PC) when cytologic features of papillary carcinoma were present but quantitatively insufficient for a definitive diagnosis.

Results: A diagnosis of CL was originally made in 117 cases. Those comprised 102 CL NOS, 8 CL ATYP, and 7 CL SUSP. F/U was available in 29/102 CL NOS, 6/8 CL ATYP, and 6/7 CL SUSP. Original and review diagnoses are summarized in Table 1. A total of fourteen cases were reclassified as Non-CL category based on cellularity. All but one of the cases originally diagnosed as CL ATYP and CL SUSP, had sufficient follicular component in addition to the cystic background to be reclassified as benign, FLUS, or SUSP PC. Four cases from CL NOS were reclassified as Benign (two) and FLUS (two). Histologic diagnosis of papillary carcinoma was made in 8/8 (100%) SUSP PC, 2/4 (50%) FLUS and 1/20 (5%) CL NOS.

Conclusions: Review of cystic lesions diagnosed at our institution emphasizes the importance of adhering to strict criteria when diagnosing CL NOS. As such, in our hands, this category is associated with a 5% risk for papillary carcinoma that falls within the range for FLUS (5-15%). Classifying CL NOS under the UNSAT category seems to contradict the goal of TBS-TC in providing useful diagnostic information and the associated risk of malignancy for optimal clinical management. It may be more informative to maintain this category as it acknowledges the presence of a lesion that can further be assessed and followed up using imaging characteristics and clinical parameters.

Table 1. Summary of CL original and review diagnosis

	Review Diagnosis
OriginalDiagnosis (n)	CL NOS	Benign	FLUS	SUSP PC	UNSAT	NA*
*Slides not available for review.
CL NOS (29)	20	2	2	0	0	5
CL ATYP (8)	0	1	2	2	1	2
CL SUSP (7)	0	0	0	6	0	1
Total (44)	20	3	4	8	1	8

60

Sonographically Guided Thyroid FNA, Experience of a Private Laboratory

Guillermo Villarmarzo, MD, Wilma Virella-Santana, MD

Hato Rey Pathologist Associates, San Juan

Introduction: Fine-needle aspiration biopsy (FNAB) is the gold standard for evaluation of thyroid nodules. It has decreased unnecessary surgery by up to 50%. Ultrasound (US) guidance has further improved the technique by allowing non palpable nodules to be detected, palpable nodules to be confirmed as true nodules, and assuring correct needle positioning within the nodule.

Our pathology practice performs US-assisted thyroid FNAB. This combination confers the advantages of US evaluation of the thyroid, immediate sample evaluation by the pathologist for repeat sample when necessary, and assessment by the pathologist with clinical and sonographic information of thyroid and surrounding neck structures.

Materials and Methods: From 2005 to 2009 a total of 37961 thyroid nodules were aspirated by three pathologists. All aspirations were performed and evaluated by the same pathologist performing the aspiration. Two passes are usually performed per nodule, more if additional material is needed. The material is expelled into one labeled glass slide using the syringe, separated into two smears and the smears are air-dried. One slide is stained with Diff-Quick for immediate evaluation and the other re-hydrated and stained by conventional pap stain for later evaluation. Adequacy criteria were to have 6 well preserved groups of at least 10 follicular cells each in two slides with some exceptions. For a subset of cases that was treated by surgery correlation was done with descriptive statistics.

Results:37,961 nodules were aspirated, of which 35,969 were performed under sonographic guidance. 86% of the US-FNAB and 92.2% of those performed by palpation were negative. Most of them were adenomatous nodules. 2.2% of the US-FNAB and 2.4% of those performed by palpation were positive, the majority being papillary carcinomas. 6.0% of the US-FNAB and 5.1% of those performed by palpation fell within the suspicious category.

For 1001 of these cases a surgical specimen was available. Using this data the calculated sensitivity was 96%, specificity 100%, positive and negative predictive values were 100% and 96% respectively, accuracy was 98.9%, false negative and false positive rates were 4% and 0% respectively.

The unsatisfactory rate was 5.0%. It is of note that we detected carcinomas within sub centimeter nodules. This is something that we plan to quantify in a future study.

Conclusions: This is a retrospective study of thyroid FNAB, most US-guided, performed in four clinics served by our laboratory. The procedures were performed by three pathologists, who each evaluated the samples and rendered the final diagnoses on their cases. Our numbers regarding distribution of thyroid diagnoses are in agreement with the literature. Our performance statistics compare very favorably with the literature. Our unsatisfactory rate is very low, compared to reported values in the literature of 20%. There are unquantified findings of many malignancies in subcentimeter nodules, which require US for detection, further strengthening the importance of US. US also avoids unnecessary FNAB, since many palpable nodules turned out not to be confirmed by US. The combination of US guidance with on-site evaluation of the sample by an experienced pathologist constitutes the best setup for thyroid nodule evaluation.

61

Carcinoma in Thyroid Nodules Diagnosed as Follicular Lesion of Undetermined Significance and Follicular Neoplasm in Fine Needle Aspiration (FNA)

Ann Walts, MD, Shikha Bose, MD

Pathology & Laboratory Medicine, Cedars-Sinai Medical Center, Los Angeles, California

Introduction: The 2007 National Cancer Institute (NCI) system for diagnosis and reporting thyroid FNAs describes six diagnostic categories each with an assigned “risk of malignancy”. Of these, “follicular lesion of undetermined significance (FLUS)” and “follicular neoplasm (FN)” with 5-15% and 15-30% risk of malignancy, respectively, are the most challenging for diagnosis and management. Our cytology laboratory has utilized the NCI scheme to report all thyroid FNAs since September 2008. This prospective study was designed to assess the risk of malignancy in lesions diagnosed as FLUS and FN in FNA.

Materials and Methods: All thyroid FNAs reported as FLUS (N=142) or FN (N=33) from 9/1/2008 to 1/1/2010 were retrieved from our departmental files. Three to 19 months (median 8 mos) later, files were searched for cytology and surgical follow-up for each patient. Demographic and pathological data were collected from clinical charts and pathology reports. Statistical analysis was performed utilizing Fisher's exact test.

Results: The 142 FLUS and 33 FN comprise 8.2% and 1.9%, respectively, of all 1722 thyroid FNAs diagnosed in our laboratory during the study period. Tissue followup became available for 37 FLUS and 19 FN FNAs. An additional 14 FLUS and 2 FN cases had only repeat FNA. As shown in Table 1, at a median followup of 8 months, there was a statistically significant difference in the percentage of FLUS and FN FNAs with tissue followup (26%FLUS vs 58%FN; P=0.0001), but no significant difference in the frequency of malignancy (38%FLUS vs 37%FN) diagnosed in the thyroid resected from the two groups. 86% of the carcinomas in the FLUS group were papillary and 14% were follicular, while in the FN group 57% of the carcinomas excised were papillary and 43% were follicular; however, this difference was not statistically significant. The interval from FNA to excision was longer in the FLUS group (median 2 vs 1 mo) and the median size of the lesion (irrespective of final diagnosis) was smaller in the FLUS group. Patient age and female predominance were similar in both groups and no statistically significant difference was observed in percentage of repeat FNAs, extent of resection, or frequency of lymph node excision, chronic thyroiditis, intraoperative frozen section, or consultation among surgical pathologists prior to final signout.

Conclusions:

• 1
  FNA diagnosis of FLUS vs. FN differentially impacts clinicians' selection of cases for surgical excision.
• 2
  At a median followup of 8 months, the frequency of malignancy in FLUS and FN exceeded that predicted by the NCI scheme and was not significantly different in the two groups. This could reflect a bias toward early excision of the subset of FLUS clinically considered at greatest risk for malignancy (history of radiation exposure, family history of thyroid carcinoma, increasing lesional size, progressive symptoms). If so, we would anticipate that the percentage of malignancy in the FN group will exceed that in the FLUS group as additional followup becomes available. Further prospective studies are needed to support the premise that the FLUS and FN categories identify groups of patients with non-overlapping risks of malignancy.
• 3
  Our findings suggest that papillary carcinomas are more likely to be interpreted as FLUS while follicular carcinomas are more likely to be interpreted as FN in FNA.

Table 1. Comparison of FLUS and FN Cases

	FNA Dx FN (N=33)	FNA Dx FLUS (N=142)	P value *
Cases with followup	21 (64%)	51 (37%)	0.0004
Surgical followup	19 (58%)	37 (26%)	0.0001
Repeat FNA only	2 (6%)	14 (10%)	0.74
Interval FNA to surgery	0.5–3.5 mos (median 1 mo)	0.5–13.5 mos (median 2 mos)
Cases with surgical followup
Patient age	25–78 yrs (median 49 yrs)	27–74 yrs (median 48 yrs)	0.75
Patient sex	13 (68%)	28 (76%)
female	6 (32%)	9 (24%)
male
Initial surgical procedure/excision of aspirated lesion			0.57
partial thyroidectomy	6/19 (32%)	15/37 (40.5%)
total thyroidectomy	13/19 (68%)	22/37 (59.5%)
Lymph nodes excised with excision of aspirated lesion			0.78
Yes	8/19 (42%)	18/37 (49%)
No	11/19 (58%)	19/37 (51%)
Weight of thyroid tissue excised	9-108 gms (median 32)	4-77 gms (median 24)
Size of lesion excised
irrespective of final diagnosis	1.1–5.8 cm (median 2.35)	0.9–5.3 cm (median 2.05)
final diagnosis carcinoma	1.1–4.5 cm (median 2.2)	0.9-5.0 cm (median 1.35)
Histological Diagnosis	19	37
Carcinoma	7	14	1
Follicular CA	3 (43%)	2 (14%)	0.28
Papillary CA	4 (57%)	12 (86%)
conventional	3	5
vollicular variant	1	7	0.77
Benign	12	23
follicular adenoma	11	12	1
nodular hyperplasia/goiter	1	11	1
CA as % of FNAs with any fup **	7/21 (33%)	14/51 (27%)
Benign as % of FNAs with any fup	14/21 (67%)	37/51 (73%)
CA as % of FNAs with tissue fup	7/19 (37%)	14/37 (38%)
Benign as % of FNAs with tissue fup	12/19 (63%)	23/37 (62%)

62

Thyroid Fine-Needle Aspiration Category of “Follicular Lesion of Undetermined Significance”- An Experience of 1382 cases in a Community Practice Setting

Howard Her-Juing Wu, MD¹, Crystal Rose, MD², Tarik Elsheikh, MD²

¹Pathology and Laboratory Medicine, Indiana University, Indianapolis, Indiana; ²Pathology, Ball Memorial Hospital, Muncie, Indiana

Introduction: National Cancer Institute (NCI) State of Science Conference on thyroid fine-needle aspiration (FNA) proposed a diagnostic category of “Follicular Lesion of Undetermined Significance” (FLUS), which includes a heterogeneous group of lesions that are not convincingly benign and are not diagnostic of a neoplasm. It was suggested that this diagnosis should be limited to less than 7% of all thyroid FNAs.

Materials and Methods: In our practice we receive thyroid FNA specimens from 7 community hospitals and many clinicians' offices. We conducted a computer search for all thyroid FNAs performed between 2006 and 2008. Histologic follow-up was retrieved. Typically, each case consisted of direct smears with or without cell blocks, prepared from 3 to 8 passes. Diagnostic terminology using the recommended 6 categories of the NCI conference classification were used, including benign (B), follicular lesion of undetermined significance (FLUS), follicular neoplasm (FN), suspicious for malignancy (SM), malignancy (M) and non-diagnostic (ND), was applied to 1382 consecutive thyroid FNAs. The female to male ratio was 5:1 and the age distribution ranged from 11 to 97 years (mean 56 years). The interpretation was rendered by 4 pathologists.

Results: FLUS diagnosis accounted for 27% (376 cases) of all thyroid FNAs. Follow up surgical excision was available in 51 FLUS cases (14%). Eleven cases (22%) were found to have malignant diagnosis including 8 papillary microcarcinomas (< 1 cm) and 3 follicular variant of papillary carcinoma. The use of FLUS terminology varied greatly among pathologists ranging from 11% to 47% (see Table 1). The pathologists with less than 5 years experience or no cytopathology fellowship training tended to more commonly use the FLUS terminology. On site evaluation was performed on 732 cases with a FLUS rate of 26%, which is slightly less than the FLUS rate in cases without on site evaluation (29%). Excluding the microcarcinoma, the cancer rate for FLUS cases was 6% among the cases with histologic followup.

Conclusions: The term FLUS was more frequently utilized in this community setting than was suggested by the NCI conference. The frequency of using this terminology varied widely among pathologists, but was higher among pathologists with less experience or no cytopathology fellowship training. The cancer rate, based on histologic followup, was relatively low for FLUS; therefore followup with repeat FNA, as suggested by NCI conference, appears appropriate.

Table 1. Thyroid FNA

Pathologist	Years experience	Fellowship training	B	FLUS	FN	SM	M	ND	Total
A	>10	Yes	216 (51%)	89 (21%)	38 (9%)	11 (3%)	9 (2%)	57 (14%)	420
B	>10	Yes	157 (42%)	43 (11%)	46 (12%)	14 (4%)	12 (3%)	105 (28%)	377
C	<5	Yes	79 (34%)	79 (34%)	17 (15%)	8 (4%)	7 (3%)	41 (18%)	231
D	>10	No	87 (25%)	165 (47%)	15 (4%)	3 (1%)	9 (3%)	75 (21%)	354
Total			539 (39%)	376 (27%)	116 (8%)	36 (3%)	37 (3%)	278 (20%)	1382

63

Review of Atypical Cytology of the Thyroid Gland According to the Bethesda System with Ultrasonographic Correlation

Young Yoo, MD, PhD¹, Hak Lee, MD¹, Han Jung, MD, PhD¹, Hyun Choi, MD, PhD¹, Young Yoo, MD, PhD¹, Jin Kang, MD, PhD¹, Suk Kang, MD, PhD²

¹Department of Pathology, St. Vincent Hospital, The Catholic University of Korea, Suwon, Gyeonggi-do, Republic of Korea; ²Department of Pathology, Yeouido St. Mary's Hospital, The Catholic University of Korea, Seoul, Seoul, Republic of Korea

Introduction: Fine needle aspiration (FNA) cytology is a gold standard for the preoperative diagnosis of thyroid nodules. Recently, the Bethesda system was introduced for reporting thyroid cytology. In this system category of atypia of undetermined significance (AUS) contains broad spectrum. The aim of this study was to select confidential cytologic features for the appropriate grouping of previously reported atypical cytology into categories other than AUS recommended in the Bethesda system by reviewing the atypical cytology performed in our institute with histologic-ultrasonographic correlation and thereby to make reference to reduce the results of AUS.

Materials and Methods: For the entire series of 3295 cases of FNA cytology of the thyroid and operated cases including 662 papillary carcinoma (PC), 714 nodular hyperplasia (NH), 68 follicular adenoma (FA), 10 follicular carcinoma (FC), 14 Hürthle cell adenoma (HCA) and 41 Hashimoto's thyroiditis (HT) which were performed from 2005 to 2009, 244 cases (7.4%) of atypical cytology were collected which could be matched with histologic results. After slide review cytologic findings were grouped according to dominant features and histologic-ultrasonographic correlation was made with these cases.

Results:244 cases generally showed low cellularity but grouping by dominant cytologic features was possible. The histologic results of these groups were as follows: All 115 cases with nuclear groove and pseudoinclusion were proved PC. 45 with nuclear groove without pseudoinclusion were proved 30 (67%) PC and 15 (33%) NH. 14 with pseudoinclusion only were 9 (64%) PC and 5 (37%) NH. 5 pseudoinclusion with Hürthle cell change (HCC) were 4 (80%) NH and 1 (20%) HT. 25 nuclear groove with HCC were 11 cases (44%) PC, (12%) HT, 2 (8%) HCA and 9 (36%) NH. 10 pseudoinclusion with psammoma body were 4 cases (44%) PC and 6 cases (56%) NH. All 10 HCC with squamoid cytoplasm were NH. 10 papillary architectures were 3 cases (30%) PC and 7 (70%) NH. Among surgically proven papillary carcinoma ultrasound provides suspicious for malignancy in 542 cases (84%), indeterminate 92 (14%) and benign 28 (3%). For the benign conditions ultrasonographic findings were well correlated.

Conclusions: Categorization of PC or suspicious for PC can be made safely with nuclear groove with pseudoinclusion even though cytologic interpretation was limited by low cellularity. The cases with nuclear groove alone also have high probability for PC. Pseudoinclusion alone is not a significant finding than nuclear groove for the diagnosis of PC. If HCC is accompanied with nuclear groove PC was probable. HCC and psammoma body without nuclear groove were not meaningful findings for the diagnosis of PC. HCC alone was more probable to benign conditions such as NH or HT. Although papillary architecture is meaningful, its nuclear feature should be examined. By cytology alone diagnostic accuracy of FA, HCA or HT is very low. Ultrasound alone is not diagnostic but can be a helpful adjuvant method for the cytologic interpretation. If refer to these conclusions incidences of AUS results can be reduced.

Laboratory Operations

64

Critical Comparison of Two Techniques for Cell Block Preparation: A Study of 117 Cases

Linda Payne, BS, Moisha Wright, BS, Ashutosh Lodhi, MBBS, Umadevi Narra, BS, Rosa Ramirez, BS, Shobha Patel, BS, Gregg Staerkel, MD, Savitri Krishnamurthy, MD

Cytopathology, MD Anderson Cancer Center, Houston, Texas

Introduction: A standardized and optimal cell block preparation technique is critical for any cytology laboratory to generate high quality cell blocks. Several techniques are available for retrieving and collecting cells and tissue fragments in liquid suspension in a tissue block. We compared two techniques for cell block preparation including a dedicated cytology cell block system (CCS) (Cellient™, Hologic) and a general histology processing system (HPS)(Tissue-Tek™ VIPS, Sakura).

Materials and Methods: Cytology specimens including fine needle aspirates (FNA) and fluids were collected in either Cytolyt Hologic™ solution or in RPMI. The specimens were centrifuged for 10 min. at 377g and were fixed in formalin/95% alcohol solution for processing with the HPS and were washed twice with Cytolyt Hologic™ and centrifuged at 960g for 5min for the CCS technique. The CCS technique employs a series of alcohol solutions and xylene which are filtered over the specimen for fixation with the final step of encasement in paraffin wax. The HPS technique uses formalin fixation, tea bagging, overnight processing and manual embedding of tissue in paraffin wax. The cell blocks prepared by both techniques were cut at 5 m thickness, stained with hematoxylin and eosin (H&E) method for conventional pathological examination without knowledge of the preparatory method. The cellularity of the specimens was recorded as either less than or more than 25%. The specimens were evaluated semiquantitatively on a scale of 1-3 for tissue preservation, sharpness of cellular details and staining quality. These parameters were compared between the two techniques by Fishers exact test.

Results: We studied 65 specimens (FNA-60, Fluids-5) by the CCS and 52 specimens (FNA-48, Fluids-4) by the HPS techniques. The tissue sections obtained by the CCS and HPS techniques showed <25% cellularity in 54% and 71% and >25% cellularity in 45% and 29%. Tissue preservation was not significantly different between the two techniques (p= 0.425). However, specimens processed by CCS technique were significantly better than HPS for staining quality (p=0.01) and sharpness of cellular details (p=0.0004). Both techniques demonstrated presence of the tissue in the center of the glass slide in majority of the specimens. Specimens processed by the CCS technique alone showed myxoid material in the background of the smear in 26% of the specimens which is related to overzealous usage of freeze spray at the step of dislodging the filter from the paraffin button prior to the introduction into the paraffin mould.

Conclusions:

• 1
  The CCS (Cellient™) technique generated cell blocks that demonstrated significantly better staining quality and sharpness of cellular details of the constituent tissue than the HPS technique.
• 2
  Tissue preservation was not significantly different between the CCS and HPS techniques.
• 3
  The CCS technique alone showed the presence of varying amounts of myxoid material (26%) in the background of the tissue section which is related to overzealous application of freeze spray.
• 4
  The CCS technique allowed preparation of cell blocks individually in ∼45 min while the HPS technique necessitates batching of the specimens and took ∼6 hours for completion.

65

Expanding the Role of Cytotechnologists: A Cost Effective Approach for Anatomic Pathology at Mayo Clinic

Renee Root, BS, CT(ASCP), Andrew Bridgeman, MBA, Angela Sorenson, MA, SCT(ASCP), Sally Nelson, BS, CT(ASCP), Johnita Ihrke, BS, CT(ASCP), Lisa Colborn, MBA, CT(ASCP), Kevin Halling, MD, PhD, Michael Henry, MD, Amy Clayton, MD

Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota

Introduction: The traditional role of the cytotechnologist (CT) typically includes screening gynecologic, non-gynecologic and fine needle aspirate cytology specimens. CTs possess valuable morphologic expertise which can be useful in other areas of the pathology practice. Increased demand for pathologists' time in conjunction with CT expertise has resulted in opportunities to integrate CTs into numerous nontraditional areas of our practice. Little is published about the actual cost savings achieved by these practices.

Materials and Methods: A review of the role of the CT in our lab was performed. A timeline was created to demonstrate how CT roles and test volumes have grown gradually over the years. Time studies were performed to evaluate whether integration of CTs has saved pathologist time. A cost analysis was performed to quantify the costs to the lab if a pathologist performed the work that CTs are performing today.

Results: The role of the CT has expanded dramatically since the mid 1990s when CTs were trained to perform digital image analysis (DIA) of tissue specimens to the most recent addition of pre-screening acid fast bacilli (AFB) stains (Table 1). Test volumes in these areas have grown and gradual integration of the CT into other areas of the pathology practice is demonstrated (Table 2). Time studies show that use of CTs for these activities significantly reduces pathologist time required to review and sign out cases. This translates into significant cost savings to the lab while freeing up additional pathologist FTE available to perform other more complex activities (Table 3).

Conclusions: Expanding the role of the CT into other areas of pathology has many benefits including labor cost savings, increased availability of pathologists for more complex testing, CT and pathologist satisfaction, and creating an optimistic outlook for the future of cytotechnology. In addition to significantly reducing labor costs to the lab, the additional pathologist time can be used for additional revenue generating activities. CTs in our lab perform a wide variety of activities contributing to increased job satisfaction and employee retention. CTs possess many valuable skills including expertise in morphologic criteria, ability to identify areas of tumor in tissue sections, screening skills, attention to detail and training in medical and pathology terminology. This makes them very suitable to expand their presence in the pathology practice and secure an important role for CTs in the future.

Table 1. 

Additional CT Role	Year
DIA	1995
EUS FNA	1998
FISH	2000
FNA Pre-screens	2003
Manual ER/PR/HER2 Quantitation	2005
Circulating Tumor Cells	2007
Tumor Identification for Microdissection	2009
Automated HER2 Quantitation	2009
AFB Stain Pre-screens	2009

Table 2. 

Test Volumes for Additional CT Roles
Year	1995	1996	1997	1998	1999	2000	2001	2002	2003	2004	2005	2006	2007	2008	2009
DIA	553	851	1260	1085	1747	1970	1794	1836	1688	2136	1933	2127	1962	1205	1977
FNAs				3427	3760	5687	5620	5989	5092	4965	4840	5051	5443	7280	7421
FISH						138	809	853	1323	2795	4289	5421	6236	6099	5261
Manual ER/PR/HER2  Quantitation											11254	8978	7085	9191	9079
Circulating Tumor Cells													5	37	669
Tumor Identification															6007
AFB															1041

Table 3. 

Test	Pathologist Workload perTest without Cytotechnologist (Minutes)	% Pathologist Time Saved with Cytotechnologist	Annual Volume (Tests)	Number of PathologistFTEs Saved
DIA	77	98	1977	1.2
FISH	38	96	5261	1.5
ER/PR/HER2	18	86	9079	1.1
Circulating  Tumor Cells	14	71	669	0.1

66

Lean Scheduling: Balancing Work and Streamlining the Cytotechnologist Scheduling Process

Gina Saccomanno, BS, CT(ASCP), Lindsey Kane, BS, CT(ASCP), Erin Amsden, BS, CT(ASCP), Andrew Bridgeman, MBA, Angela Sorenson, MS, SCT, Renee Root, BS, CT(ASCP), Kara Hansing, BS, SCT, Johnita Ihrke, BS, CT(ASCP), Sally Nelson, BS, CT(ASCP), Angela Zieman, BS, CT(ASCP), MB, Kristina del Rosario, BS, CT(ASCP), Delores Tuve, BS, CT(HEW), Michael Henry, MD

Cytopathology, Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota

Introduction: The responsibilities of the Cytotechnologist (CT) are evolving. CTs are being asked to do far more than gynecological specimen screening. CTs are now commonly involved in fine needle aspirations, and trained in a number of molecular tests. With the additional work foci, managing the scheduling of additional rotations becomes more complex, requiring substantial attention. Constructing the calendar using existing methods used nearly 50 hours of management, CT, and administrative assistant time each month. To address the increasing complexity of scheduling rotations a team was assembled to develop an improved scheduling process using LEAN principles.

Materials and Methods: The existing scheduling process was analyzed to identify the steps being used to build the schedule. Each rotation was examined to identify the time, tasks, location, coverage needs, specimen volumes and the percent effort needed to complete tasks. Opportunities for improvement were identified. A number of important metrics were identified in the existing scheduling process, including:

• Keeping the schedule up to date in response to day to day changes, vacation requests, etc.

• Creating equal distribution of rotations among employees.

• Identifying balance between rotational skill competency and foundational cytology skills.

• Reducing the administrative time required to maintain the schedule.

Results: The duties and limits of each rotation were examined. Areas of redundancy were identified as were obvious opportunities to combine rotations. By examining each rotation closely, the required full time equivalent (FTE) required to complete the schedule could be accurately calculated, and compared to the available FTE at any given point on the schedule.

A pilot schedule was devised that would assign staff rotations. A custom spreadsheet was built to manage the creation and display of the rotation schedule. The automation of scheduling rotations and balancing of the rotation numbers done by each employee reduced the total effort to complete construction of the calendar by management, CT, and administrative assistants to 14.5 hours.

In response to the positive outcome of the pilot study, the laboratory introduced a commercial web-based scheduling system (ClairVia®). While the software initially requires the manual entry of data as well as minor changes each scheduling cycle (new vacation requests, etc), once up and running, the software can produce a schedule that balances the rotations equally across the employee pool in as little as 3 hours per month.

Conclusions: The use of a scheduling software package eliminates several steps for management, CT, and administrative assistants. It is no longer necessary to have rotation sign up far in advance to allow time for planning or to seek out employees to fill vacant rotations. Additionally, the software eliminates the need for month's end data analysis of CT time spent on rotations. While CTs may lose some upfront scheduling flexibility (i.e. the ability to pick and choose rotations by day) when using an automated scheduling system, trading rotations and unplanned time away can still be accommodated. Utilizing the LEAN process on rotation scheduling resulted in a significant reduction in the amount of time required to complete the monthly scheduling task.

67

Assessment of Immunohistochemical Stains on the Cellient™ Automated Cell Block System

Sandra Dolar, CT(ASCP), Michael Mawhorter, HS, Peter Rosso, HS, Dawn Underwood, CT(ASCP), Jennifer Brainard, MD, Bin Yang, MD, PhD

Pathology & Laboratory Medicine Institute, Cleveland Clinic, Cleveland, Ohio

Introduction: Adequate cell block preparation plays an important role in facilitating accurate cytologic diagnosis, especially when immunohistochemical stains are needed. It has been shown that the Cellient™ automated cell block system (Hologic, Marlborough, MA) has produced consistent results with a higher yield of target cells compared to the conventional thrombin clot method. However, its adequacy for immunohistochemical stains has not been systematically studied given its methanol-based fixation protocol which differs from the formalin-based protocol validated for most antibodies in immunohistochemistry. We evaluated commonly used antibodies immunohistochemically in cell blocks prepared by the Cellient™ system.

Materials and Methods: Cellient™ cell blocks were prepared from 67 cases of non-gynecologic cytologic specimens, including 24 body fluids (pleural, peritoneal and pericardial effusions) and 43 FNA samples from different body sites. Immunostaining with three fixation protocols were applied: 1) cells fixed in Cytolyte following manufacturer's instruction; 2) cell pellets further fixed with 10% neutral formalin for 30 minutes before cell block preparation; 3) slides made with method 1 followed by immersion in formalin for 1 hour before immunostaining. Thirty-eight antibodies were applied to the study using the Vantana Benchmark XT® instrument. Antibodies include 30 membrane/cytoplasmic-staining antibodies (CD3, CD10, CD20, CD30, CD45, CD68, CD138, kappa, lambda, CK5/6, CK7, CK20, AE1/3, CAM5.2, CDX2, CEA, B72.3, Ber-EP4, PSA, GCDFP, S100, HMB45, melan A, tyrosinase, NSE, synaptophysin, chromagranin, calretinin, thyroglobulin, calcitonin) and 8 nuclear-staining antibodies (p16, p53, p63, TTF-1, WT-1, ER, PR, and Mum-1). Immunohistochemical staining intensity was evaluated using a four-grade system (0, 1+, 2+ and 3+).

Results: All 30 membrane/cytoplasmic antibodies showed 2-3+ staining intensity in Cellient cell blocks using manufacturer's protocol. Most nuclear-staining antibodies (p16, p53, TTF-1, WT-1, ER and PR) failed to stain target cells in Cellient™ cell blocks. Two nuclear-staining antibodies (p63 and Mum-1) showed weak 1+ staining pattern. Additional formalin treatment (fixation method 2 and 3) slightly enhanced the staining intensity for all the membrane/cytoplasmic antibodies and nuclear antibodies p63 and Mum-1. Formalin fixation did not result in a significant change in immunoreactivity for the other nuclear antibodies. The lost immunoreactivity of these nuclear-staining antibodies is independent of specimen types (body fluid vs FNA) and of body sites.

Conclusions: Although the Cellient™ automated cell block system retains the immunoreactivity of membrane/cytoplasmic antibodies, the immunoreactivity of the majority of nuclear-staining antibodies was lost using a methanol fixation protocol. Additional formalin treatment cannot rescue the lost immunoreactivity of most nuclear-staining antibodies. It is imperative to know this technical pitfall when applying nuclear-staining antibodies to prevent false-negative interpretation of immunostaining results.

Lung and Mediastinum 68

Cytologic Examination of Bronchoalveolar Lavage Fluid in Patients with Acute Leukemia

Michelle Collier, MD, Larry Kluskens, MD, PhD, Paolo Gattuso, MD

Pathology, Rush University Medical Center, Chicago, Illinois

Introduction: Bronchoalveolar lavage is often performed in patients with acute leukemia who develop respiratory failure or pulmonary infiltrates. Patients usually undergo BAL to rule out infection. We present a series of 12 cases in which the diagnosis of leukemia was made on the bronchoalveolar lavage material.

Materials and Methods: We retrospectively reviewed all bronchoalveolar lavage samples reviewed by our cytopathology department from 1/1/2006 to 12/31/2008. There were a total of 1130 cases, of which 139 showed malignant cytology (33 adenocarcinoma, 15 squamous cell carcinoma, 10 leukemia, 10 neuroendocrine, 57 non-small cell type, and 2 lymphoma) and 71 were suspicious for carcinoma. Sixteen samples were unsatisfactory and 904 were benign of which 33 had identifiable microorganisms (17 pneumocystis, 8 Candida, 3 aspergillus, 2 actinomyces, 1 blastomyces, 1 histoplasma, 1 herpes).

In addition to the 10 leukemia cases identified, 2 more cases were reviewed after the search criteria.

Results: The demographics for the 12 patients included 7 men (age range 22-75, mean 39 years) and 5 women (age range 44-73, mean 63.2 years). Ten of the 12 patients had bone marrow biopsy proven acute leukemia reviewed at our institution (2 acute myelomonocytic leukemia, 2 acute promyelocytic leukemia, 2 AML with inv16, 1 therapy related AML, 1 acute monocytic leukemia, 1 chronic myeloid leukemia in blast crisis, 1 AML with maturation (M2)). The time range from diagnosis of acute leukemia to positive BAL results was -1 to 8 days, with a mean of 2 days. For the other two patients, one was diagnosed with large granular lymphocyte leukemia in a splenectomy, had a positive BAL 233 days later during relapse. The other had MDS with excess blasts, started treatment for AML, and had a positive BAL about 90 days later during relapse. Four of the patients had a preceding myelodysplastic syndrome (MDS). Interestingly, 2 of the 12 patients also had a biopsy proven leukemic infiltrate of solid organs at the time of presentation, one in the pancreas and another in the skin.

Most of the patients had been started on chemotherapy or other treatment at the time of the BAL. The patient symptoms that prompted the BAL included: (8 with shortness of breath/respiratory distress/hypoxia, 3 with fever, 2 with chest pain and 2 with cough). All had chest x-rays at the time of procedure which all showed opacities or consolidation (6 with bilateral findings and 6 with localized findings), mimicking an inflammatory process.

The outcomes of the patients varied (7 patients are deceased, 3 are alive and in remission, and 2 were lost to follow-up). Of the 7 patients who have died, 3 died during their initial hospital course (time range from time of BAL to death was 4, 6, and 10 days). The other 4 patients time range from BAL to death was 1.5, 6, 7, and 40 months, respectively.

Conclusions: Bronchoalveolar lavage is a relatively safe procedure that is commonly performed in our institution. The presence of a leukemic lung infiltrate can mimic infection and bronchoalveolar lavage is a useful diagnostic tool in this setting. The prognosis of patients with lung involvement by acute leukemia is poor, as our study shows.

69

Specific Objective Histologic Classification in Non Small Cell Lung Cancer (NSCLC) Cytology Specimens

Anastasios Dimou, MD¹, Valsamo Anagnostou, MD¹, Megan Selbst, MD¹, Konstantinos Syrigos, MD, PhD², David Rimm, MD, PhD¹, Malini Harigopal, MD¹

¹Pathology, Yale University School of Medicine, New Haven, Connecticut; ²Internal Medicine, Section of Medical Oncology, Yale Cancer Center, Yale University School of Medicine, New Haven, Connecticut

Introduction: Histology has become critical for optimal therapeutic decisions in patients with NSCLC. Minimally invasive techniques and cytology specimens in particular, provide scant material for classification of NSCLC into squamous cell carcinoma or adenocarcinoma, the two major histology types. A histology prediction model based on quantification of expression of four biomarkers, TTF1, CK13, CK5 and EGFR was developed in a cohort of patients with NSCLC organized in tissue microarray (TMA) format and optimized to predict histology with high sensitivity and specificity (>95%). Subsequently, it was validated in two independent cohorts of patients. In this study, we assess this method on cytology cell block specimens with a diagnosis of NSCLC.

Materials and Methods: Quantification of TTF1, CK13, CK5 and EGFR using the AQUA method of quantitative immunofluorescence was used as described previously. Scores for each case were normalized for run to run variability by using an index TMA and entered into a logistic regression model for prediction of adenocarcinoma histology.

Results: Initial retrospective collection included 20 cases that met the inclusion criteria. Ten cases were excluded as having no tumor in the H&E cell block slide. An additional case was excluded as there were no tumor cells in the slide stained with one of the markers. The remaining 9 cases consisted of 6 squamous cell carcinomas and 3 adenocarcinomas. Histology was known from the cytology specimens themselves in 3 cases; in the remaining 6 cases the final adjudicated diagnosis was obtained from a subsequent procedure. The model was able to predict histology correctly in 8/9 cases. In the misclassified case, a prediction of squamous histology was made. The pathologist that reviewed the cytology specimen could not classify the case in terms of histology and a decision of adenocarcinoma was made, based on an endobronchial biopsy.

Conclusions: Cytology specimens that are diagnostic for NSCLC are often insufficient to tell the difference between adenocarcinoma and squamous cell carcinoma. Our histology prediction model was able to classify correctly the majority of cases. We are currently validating these preliminary results in a larger cohort of cytology specimens from patients with NSCLC, including a prospective analysis.

70

The Accuracy of Fine Needle Aspiration and Immunocytochemical Stains (TTF-1, P63, CK5/6 and MUC-1) in Distinguishing Adenocarcinoma from Squamous Cell Carcinoma of the Lung

J. Akinfolarin, MD¹, T. Giorgadze, MD, PhD¹, S. Bandyopadhyay, MD¹, F. Lonardo, MD¹, D. Shi, MD¹, S. Sethi, MD², M. Husain, MD¹, J. Feng, MD, PhD¹

¹Pathology, Wayne State University, Detroit, Michigan; ²Pathology, Karmanos Cancer Institute, Detroit, Michigan

Introduction: Fine Needle Aspiration (FNA) is increasingly used for both the diagnosis of lung carcinoma and the classification of the specific tumor type. The diagnosis of non-small cell carcinoma (NSCLC) on FNA specimens is no longer sufficient due to the revealed differences in treatment and side effects of nonsquamous versus squamous cell carcinoma (SCC) by recent clinical trials. Also, EGFR dependent therapy for adenocarcinoma (ADC) makes it imperative to identify patients with this subtype of carcinoma. The purpose of this study is to determine the extent to which combined cytomorphology and immunocytochemistry can distinguish ADC from SCC of lung using follow-up surgical pathology specimen as gold standard.

Materials and Methods: We identified 45 consecutive patients who had FNA diagnosis of ADC, SCC and NSCLC of the lung and also had follow up surgical pathology from January 2007 to July 2009 at our institution. Small cell carcinomas and metastastic carcinomas to the lung were excluded in this study. We correlated the FNA results with follow-up histologic diagnosis which was defined as the gold standard. Immunocytochemical stains for TTF-1, P63, CK5/6 and MUC-1 were performed on 38 cases that had adequate material on cell blocks. Nuclear staining for TTF-1 and P63, and membranous and cytoplasmic staining for CK5/6 and MUC-1, in at least two clusters of malignant cells are considered positive. All the stains were interpreted independently by 2 cytopathologists and tabulated according to follow-up histologic diagnosis as SCC, ADC and NSCLC.

Results: On surgical specimens, 20/45 (44%) were diagnosed as ADC, 14/45 (31%) were diagnosed as SCC and 11/45 (24%) were diagnosed as NSCLC. 34/45 (76%) of the FNA cases were correctly classified using only cytomorphologic features: 12/20 (60%) of ADC, 11/14 (79%) of SCC and 11/11 (100%) of NSCLC. The results of the immunocytochemistry for TTF-1, P63, CK5/6 and MUC-1 are shown in Table 1 below.

Conclusions: Classification of lung cancer based on cytomorphologic features is fairly accurate. The category of NSCLC can by further subclassified with the help of the immunocytochemical panel of TTF-1, P63 and CK5/6. Dual negativity for P63 and CK5/6 with TTF-1 expression was most specific for ADC while dual positivity for P63 and CK5/6 with negative TTF-1 expression highly favored SCC. MUC-1 expression was not significant in differentiating ADC from SCC.

Table 1. 

Histologic Diagnosis (n= 38)	TTF-1	P63	CK5/6	MUC1	TTF-1-/ P63+/ CK5/6+	TTF-1+/ P63-/ CK5/6-
ADC (n=18)	14 (78%)	2 (11%)	2 (11%)	14 (78%)	1 (6%)	12 (67%)
SCC (n=13)	1 (8%)	11 (85%)	10 (77%)	7 (54%)	8 (61%)	0
NSCLC (n=7)	1 (14%)	5 (71%)	5 (71%)	6 (86%)	4 (57%)	0

71

Endobronchial Ultrasound-Guided Transbronchial Needle Aspiration Cytology—the Memorial Sloan-Kettering Cancer Center Experience

Maria Friedlander, MPA, CT(ASCP), Suzanne Brandt, MD, Carlie Sigel, MD, David Finley, MD, Nabil Rizk, MD, Andre Moreira, MD, PhD, Maureen Zakowski, MD, Natasha Rekhtman, MD, PhD

Cytology Service, Department of Pathology, Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, New York

Introduction: Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is becoming a modality of choice for sampling of mediastinal and hilar lymph nodes. The objective of this study was to review our institution's experience with EBUS-TBNA, with a focus on sensitivity/specificity, tumor typing accuracy, and suitability for EGFR molecular testing.

Materials and Methods: A retrospective review of all EBUS-TBNAs performed during a 6-month period (Jan - Jun 2009) was conducted, which included a total of 278 cytology cases representing 148 patients (71 male:77 female; age range 26-86; average age 64 years). Cytologic results were correlated with histologic follow-up on surgical specimens.

Results: Of the 278 cases, 121 (44%) were classified as malignant or suspicious, 1 (0.4%) atypical, 106 (38%) negative for malignant cells, and 50 (18%) non-diagnostic/unsatisfactory. Malignant cases were sub-typed as adenocarcinoma (n=54), squamous cell carcinoma (n=13), small cell/neuroendocrine carcinoma (n=10), poorly-differentiated non-small carcinoma (n=4), poorly-differentiated carcinoma (n=9), mesothelioma (n=1), extra-pulmonary malignancy (n=15), lymphoma (n=6), and malignant neoplasm, cannot classify further (n=9). Of 106 negative cases, 8 (8%) were diagnosed as granulomatous inflammation, and the rest (92%) as benign lymph nodes. Fifty cases (18%) had same-site histologic follow-up (Table I), which showed that negative EBUS-TBNA was 100% predictive of benign histology. There were no false-positive cytologic diagnoses: 14 positive EBUS-TBNAs were confirmed histologically, and 2 positive EBUS-TBNA with negative surgical biopsy were confirmed as biopsy sampling errors. Non-diagnostic/unsatisfactory EBUS-TBNA was more likely to have a benign (87%) than malignant (13%) histologic follow-up. Overall, EBUS-TBNA procedure had sensitivity for malignant disease of 89% (equivalent to surgical biopsy in this study) and specificity of 100%. Thirty malignant cases had concurrent or subsequent histologic follow-up (at same or different site) revealing no major discrepancies in tumor subtypes rendered by EBUS-TBNA cytology and histology. Molecular testing was performed on 8 cases, and identified EGFR mutation in 2 patients.

Conclusions: This is one of the largest reviews of EBUS-TBNA in modern cytology practice. We find that this procedure has overall sensitivity and specificity for malignancy that is equivalent to surgical biopsy. In addition, these specimens permit high accuracy of tumor subtyping, and are suitable for EGFR mutational analysis. Combined with increased patient comfort, lower morbidity and possible greater cost-effectiveness compared to surgical approaches, our findings lend strong support to EBUS-TBNA becoming the method of choice for assessment of mediastinal and hilar lymph nodes. This is the first report of EGFR testing on EBUS-TBNA specimens in routine clinical practice, which is of great importance for guiding EGFR-targeted therapy in advanced-stage patients for whom EBUS-TBNA may be the only available diagnostic modality.

Table 1. Summary of Same-Site Cytologic-Histologic Correlation

	Histology Positive	Histology Negative
Cytology Positive/  Suspicious (n=16)	14	2 (surgical biopsy  sampling errors)
Cytology Negative (n=19)	0	19
Cytology Non-Diagnostic/  Unsatisfactory (n=15)	2 (EBUS-TBNA  sampling errors)	13

72

Diesel Exhaust Particle Exposure Influences Bleomycin-Induced Lung Fibrosis in Mice

Ying-Ji Li, PhD¹, Takako Shimizu, B.Sc¹, Yukiyo Hirata, PhD¹, Hirofumi Inagaki, PhD¹, Yusuke Shinkai, PhD², Ken Takeda, PhD², Tomoyuki Kawada, PhD¹, Arata Azuma, PhD³, Hajime Takizawa, PhD⁴, Shoji Kudoh, PhD⁵

¹Hygiene and Public Health, Nippon Medical School, Tokyo, Japan; ²Hygiene Chemistry, Faculty of Pharmaceutical Sciences, Tokyo University of Sciences, Noda-shi, Japan; ³Pulmonary Medicine/Infection and Oncology, Nippon Medical School, Tokyo, Japan; ⁴Fourth Department of Internal Medicine, Teikyo University, Kawasaki, Japan; ⁵Fukujuji Hospital, Anti-Tuberculosis Association, Tokyo, Japan

Introduction: Epidemiological studies suggest that particulate air pollutants, including diesel exhaust particles (DEP), may play a role in the recent increase of respiratory morbidity and mortality. The experimental studies showed that DEP induced oxidative stress plays an important role in proinflammatory reaction on airway. Recent studies suggest that N-acetylcysteine improve the idiopathic pulmonary fibrosis. In the present study, we studied the effect of DEP on an experimental model of bleomycin (BLM)-induced lung injury and subsequent fibrosis in mice.

Materials and Methods: BLM was administered IV to C57BL/6J mice on day zero. Mice were exposed to 1mg/m3 DEP for 8 hrs/day and 5 days/week. We designed five experimental groups as follows: group 1, control; group 2, BLM alone; group 3, BLM plus pre-4wks-DEP exposure; group 4, BLM plus syn-DEP exposure; group 5, BLM plus post-1wk-DEP exposure. Cell populations in bronchoalveolar lavage (BAL) fluid were examined at 1 week after BLM injection in groups 1, 2, 3 and 4. At 4 weeks after bleomycin injection, and last DEP exposure, fibrotic foci were histologically observed by Ashcroft score in all the groups.

Results: In the DEP pre-exposed group, there were DEP-laden alveolar macrophages in the BAL fluid; macrophages number was increased and neutrophil number was decreased in the BAL fluid compared with BLM alone group. In the DEP syn-exposed group, neutrophil number in the BAL fluid was increased. Bleomycin-induced pulmonary fibrosis observed by Ashcroft score was increased in all the DEP exposed groups compared with BLM alone group.

Conclusions: These findings suggest that DEP is an important risk factor in the generation and development of BLM induced lung fibrosis in mice.

73

Fine Needle Aspiration (FNA) Cytology of Lung Carcinoids: A Retrospective Study of 63 Cases with Histological Correlation

Lisa Stoll, MD, MPH, Michael Johnson, MD, PhD, Frances Burroughs, SCT(ASCP), Qing Li, MD, PhD

Pathology, The Johns Hopkins Hospital, Baltimore, Maryland

Introduction: Bronchopulmonary neuroendocrine (NE) neoplasms are a spectrum of tumors separated into 4 categories by WHO classification: low-grade typical carcinoid (TC), intermediate-grade atypical carcinoid (ACT), and high-grade small-cell lung carcinoma (SCLC) and large-cell neuroendocrine carcinoma (LCNEC). Given the overlapping features within these tumors, misclassification is a known diagnostic pitfall with significant treatment and prognostic consequences. The morphologic distinction between high and low-grade NE neoplasms is well recognized and usually is not a source of significant difficulty in daily practice. However, some differentiating criteria may not be appreciated in cytology specimens since small amounts of material or tissue fragments are commonly obtained at the time of FNA. In this study, we have correlated the FNA cytological features of TC with the corresponding surgical specimen. The purpose of our study is to critically review the diagnostic features of TC and discuss the potential diagnostic pitfalls, particularly in those cases that have been misclassified on FNA cytology.

Materials and Methods: The archives of the Division of Cytopathology and the Department of Surgical Pathology at The Johns Hopkins Hospital were searched for lung carcinoid tumors from 1990 to 2010. A total of 390 cases were found with 63 cases having both FNA cytology and corresponding surgical specimen. Cytology specimens consisted of 20 lung ultrasound/CT-guided FNAs, 24 transbronchial needle aspirations (TBNA), 14 lung brushings and/or washings, 4 mediastinal mass FNAs, and one bronchioloalveolar lavage (BAL). The cytological and corresponding surgical material were reviewed, and correlated with immunohistochemistry (IHC) and clinical information.

Results: The patients in our study ranged in age from 8 to 81 years old with a median age of 55 years. The male to female ratio was 1:2.9. Nineteen of 63 patients (30%) were smokers. Among 63 paired cases, 32 cases (51%) showed concordant cytological and histological diagnoses of TC; whereas 31 cases (49%) showed discordant diagnoses. Table 1 summarizes the cytological and histological diagnoses of these 31 cases.

Conclusions: The important morphologic factors for distinguishing low-grade NE neoplasm TC from ACT, high-grade SCLC or non-small cell carcinoma remain an accurate evaluation of the nuclear features, chromatin patterns and particularly, the assessment of nucleoli. The presence of nucleoli or dot-like eosinophilic or basophilic foci within the nucleus is the feature most suggestive of TC. However, nuclear molding and crowding are not discernible features, since they may be found on smears with increased cellularity or in poorly differentiated carcinomas. Crush artifacts can occur in both low- and high-grade NE neoplasms and may cause a misinterpretation of SCLC. Other artifacts due to delayed fixation or poor processing are causes of incorrect interpretations in some cases. IHC stain of Ki67 may be high and useful in some of the cases.

Table 1. Cytological cases with discordant histological diagnoses

Cyto diagnoses	Histo diagnoses
Reactive bronchial cells (n=6)	TC (n=6)
Non-diagnostic (n=4)	TC (n=4)
Atypical cells (n=2)	TC (n=2)
Favor TC (n=3)	PD ca with basaloid features (n=1)
	PDADC (n=1)
	Reserve cell hyperplasia (n=1)
Low-grade NE (n=1)	Low-grade sarcoma (n=1)
ACT or ACT vs TC (n=4)	TC (n=4)
SCLC or SCLC vs TC (n=5)	TC (n=5)
ADC vs TC (n=1)	TC (n=1)
PD ca with NE features (n=5)	TC (n=5)

74

Improvement of Cellular Yield by Using So-called “Blood Clot” Method in Cell Block Preparation during Endobronchial Ultrasound-Guided Transbronchial Needle Aspiration (EBUS-TBNA)

Qing K Li, MD, PhD¹, Susan Otell, CT(ASCP)¹, John Kirby, CT(ASCP)¹, Frances Burroughs, SCT(ASCP)¹, Peter Illei, MD¹, David Feller-Kopman, MD², Dorothy Rosenthal, MD¹, Douglas Clark, MD¹, Rex Yung, MD²

¹Pathology, The Johns Hopkins Hospital, Baltimore, Maryland; ²Pulmonary and Critical Care Medicine, The Johns Hopkins Hospital, Baltimore, Maryland

Introduction: In lung cancers, our study and others have shown that identification of biomarkers has changed the diagnostic and therapeutic paradigm from primarily morphological assessment to morphological assessment plus molecular analysis of a patient's tumor. EBUS-TBNA is a minimally invasive procedure which is increasingly used for obtaining diagnostic material from lung cancer patients. Several types of cytological specimens can be obtained by the procedure. Among them, cell block (CB) preparation from the aspirate needle plays an important role in the recovery of cellular material which can be used for molecular study, as clinical trials and treatment options not only depend upon morphological assessment but also on molecular characteristics of the tumor. However, the efficiency of the recovery of cellular material by conventional needle rinse method is low. In this study, we routinely apply a novel “blood clot” technique to recover cellular material from aspirate needle for CB preparation. We have compared the efficiency and diagnostic yield of this technique with conventional aspiration needle rinse method in CB preparation for diagnosing and staging lung cancer.

Materials and Methods: A total of 84 consecutive patients who underwent EBUS-TBNA for the evaluation of suspected thoracic malignancies with nodal involvement were included in our study over a period of 12 months during 2009 at The Johns Hopkins Hospital. All cases were performed with on-site evaluation by cytopathology personnel; and aspirations were performed by two experienced bronchoscopists using a 22 gauge aspirate needle. To prepare CB using this novel ‘blood clot’ technique (CB-BC), the cellular material within the aspirate needle was allowed to form a blood clot, then gently pushed out using the wire trocar stylet onto filter paper. Subsequently, the material was fixed and processed for CB preparation as routine in the cytopathology laboratory. In addition, 28 consecutive patients, in whom CBs were prepared by conventional aspiration needle rinse method (CB-NR), were also included as controls. The size and location of lymph nodes were also correlated.

Results: Among 84 patients, 107 lymph nodes and 13 lung lesions were sampled. In lymph node specimens, 95 of 107 CB-BCs (88.8%) yielded sufficient diagnostic material, whereas 11 of 13 lung CB-BCs (84.6%) yielded sufficient diagnostic material. In addition, ancillary studies were performed in 40 cases. 25 of 26 immunohistochemistry (IHC) studies and 14 of 14 EGFR/Kras mutational analyses had sufficient material. However, in the control CB-NR group, 22 of 39 (56.4%) lymph nodes and zero of 3 (0%) lung specimens provided sufficient diagnostic material. The average size of lymph nodes in the study group (1.76 cm) did not show significant difference than the control group (1.82 cm, p>0.05).

Conclusions: Our data show that the overall non-diagnostic rate of CB-BC in lymph node specimens is 11.2% vs 43.6% of conventional CB-NR (p<0.05). The non-diagnostic rate of lung specimens is 15.4% of CB-BC vs 100% of CB-NR. The CB-BC method markedly increases the cellular yield of CB preparation without compromising cytomorphological characterization of tumor cells. The technique is cost-effective and reproducible. Most importantly, it is able to provide sufficient material for a variety of ancillary studies, including IHC and molecular analysis of tumor cells.

75

Cytological Evaluation of Mediastinal Lesions in the Pediatric Population

Maximo Llaudes, MD, Morris Edelman, MD, Patricia Wasserman, MD, Nora Morgenstern, MD

Cytopathology, Long Island Jewish Medical Center, New Hyde Park, New York

Introduction: Lesions of the mediastinum in the pediatric population are represented by a variety of conditions that are responsive to different diagnostic and treatment modalities. The diversity of these lesions ranges from congenital to neoplastic masses that can be further classified according to age groups. The management and diagnostic challenges of these lesions can be simplified considerably if the location within the mediastinum and age of the child are considered. Traditionally, mediastinal masses in children have required open thoracotomy, or sternotomy. With the advent of minimally invasive methods, the principles of management of masses in the mediastinum have been revised to include fine needle aspiration biopsy (FNAB) and tomography-guided transthoracic core biopsy. These techniques are well established in the adult population, and are now increasingly being applied in the management of childhood mediastinal lesions reducing the cost and the trauma associated with a more invasive procedure, therefore allowing for rapid diagnosis and limiting the duration of hospital stay. The objective of this study was to cytomorphologically evaluate mediastinal lesions in the pediatric sector obtained through different surgical and radiological methods.

Materials and Methods: A 15-year retrospective study of masses of the mediastinum was performed between the years 1995-2009. The patient population ranged between 0-19 years of age. There were 29 cases obtained through radiologically guided percutaneous FNAB, and 46 through thoracoscopy or thoracotomy procedures. Immediate adequacy was assessed using a Romanowsky stain (DifQuik), and additional slides were prepared using the Papanicolaou stain. The cell block material was used for Hematoxylin & Eosin (H&E) slides, and immunohistochemistry. Flow cytometry and molecular cytogenetics (FISH on air dried smears or cytogenetics on fresh specimen in RPMI medium) were used in the diagnosis of lymphoproliferative disorders and small round blue cell tumors.

Results: There were 32 cases with a benign diagnosis (43%), 39 malignant cases (52%), and 4 non-diagnostic cases (5%) (Table 1). Most (62%) of the malignant cases were lymphoproliferative lesions (13 Hodgkin's Lymphoma (Figure 2), 11 Non-Hodgkin's lymphoma), with the rest of the malignant lesions including neuroblastoma (7), immature teratoma (2), Ewing sarcoma (1), rhabdomyosarcoma (2), Langerhans cell histiocytosis (2), and one case of metastatic papillary thyroid carcinoma to mediastinum). The distribution of mediastinal lesions according to ages is shown in Figure 1. Overall, cytological adequacy evaluation was performed in 49 cases (64%) with 96% agreement between the cytologic and subsequent surgical specimen. The benign lesions (43%) were comprised of cystic congenital lesions and reactive processes. All the FNAB procedures had cytological adequacies performed, followed by diagnostic core biopsy with 90% agreement between cytologic diagnosis and diagnosis in the core biopsy.

Conclusions: Awareness of cytomorphologic features at the time of adequacy evaluation and triage of the fresh FNA and core biopsy specimen results in accurate final cytological diagnoses of mediastinal lesions without the need for more invasive procedures. FNA and core biopsy are shown to be complementary techniques when used together as a single procedure for an integrated diagnosis. See Table 2.

Figure 1. Distribution of mediastinal lesions according to ages.

Figure 2. Reed Sternberg Cell (Dif-Quik x 400).

Table 1. Histopathologic diagnosis of pediatric mediastinal masses

Diagnosis	No.
Non-neoplastic	25 (34%)
Inflammatory/reactive	8
Bronchogenic cyst	4
Pulmonary sequestration	2
Reactive lymph node	5
Lymph node with transformation	1
Esophageal duplication	1
Cystic thymus	2
Vascular malformation	1
Castleman's disease	1
Neoplastic-benign	9 (12%)
Ganglioneuroma	5
Mature teratoma	2
Lymphangioma	1
Schwannoma	1
Neoplastic-malignant	39 (54%)
Neuroblastoma	6
Hodgkin lymphoma	17
Non-Hodgkin lymphoma	9
LCH	2
Ewing's sarcoma	1
Immature teratoma	2
ERMS	1
Metastatic	1

Table 2. Correlation between cytologic diagnosis obtained through different methods, and histologic diagnosis.

Cytologic Diagnosis
	Intraoperative			FNA			CRP
	TP	FN	Nd	TP	FN	Nd	TP	FN	Nd	Total
TP: true positive; FN: false negative; Nd: non-diagnostic; CRP: Core-Roll Procedure.
Histologic diagnosis
TP	15	1	2	2	0	0	14	0	2	36
FN	0	0	0	0	0	0	0	0	0	0
Nd	0	0	0	0	0	1	0	0	1	2
Total	15	1	2	2	0	1	14	0	3	38

76

CD56/Ki-67 a More Helpful Immunostain Panel than Synaptophysin/Chromogranin to Distinguish Small Cell Carcinoma from Non-small Cell Carcinoma

Zahra Maleki, MD

Pathology, Johns Hopkins University, Baltimore, Maryland

Introduction: The basic approach in the cytological categorization of lung cancers is to distinguish small cell carcinoma (SMCC) from non-small cell carcinoma (NSMCC). Distinction between these two entities is clinically essential, because prognosis and treatment are substantially different for these cancer types. Fine needle aspiration (FNA) is considered highly accurate and is a first-line diagnostic approach to work-up pulmonary masses and furnishes vital information for therapeutic management. Currently, synaptophysin and chromogranin are routinely used to distinguish SMCC from NSMCC. However, these neuroendocrine markers may have variable expression in SMCCs and cannot distinguish atypical carcinoid from SMCC. Lack of expression of these two markers in SMCCs is not uncommon and neuroendocrine expression in carcinoids and some NSMCCs including squamous cell carcinomas is well documented. Herein, we found that a panel of CD56 and Ki-67 can be more helpful to differentiate SMCC from NSMCC and atypical carcinoid.

Materials and Methods:40 cases of SMCC of a lung primary were retrospectively retrieved from the cytopathology archives of a large tertiary care center. The cytologic slides, consisting of both conventional smears and liquid-based preparation and immunostains were reviewed. Patients' medical records were reviewed.

Results: There were 22 females and 18 males (F:M= 1.2) and their ages ranged from 45 to 81 years (mean=64.1 years). There were 20 lung cases and 20 metastatic lesions of a lung primary (Table 1). The specimens were obtained by FNA (n=35) or thoracentesis (n=5). CD56 was strongly expressed in all 40/40 cases and Ki-67 showed high proliferation rate (>70%) in all 10/10 cases. Both synaptophysin and chromogranin were available for 27 cases (Table-2). Focal to strong positivity was noted in positive cases of synaptophysin and chromogranin.

Conclusions: A panel of synaptophysin and chromogranin is commonly used as neuroendocrine markers to distinguish SMCC from NSMCC. Our study suggests that a panel of CD56/ Ki-67 is superior to synaptophysin/chromogranin to differentiate SMCC form NSMCC and atypical carcinoid.

Table 1. Anatomic site and the procedure type

Anatomic Site	Procedure
Lung (n=20)	FNA (n=20)
Pleural fluid (n=5) Lymph node  (n=5) Liver (n=4) Soft tissue (n=3) Bone (n=1) Pancreas (n=1) Adrenal (n=1)	FNA (n=15) Thoracentesis (n=5)

Table 2. Synaptophysin, Chromogranin expression in SMCC

Immunostain	Chromogranin (negative)	Chromogranin (positive)
Synaptophysin (negative)	N=3	N=1
Synaptophysin (positive)	N=10	N=13

77

Reducing False Negative Results through Proposed Quantitative Criteria for Determining Lymph Node Sampling Adequacy in Endobronchial Ultrasound Guided Transbronchial Fine Needle Aspiration (EBUS-TBNA)

Carrie Marshall, MD¹, John Stewart, MD PhD²

¹Department of Pathology, University of Colorado Denver Health Science Center, Aurora, Colorado; ²Department of Pathology, University of Texas MD Anderson Cancer Center, Houston, Texas

Introduction: Endobronchial ultrasound guided transbronchial fine needle aspiration (EBUS-TBNA) is a well established and cost-effective method to stage patients with thoracic cancers. EBUS-TBNA is considered not as accurate as mediastinoscopy due to an increased number of false negative diagnoses. Previous publications have reported the negative predictive value of EBUS-TBNA to range between 76% and 86%. Can the negative predictive value of EBUS-TBNA be improved by establishing criteria to determine what constitutes an adequate lymph node sample?

Materials and Methods: The pathology database was searched for patients who had EBUS-TBNA for pre-operative staging for a thoracic malignancy between 10/01/2006 and 12/31/2009. The searches were limited to procedures performed by a single surgeon who is well-experienced in EBUS-TBNA. Only patients who had subsequent surgical resection without intervening treatment or therapy were included in the study. Each cytology slide was reviewed by a single pathologist and given a numerical lymphoid score (Table 1). The scores of the Pap stained aspirate smears were summed for a total case score. Based on the surgical pathology of the resected lymph nodes, or clinical follow-up of non-resected nodes, the cases were categorized as true negatives or false negatives. The case score was compared to the categorization of the case in order to determine if the false negative cases had a lower score than the true negative cases.

Results: Of 315 EBUS-TBNAs performed during the study period, 42 patients (totaling 118 lymph node samples) had subsequent surgical resection without preoperative chemotherapy or radiation therapy. The underlying diagnosis was either mesothelioma (12/42 cases, 28%) or non-small cell lung carcinoma (30/42 cases, 78%). 5 cases were false negatives and 92 were true negatives (4 cases were true positives, and 17 were non-diagnostic). The mean lymphoid score for true negatives was 45 (range 4-169). The lymphoid scores for the false negatives were 11, 13, 79 and 130. The resected nodes of the false negative cases with scores of 79 and 130 did not show evidence of aspiration sampling.

Conclusions: By using a quantitative scoring system to determine the amount of lymphoid tissue sampled in EBUS-TBNA, the amount of false negatives could be reduced. Cases with a lymphoid score below 5 should be called non-diagnostic rather than negative. Cases with a lymphoid score between 5 and 20 should be qualified with a comment that they contain scant lymphoid tissue and are at a higher risk to be falsely negative. With immediate assessment, these results can be communicated to the bronchoscopist at the time of procedure and additional passes can be obtained. Specimens with abundant lymphoid tissue may still be false negatives due to sampling error (i.e., partial nodal involvement by metastasis or sampling a negative node when adjacent nodes are positive), but by following adequacy criteria the risk of an interpretive error (i.e., overlooking rare tumor cells in a hypocellular specimen) could be reduced. A future goal would be to have acceptable interobserver agreement on what is an adequate lymph node sample on EBUS-TBNA (perhaps with the help of reference images).

Table 1. 

Lymphoid Scoring Criteria
For each low power (4x) field of view
0 points	• Lymphocyte density of a peripheral blood smear	• Bloody specimen with >50% neutrophils
1 point	• Lymphocyte density of a buffy coat blood smear	• Cluster of 10 lymphocytes or greater
2 points	• Cluster of 200 lymphocytes or greater	• High density of single lymphocytes <1/3 FOV
3 points	• High density of single lymphocytes >1/3 FOV
-1 point	• Obscuring blood clot

78

EGFR and KRAS Mutational Analyses in Metastatic Lung Adenocarcinoma

Delicia Munfus-McCray, MD, PhD, Frederic Askin, MD, Douglas Clark, MD, Christina Adams, CT(ASCP), Edward Gabrielson, MD, Qing Kay Li, MD, PhD

Pathology, Johns Hopkins Medical Institutions, Baltimore, Maryland

Introduction: In primary non-small cell lung cancers (NSCLC), particularly adenocarcinoma, EGFR and K-RAS mutations are found in approximately 10-20% and 20%, respectively; and they have been shown to be mutually exclusive. Patients with EGFR mutations have a 70-80% response rate and longer progression free survival associated with chemotherapy and tyrosine kinase inhibitor (TKI) therapy. KRAS mutations are associated with virtually no response to TKI and occur in patients with significant tobacco exposure. In this study, we have investigated the EGFR and KRAS mutational status in metastatic adenocarcinoma of lung.

Materials and Methods: Using Johns Hopkins Hospital archives, 1,729 cases of lung adenocarcinoma diagnosed either by cytological fine needle aspiration (FNA) or surgical material, were identified between 2007 and 2010. A total of 40 cases of metastatic adenocarcinoma associated with both EGFR and KRAS molecular studies were identified. Cases consisted of 22 surgical and 18 cytological specimens. In addition, 14 pairs of primary and metastatic lung adenocarcinoma lesions were also included. DNA was isolated from the samples, quantified and amplified by polymerase chain reaction (PCR) using primers to exons 18-21 of the EGFR gene and exon 2 of the KRAS gene. EGFR PCR products were analyzed by Genzyme, using bi-directional direct DNA sequencing, capillary gel electrophoresis and fluorescence detection. Mutations in codons 12 and 13 of KRAS were detected by melting curve analysis using FRET technology on the ROCHE Lightcycler. Each sample was analyzed in duplicate, with comparison of melting curves from known positive and negative specimens.

Results: Patient ages ranged from 43 to 82 years old with a median of 60 years old, 50% were Caucasian and the male:female ratio was 1:0.83. 60% (24/40) patients were current or former smokers. The metastatic sites in descending order were: lymph nodes (57.5%), brain (15%), pleura (12.5%), liver (2.5%) and others (12.5%). EGFR mutations were found in 6 of 40 cases (15%) and KRAS mutations in 11 of 40 cases (27.5%). Out of the 6 cases with EGFR mutations, 5 (83%) patients were non-smokers and 1 (17%) patient had a 20 pack year smoking history. EGFR specific mutations did not reveal a significant gender, or presentation stage correlation. However, KRAS mutations occurred exclusively in smokers of 20 pack years or greater. Of the KRAS mutation cases, 8 of the 11 (73%) presented as stage IV cancer and 3 (27%) cases presented as stage I or II. No gender correlation was noted in KRAS mutation analysis. Of the 14 paired primary and metastatic lung adenocarcinoma lesions, one case displayed acquisition of KRAS mutation in the metastatic tumor.

Conclusions: KRAS mutation is more commonly associated with advanced stage of initial cancer presentation and correlates with cumulative smoking dosage in contrast to EGFR mutation. The observation that KRAS mutation can be discordant between primary and metastatic lesions, suggests that other cellular signaling pathways may be involved and need to be investigated in a larger scale study.

79

Influence of Histologic Type of Pulmonary Non-small Cell Carcinoma on the Yield of Bronchoscopic and Transthoracic Pre-resection Diagnostic Procedures

Naobumi Tochigi, MD, PhD, N. Paul Ohori, MD

Pathology, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania

Introduction: With recent advances in surgical and multimodality therapy, the demand for the accurate cytology or biopsy diagnosis of pulmonary non-small carcinoma has increased. However, a number of complex issues exist and the diagnostic yield may depend on a variety of factors. In this study, we investigated the influence of the histologic type of non-small cell carcinoma on the diagnostic yield of bronchoscopic and transthoracic procedures.

Materials and Methods: The pathology files were searched for surgical pathology cases of resected primary pulmonary non-small cell carcinomas with corresponding pre-resection cytology and/or biopsy material for the period, January 1999 to December 2009. The resected cases were categorized as adenocarcinoma (AD), squamous cell carcinoma (SQ), adenosquamous carcinoma (AS), large cell carcinoma (LCC), large cell neuroendocrine carcinoma (LCNEC) or carcinoid tumor (CT). The preceding cytology and/or biopsy procedures were categorized as bronchoscopic (bronchial washing - BW, bronchial brushing - BB, endobronchial or transbronchial biopsy - ETBX) and transthoracic (transthoracic fine needle aspiration - TTFNA, transthoracic needle core biopsy - TTNCB) procedures. Primary diagnosis was documented as positive for malignant cells (PMC), suspicious for malignant cells, atypical cells present, negative-benign, or unsatisfactory. The secondary diagnosis included the specific classification of the neoplasm when available.

Results: Among the bronchoscopic procedures, ETBX had the highest yield for obtaining the primary diagnosis of “positive for malignant cells” (58.6%) and specific pre-resection classification. By EBTX, 65.6% of SQ cases and 100% of CT cases resulted in a primary diagnosis of PMC. Pre-resection histologic classification was made in 82.1% and 100% of SQ and CT cases, respectively, by ETBX. By contrast, only 44.2% of AD resulted in a primary diagnosis of PMC and in these cases, 62.3% were pre-surgically classified as AD. For transthoracic procedures, TTFNA had the highest yield for obtaining the primary diagnosis of “positive for malignant cells” (82.6%). However, specific pre-resection classification was most reliably performed by TTNCB - 82.6% for AD and 73.7% for SQ (in contrast to 47.1% for AD and 50.6% for SQ by TTFNA).

Conclusions: Bronchoscopic and transthoracic procedures have different performance characteristics depending on the histologic type of the neoplasm. The amount and type of diagnostic information required may influence the choice of pre-resection procedure.

80

Predicting Lung Adenocarcinoma Outcome Based on a Cytologic Scoring System

Carlie Sigel, MD¹, Dorota Rudomina, CT¹, Natasha Rekhtman, MD, PhD¹, William Travis, MD², Kim Geisinger, MD³, Andre Moreira, MD, PhD¹

¹Surgical Pathology and Cytopathology, Memorial Sloan-Kettering Cancer Center, New York, New York; ²Surgical Pathology, Memorial Sloan-Kettering Cancer Center, New York, New York; ³Surgical Pathology and Cytology, Wake Forest University, Winston-Salem, North Carolina

Introduction: Adenocarcinoma (AD) mixed subtype, the most common histologic type of lung cancer, is characterized by a variety of different architectural patterns. Recently, a three-tiered histologic pattern-based grading system was developed for stage I lung AD which stratified patients into low, intermediate, and high risk categories for disease recurrence. However, many lung cancer patients present with inoperable disease with cytology serving as the primary and only method for diagnosis. Attempts to correlate architectural arrangements between parallel cytologic and histologic preparations have been unsuccessful. Thus, we evaluated the cytomorphologic features of AD in patients previously scored by a histologic grading scheme. Features of potential prognostic significance were used to create a cytologic scoring system.

Materials and Methods: Retrospectively, we reviewed fine needle aspiration (FNA) specimens from 73 patients with mixed subtype ADs, blinded to a risk-stratification score previously assigned using a histologic pattern-based grading system. Selected cytomorphologic features were evaluated on ThinPrep® stained slides. Then, we grouped the specimens according to the predominant histologic pattern (bronchoalveolar, acinar and papillary, or solid and micropapillary) corresponding to histologic grade (low, intermediate, and high) and analyzed the distribution of these cytologic characteristics. Significant cytologic features were assigned a point value and the sum of these points was analyzed for the relationship to oncologic outcome data.

Results: Histologically low grade tumors correlated significantly with the following cytomorphologic characteristics: small nuclear size (<5x area of a resting lymphocyte) (p<0.001), nuclear size uniformity (p<0.001), and smooth nuclear contours. Features correlating with histologic high grade included large nuclear size (p<0.001), wide nuclear size variation (p<0.001), and presence of giant tumor nuclei (p=0.005). The cytologic scoring system showed a trend in separating histologically low and high grade tumors into two tiers of patient survival. Intermediate grade AD spanned the entire cytologic scoring spectrum.

Conclusions: Our cytologic scoring system can separate AD into two tiers; those with low risk of recurrence and those with high risk of tumor recurrence.

81

Fine Needle Aspiration Cytology of Basaloid Squamous Cell Carcinoma: A Comparison Study with Small Cell Carcinoma

Howard Wu, MD, Rebecca Marks, MD, Harvey Cramer, MD

Pathology and Laboratory Medicine, Indiana University School of Medicine, Indianapolis, Indiana

Introduction: Basaloid squamous cell carcinoma (BSCC) is an aggressive variant of squamous cell carcinoma that has a predilection for the upper aerodigestive tract and frequently metastasizes to cervical lymph nodes. Differentiating BSCC from small cell carcinoma on the basis of fine needle aspiration (FNA) cytomorphology is sometimes problematic. In this study, we report the FNA findings of 16 BSCC cases and compare those cases to 16 cases of small cell carcinoma. To our knowledge, this is the largest series of BSCC FNA cases ever reported.

Materials and Methods: A search of our laboratory information system was performed to identify all FNA cases of BSCC that had correlating histopathology. The Diff-Quik and Papanicolaou-stained direct smears and all available cell block sections from the FNAs and the histology slides from the corresponding biopsies or resections were retrieved and retrospectively reviewed. In addition, an equal number of consecutive FNA cases of small cell carcinoma with histologic confirmation were selected for comparison. The following FNA cytomorphologic features were compared for both tumors: cohesive tissue fragments, single cells, adenoid cystic-like features (cribriform pseudoglandular lumina with hyaline materials), necrosis, nuclear size, nuclear molding, nucleoli, cytoplasm and the presence of single keratinized cells. The data were analyzed with a t-test.

Results: There were 16 cases of BSCC originating from 13 patients, including 9 males and 4 females with an age range of 46-78 years (mean: 60 years). The original sites of the patients' primary BSCC included the tonsil (4 patients), tongue (1 patient), pharynx (2 patients), larynx (1 patient), anus (1 patient), lung (2 patients), esophagus (1 patient) and unknown site (1 patient). FNA was performed on 14 metastatic sites and on 2 primary tumors. The metastatic sites included cervical lymph nodes (10 cases), lung (2 cases) bone (1 case) and chest wall (1 case). The two primary BSCC tumors sampled by FNA originated from the lung. FNA features characteristic for BSCC include a smear pattern with a predominance of tightly cohesive tumor cell clusters, small to medium sized nuclei, single small nucleoli, nuclear molding and scant cytoplasm. These features were identified in all 16 cases. Necrosis was present in 13 cases, adenoid cystic-like features were identified in 5 cases and single keratinized cells were present in 3 cases. Sixteen cases of small cell carcinoma were retrieved for comparison. Cytomorphologic features common to both BSCC and small cell carcinoma included the presence of small to medium sized tumor nuclei with small or indistinct nucleoli, scant cytoplasm, nuclear molding and necrosis. Small cell carcinoma tends to have a predominantly dyshesive single cell pattern, while the smears from BSCC show a predominance of the tightly cohesive tumor cell clusters. However, differences in cellular cohesion between the two tumor types were not found to be statistically significant. Adenoid cystic features and the presence of single keratinized cells were specific for BSCC (p < 0.05).

Conclusions: BSCC is an unusual variant of squamous cell carcinoma with many cytologic features that overlap with small cell carcinoma. Specific FNA cytomorphologic features that can be used to differentiate BSCC from small cell carcinoma include the presence of adenoid cystic-like areas and single keratinized tumor cells.

Lymph Node/Hematopoietic 82

Cytologic Findings of Anaplastic Lymphoma Kinase (ALK)-Positive Large B-Cell Lymphoma

Suzanne Brandt, MD², Aashiyana Koreishi, MD¹, Julie Teruya-Feldstein, MD¹, Oscar Lin, MD, PhD¹

¹Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, New York; ²Department of Cytopathology, Weill Cornell Medical College, New York, New York

Introduction: ALK-positive large B-cell lymphoma (ALK-positive LBCL) is a rare variant of diffuse large B-cell lymphoma (DLBCL) that demonstrates strong granular cytoplasmic immunoreactivity for ALK and, commonly, the recurrent cytogenetic translocation, t(2;17)(p23;q23). Also, the cells might be immunoreactive for one or more of the markers of plasmacytic differentiation (CD138, CD38, and MUM1) and EMA. Only rare cases express CD45 or B-cell associated antigens (CD20, CD79a and PAX5). CD30 expression is absent. Cytokeratin has been shown to be focally positive in rare cases. Most patients present as stage III/IV disease. The objective of this study is to evaluate the cytologic findings of this recently described entity.

Materials and Methods: Fine needle aspiration (FNA) specimens from cases of ALK-positive LBCL with histologic confirmation were retrieved from our archives. We evaluated air-dried and alcohol fixed slides stained with Diff-Quick and/or H&E. Cell blocks and immunocytochemical stains were also reviewed. The following morphologic criteria were assessed: cellularity, cluster characteristics, cell size, cytoplasmic and nuclear characteristics and background composition.

Results: FNA specimens from 3 male patients ranging in age from 29 to 72 years old were evaluated. The sites sampled were cervical lymph node (n=2) and retroperitoneal lymph node (n=1). All specimens were moderately to highly cellular and composed of a population of large cells present in a background of small lymphocytes and lymphoglandular bodies. Single cells were present in all 3 cases, but 2 cases also showed the presence of clusters with acinar and papillary configurations. The cells were large with nuclei measuring more than 5 times the size of a mature T-lymphocyte. The cytoplasm was dense and the nuclei were mostly oval with occasional interspersed pleomorphic nuclei. Multinucleation was noted in all 3 cases. Single prominent nucleoli were commonly seen. Immunocytochemical studies showed cytoplasmic immunoreactivity for ALK in all 3 cases and CD45 expression in only 1 case. CD20 and cytokeratin immunostains were negative in all 3 cases.

Conclusions: ALK-positive LBCL is a rare variant of large cell lymphoma that can simulate carcinomas, both morphologically and immunophenotypically. The absence of CD45, CD20 and CD30 expression in neoplasms with at least focal cluster formation might lead to a misdiagnosis of carcinoma, particularly in FNA specimens. Analysis of ALK expression is critical to achieve the correct diagnosis.

83

Fine Needle Aspiration Diagnosis of Hodgkin Lymphoma: The Utility of Ancillary Studies and Core Biopsy

S. Kamat, MD, R. Felgar, MD, W. Khalbuss, MD, S. Monaco, MD

Pathology, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania

Introduction: Hodgkin Lymphoma (HL) is a cytologic diagnosis that often relies on finding characteristic Reed-Sternberg or Hodgkin cells in a mixed inflammatory background. Thus, the combination of cytomorphology and immunohistochemical (IHC) stains in fine needle aspiration (FNA) biopsies are helpful in establishing the diagnosis. Our aim was to look at the FNA diagnoses of HL in order to look at the ability of IHC, flow cytometry (FC), and core biopsy results to enhance the ability to make a definitive diagnosis.

Materials and Methods: Forty-one FNA cases from 35 patients with a cytologic diagnosis of HL or atypical/suspicious for HL were identified over a period of 3 years (May 2006-May 2009). We retrospectively reviewed these cases to look at the use of ancillary studies and core biopsy in enhancing the ability to make a definitive diagnosis, in addition to correlating with the available histopathological follow-up.

Results: The cytologic diagnoses, along with the ancillary studies performed and the number of cases with core biopsy and histological follow up are seen in Table 1. Of note, 6 (37%) of the 16 atypical/suspicious cases had no IHC due to insufficient cell block material; whereas all 20 cases with a definitive diagnosis of HL had sufficient cell block or core biopsy material for IHC. A core biopsy was performed with the FNA in 8 cases (20%), of which 6 (75%) were able to make a definitive diagnosis, compared to 19 of the 33 (58%) FNA cases without a core biopsy. A total of 27 (66%) cases had histological follow-up, which confirmed the diagnosis in 10/11 (91%) atypical/suspicious cases and 16/16 (100%) positive cases. One atypical/suspicious FNA case had granulomatous lymphadenitis on follow-up without evidence of HL.

Conclusions: Our data shows that concurrent core biopsies can be beneficial in limited FNA cases suspicious for HL to obtain material for confirmatory IHC stains, which can aid in making a definitive diagnosis. FC is important in some cases to exclude a non-Hodgkin lymphoma, including the possibility of CLL/SLL with a HL transformation, particularly in older patients with HL.

Table 1. 

FNA Diagnosis	No. of cases	Avg Age (yrs)	Cases w/ IHC	Cases w/ FC	Cases w/ Core Bx	Histologic follow-up(core or excision)
Atypical/ Suspicious  for HL	16 (39%)	47	10 (63%)	6 (38%)	2 (13%)	11 (69%)
HL	20 (49%)	47	20 (100%)	6 (30%)	6 (30%)	15 (75%)
HL with CLL/SLL	5 (12%)	72	5 (100%)	5 (100%)	0 (0%)	1 (20%)
Total	41	50	35 (85%)	17 (41%)	8 (20%)	27 (66%)

84

The Diagnostic Yield of Banded Karyotype Analysis is Low for Hematologic Fine Needle Biopsy Specimens

Sarah Kerr, MD, Kevin Meyer, BA, Robert Jenkins, MD, PhD

Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota

Introduction: Chromosomal abnormalities have been heavily incorporated into the World Health Organization classification of hematologic malignancies. The presence of key chromosomal abnormalities can be used by the cytopathologist to substantially narrow the differential diagnosis when hematologic malignancy is suspected, potentially obviating the need for an excisional biopsy. While many chromosomal abnormalities can be detected by fluorescence in situ hybridization (FISH), banded karyotype analysis does not require knowledge of the molecular target of interest, thus providing the attractive potential of detecting key abnormalities when the diagnosis is uncertain. Because our cytogenetics laboratory has seen an increase in requests for karyotypes on hematologic specimens, we sought to analyze the diagnostic yield of this expensive and labor-intensive procedure on small biopsy specimens.

Materials and Methods: Nineteen probable biopsy specimens were identified among 142 consecutive cases referred for banded karyotype analysis of potential hematologic malignancy over an 11-month period (blood and bone marrow specimens excluded). Of these, 9 were definitively identified on the requisition paperwork as fine-needle biopsies (FNBs). Ten additional specimens were assumed to be FNBs, because the gross morphology and specimen size (2 mm³ or less) were compatible with FNB. The remaining 123 specimens were tissue fragments compatible with excisional biopsy material. All specimens had been processed per routine laboratory procedure using an enzyme dissolution method, liquid media culture, and G-banded metaphase chromosome analysis. Concurrent FISH results were available for some cases.

Results: Detailed results for the FNB specimens are shown in Table 1. Median specimen size for the 19 FNBs was 2 mm³, in contrast to a median size of 8 mm³ for the excision group. Thirteen of the FNB cases were cancelled due to culture failure. Of the 6 cases for which metaphases could be analyzed, 4 demonstrated clonal abnormalities and 2 yielded normal metaphases. These results give a FNB reportable rate of 32%, a clonal rate of 21%, and a clonal/reportable ratio of 0.67. This is in contrast to the same statistics for the 123-case excision series of 80%, 37%, and 0.47, respectively. Concurrent FISH was available for 4 FNBs. In 3 of these cases, FISH identified a key abnormality when the culture failed to yield metaphases; in one case, FISH confirmed a key abnormality in a complex karyotype.

Conclusions: Banded chromosome analysis can be used as a tool in the diagnosis of hematologic malignancy. However, an adequate specimen is needed to produce meaningful results. The culture failure rate for FNB's is much higher than the average tissue hematologic specimen in our hands. When a reportable result is obtained, however, the clonal rate is similar to (or possibly higher than) our laboratory's clonal rate for larger specimens. Directed FISH analysis may provide a higher diagnostic yield at a lower cost for limited-size specimens.

Table 1. Consecutive Cytogenetics Results for Hematologic Fine-Needle Biopsies

Reason for Referral	Specimen Size (mm^3)	Interpretation	Karyotype	FISH results
DLBCL = diffuse large b-cell lymphoma; APL = acute promyelocytic leukemia; B-ALL = B-lymphoblastic leukemia/lymphoma; T-lymphoblastic leukemia/lymphoma; PTLD = post-transplant lymphoproliferative disorder; CLL = chronic lymphocytic leukemia.
Lymphoma	5	Normal	46,XX[20]
Lymphoma vs  Merkel Cell	2	Clonal	46-49,XY,+12[3] [cp4]/47,XY,  +6[3]/46,XY[13]
Lymphoma	2	Normal	46,XY[30]
Lymphadenopathy	1	Clonal	47,XY,+6[8]/48,idem, +8[2]/46,XY[30]
Lymphoma	0.05	Failed
DLBCL	0.5	Failed
Lymphoma	2	Failed
APL	2	Clonal	46,XY,psu dic(6;1) (q25;p13),  t(15;17)(q22;q21)  [15]/46,XY[5]	PML/RARA  fusion
Follicular Lymphoma	2	Failed		Three  copies of BCL2
B-ALL	3	Failed
Possible Recurrent  T-ALL	2	Failed		Three  copies of the 5'TCRB region
PTLD	2	Clonal	45-46,XY,-1,  der(6)t(1;6)  (q25;q21),add(7)  (q32),-8,-8,del(9)  (p22),+12,add(14) (q32),add(15)(q13), -16,-16,-17,-17,  -18,-20,  +8mar[cp3]
Possible Lymphoma	2	Failed
Possible Lymphoma	2	Failed
B-ALL	2	Failed
CLL, Lymphadenopathy	1	Failed		Trisomy 12
Lymphadenopathy	2	Failed
Lymphoma	2	Failed
Non-Hodgkin  Lymphoma	2	Failed

85

Evaluation of a Multiparametric Approach for the Diagnosis and Subtyping of B-Cell Non-Hodgkin Lymphoma in Fine Needle Aspirates: Cyto-histologic Correlation with Review of Discordant Cases

Hyangmi Ko, PhD, MD, William Geddie, MD, FRCPC, Scott Boerner, MD, FRCPC, Denis Bailey, PhD, MD, FRCPC, Gilda Santos, PhD, MD, FRCPC

Laboratory Medicine Program, University Health Network, Toronto, Ontario, Canada

Introduction: Cytogenetic findings have been recently incorporated to cytology reports for diagnostic interpretation of non-Hodgkin lymphoma (NHL) as recommended by the current WHO classification. Studies addressing the limitations and diagnostic efficiency of NHL using fine needle aspirates (FNA) have used a combination of immunophenotypic and cytomorphologic data. We compared the cytologic diagnosis of NHL in FNA using a multiparametric approach, which included cytogenetic findings, with corresponding core biopsies or surgical specimens, to establish diagnostic limitations.

Materials and Methods: FNA cases with a final diagnosis of B-cell non-Hodgkin and available histopathological reports were selected during the period from October 2003 to September 2009. Only FNAs with both immunophenotypic and cytogenetic data were included. Histopathological diagnoses were retrieved from the reports of concurrent, previous or subsequent core biopsies or surgical specimens. Cytology diagnoses were correlated with histopathology in order to identify discordant cases. All discordant cases were reviewed to determine the reasons for disagreement and were further divided into major and minor discordances according to potential impact in clinical treatment.

Results: A total of 185 FNAs from 176 patients with a final diagnosis of lymphoma and corresponding core biopsies and/or surgical specimens were available. Among them, specific subclassification was possible in 174 cases. There were 46 cases with concurrent histology; the other cases had either prior or follow-up surgical biopsies. Cyto-histologic correlation revealed diagnostic agreement in 140 cases (75.7%). Review of 45 discordant cases showed that 25 biopsies were from a site different than the aspiration site. Analysis of the reasons for the discordance revealed that in 11 cases with different location and 1 case with same location, the biopsies were taken with an interval of at least 7 months and showed transformation to diffuse large B-cell lymphoma (DLBCL). In 4 additional cases, cytogenetic analysis failed or was not done in biopsy specimens. No lymphoma was detected in 3 cases in which the biopsy was performed in a different location several months before or after the cytologic diagnosis. Those cases were excluded from the analysis for clinical impact. Other reasons were: no subtyping in cytology (7 cases), sampling problems with no DLBCL in FNA (7 cases), problems in differentiating follicular lymphoma grade 3 and DLBCL (6 cases), wrong subtyping (3 cases), false positive (2 cases) and different grading for follicular lymphoma (1 case). Major disagreement was detected in 10 cases (5.4%).

Conclusions: A multiparameter approach for FNA interpretation provides specific subclassification of B-cell NHL with a high concordance rate with histological diagnosis. The discordances are mainly related to difficulties in the differential diagnosis between DLBCL and follicular grade 3 and sampling of low-grade areas. In a small proportion of cases, the disagreement could have impacted clinical treatment and were attributed to sampling problems. In patients with multiple lymph nodes, FNA should be performed in different sites.

86

Fine Needle Aspiration Cytology of Large Cell Lymphoma: A Study of 177 Patients

Rebecca Marks, MD, Harvey Cramer, MD

Pathology and Laboratory Medicine, Indiana University School of Medicine, Indianapolis, Indiana

Introduction: The use of fine needle aspiration (FNA) cytology to diagnose non-Hodgkin lymphoma (NHL) has been considered controversial, largely due to the lack of architectural information that can be obtained from cytologic material. While the reported sensitivity and specificity for the FNA diagnosis of NHL ranges between 80 and 90 percent higher accuracy rates are anticipated for the FNA diagnosis of large cell lymphoma (LCL).

Materials and Methods: A search of our laboratory information system was performed for the 8-year period ending June 30, 2009. All FNA reports diagnosed as LCL, all hematopathology (HP) cases diagnosed as diffuse large B-cell lymphoma (DLBCL), all correlating FNA and HP reports and all flow cytometry reports and selected microscopic slides were reviewed.

Results: There were a total of 177 patients included in the study, 30 of which had more than one corresponding FNA. A prior history of NHL was noted in 56 patients (31%). Less than half of our patient population (77/177, 44%) had a concurrent HP biopsy at the time of the FNA. Of the 177 patients, 38 (21%) had an FNA diagnosis of LCL that was not subject to HP confirmation and 25 of these patients (14%) appear to have been treated based solely on the FNA diagnosis. There were 63 FNA cases diagnosed as LCL with HP follow-up and the FNA diagnosis was confirmed in 58 of 63 cases (92%). The 5 FNA errors included 2 cases of Hashimoto's thyroiditis, 2 cases of follicular lymphoma and 1 case of MALT lymphoma that had been misdiagnosed as LCL by FNA. There were 76 cases of HP-proven DLBCL that were not diagnosed as LCL by FNA. The FNA diagnoses for these 76 cases included: malignant lymphoma NOS (21 patients, 28%) suspicious for LCL (7 patients, 9%), follicular lymphoma (3 patients, 4%), atypical lymphoid cells present (17 patients, 22%), suspicious for malignancy (2 patients, 3%), negative for malignancy (15 patients, 20%) and unsatisfactory sample (11 patients, 14%). Flow cytometry had been performed on FNA material in 121 patients (68%). B-cell monoclonality was detected in 76 patients (63%), not detected in 12 patients (10%) and flow cytometry was unsatisfactory in 33 patients (27%). Immunocytochemistry was performed on 20 FNA samples, with almost half (9) of these being FNA-only cases.

Conclusions: Failure rates for the FNA diagnosis of LCL were much higher than expected. A definitive diagnosis of LCL was established by FNA in 57% of our cases and the accuracy for those cases with HP follow-up was only 92%. These results appear to undermine the policy of treating selected patients with an FNA-established diagnosis of LCL without obtaining a confirmatory surgical biopsy.

87

Cyto-Immunostaining of Aspirate Smears Can Aid in the Diagnosis of Hodgkin Lymphoma by FNA Biopsy

Jill Ono, MD¹, Stewart Knoepp, MD, PhD², Ronald Balassanian, MD¹

¹Pathology, Massachusetts General Hospital, Boston, Massachusetts; ²Pathology, University of Michigan, Ann Arbor, Michigan

Introduction: The diagnosis of Hodgkin lymphoma (HL) by FNA biopsy can be challenging and often requires immunostaining when the diagnosis is suspected by cytomorphology. Immunohistochemical (IHC) staining can be done on the cell block sample; however, the large size of the targeted Reed-Sternberg (RS) cell may result in the RS cells not being well represented on the sections submitted for IHC. Additionally, obtaining an adequate cell block, particularly in cases with extensive sclerosis, can be difficult. We performed immunocytochemical (ICC) staining on air-dried aspirate smears, obtained at the time of FNA biopsy, for a series of patients for whom HL was suspected based on rapid interpretation at the time of the FNA biopsy.

Materials and Methods: Seven patients were seen in the Massachusetts General Hospital or University of Michigan FNA clinics, and FNA biopsy was performed and interpreted by a cytopathologist. In all cases, the rapid interpretation was either “Hodgkin lymphoma” or “suspicious for Hodgkin lymphoma” based on review of Diff-Quik and Toluidine blue stained slides. The aspirate material from FNA passes was divided and separated into unstained air-dried aspirates and slides for routine cytochemical staining. The unstained slides were then submitted for ICC. In all cases, ICC for CD15 and CD30 was performed. In 3 cases, ICC for PAX5 was also performed. In all 7 cases, a formalin fixed cell block was also obtained and in cases with adequate material, IHC was performed. IHC on the cell block served as a control sample for ICC/IHC. For the final diagnosis, all slides were reviewed, including routine Papanicolaou and Diff Quik stains, as well as ICC on aspirate smears and IHC and routine H&E on sections of the cell block sample. The ICC stained smears were compared to the IHC of the cell block and subsequent lymph node excisions or core biopsy specimens.

Results: ICC for CD15 and/or CD30 on aspirate smears was positive in large RS cell in 7 of 7 cases. In 6 of 7 cases, the membrane and Golgi staining could be seen more clearly in ICC stained slides. ICC staining for PAX5 showed loss of staining in RS cells compared to B-cells in 3 cases. IHC for CD15 and CD30 and/or PAX5 was performed on the cell block in 4 cases, and on the excision specimen in 5 cases. RS cells identified by ICC on smears correlated well with RS cells seen on the surgical excision or core biopsy.

Conclusions: RS cells were seen to better advantage and were more numerous and more easily identified, with membrane and Golgi staining, by ICC on aspirate smears, when compared to IHC on cell block sections. We attribute this finding to the large size of the RS cells, which may not be well represented on select sections of a small cell block sample. Our study demonstrates a relatively quick and simple procedure for performing ICC on FNA samples when HL is the suspected diagnosis. ICC on aspirate smears has a high diagnostic yield and may aid in the diagnosis of HL by FNA biopsy.

88

Diagnosis of Adult T-Cell Leukemia/Lymphoma from Cytological Specimens Using Combined Cytopathologic, Flow Cytometric and Antigen-Receptor Rearrangement Studies

Yaxia Zhang, MD, PhD¹, Phillip Ruiz, MD, PhD², Parvin Ganjei-Azar, MD¹

¹Pathology, University of Miami, Miller School of Medicine-Jackson Memorial Hospital, Miami, Florida; ²Surgery, University of Miami, Miller School of Medicine, Miami, Florida

Introduction: Adult T-cell leukemia/lymphoma (ATLL) is a rare and lethal T-cell neoplasm caused by human T-cell lymphotropic virus, type I (HTLV-1). Histopathologically, ATLL is characterized by a broad spectrum of morphological appearances. Diagnosis of ATLL from cytological specimens either from exfoliated fluid or fine needle aspiration of lymph node is challenging and few studies have addressed this rare entity.

Materials and Methods: To identify the cytological diagnostic features of ATLL in cytologic samples, we retrospectively reviewed all cases of ATLL at Jackson Memorial Hospital between 2003 and 2009.

Results: Ten patents were diagnosed with ATLL. Four cytological samples from four different patients were found: one cerebrospinal fluid, two peritoneal effusions, and one fine needle aspirate from a neck mass. In cerebrospinal fluid, the cytology sample tends to be low in cellularity. However, flow cytometry was able to identify a minimal CD2+, CD3+, CD4+, and CD25 T population. PCR for T cell/B cell antigen receptor rearrangement did not detect monoclonality due to insufficient DNA quality from this sample. In two peritoneal effusions and one FNA sample, we noted a pleomorphic polymorphous population of lymphocytes that ranged in size from small to intermediate to large. The nuclei varied from round to oval to irregular with some showing highly convoluted, flower-like shapes. The irregular nuclear contours were observed in small lymphocytes as well as in larger ones. Variably prominent nucleoli and apoptotic bodies were commonly seen. Flow cytometry study revealed the atypical T cell population with CD2+, CD3+, CD4+, CD5+ and CD25+ expression. PCR studies showed gene rearrangements in the gamma chain gene of the T-cell receptor.

Conclusions: Despite the polymorphous nature of ATLL, diagnosis of ATLL from cytological samples can be established by close attention to nuclear cytologic features in conjunction with flow cytometry and PCR studies.

Pancreas/Liver

89

Apomucins and Diagnostic Yield of Pancreatic EUS-FNA Cytologic Material

Wipapat Apichairuk, MD¹, Kalyan Ray Parashette, MD², Allen Halline, MD³, Odile David, MD¹

¹Pathology, University of Illinois at Chicago, Chicago, Illinois; ²Pediatrics, University of Illinois at Chicago, Chicago, Illinois; ³Digestive Disease and Nutrition, University of Illinois at Chicago, Chicago, Illinois

Introduction: A variety of mucin (MUC) glycoproteins have been identified in normal human tissues and aberrant expression of some of these proteins has been implicated in the pathogenesis of various precancerous and cancerous conditions. Recent studies have shown MUC1 and MUC4 to be promising biomarkers of pancreatic ductal adenocarcinoma. To date, these two biomarkers have not been studied together in pancreatic EUS-FNA material. The purpose of this study is to evaluate whether ancillary use of MUC1 and MUC4 immunostaining increases the diagnostic yield of pancreatic ductal adenocarcinoma in EUS-FNA samples.

Materials and Methods: Of the 70 EUS-FNAs of pancreas identified in the files of the University of Illinois Medical Center from January 2007 to September 2008, 34 specimens had cell blocks that could be retrieved. The original H&E slides were reviewed and additional sections were cut for MUC1 and MUC4 staining. Of these 34 specimens, 6 specimens were censored for suboptimal cellularity and 2 for clinical histories of metastatic malignancies. An additional 4 specimens had to be excluded from the evaluation of MUC4 staining due to exhaustion of the tissue blocks. Staining of both benign and atypical/malignant groups, if present, was reported separately for each specimen as positive or negative. A specimen was reported as positive if either benign or atypical/malignant groups were positive. A specimen was reported as negative only if all groups were negative. Appropriate positive and negative controls were performed.

Results: The results are summarized in Table 1.

Note: Specimens originally classified as atypical were included for statistical purposes under adenocarcinoma only if a subsequent surgical specimen demonstrated malignancy. Two atypical cases were censored from the study due to lack of outcomes data.

MUC1 is more consistently positive than MUC4 in EUS-FNA samples of pancreatic ductal adenocarcinoma. For MUC1, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) are 1. For MUC4, sensitivity and specificity = 0.8, PPV = 0.92 and NPV = 0.57.

Conclusions: In this small study, MUC1 was more sensitive and specific than MUC4 for the identification of pancreatic ductal adenocarcinoma in EUS-FNA samples. Both MUC1 and MUC4 seem to be promising biomarkers that might increase the diagnostic yield of pancreatic EUS-FNA. Larger studies are indicated.

Table 1. 

Diagnosis(Dx)	MUC1 (total evaluable n=24) # of positive cases/total # of cases of corresponding Dx	MUC4 (total evaluable n=20) # of positive cases/ total # of cases of corresponding Dx
Pancreatic ductal  adenocarcinoma	19/19	12/15
Benign	0/5	1/5

90

Performance of PathfinderTG® CEA and Amylase Pancreatic Cyst Fluid Analysis

Ivan Chebib, MD, FRCPC¹, Brian Turner, MD², William Brugge, MD², Martha Pitman, MD¹

¹Cytopathology Unit, James Homer Wright Pathology Laboratories, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts; ²Division of Gastroenterology, Department of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts

Introduction: Endoscopic ultrasound fine needle aspiration (EUS-FNA) of pancreatic cyst fluid utilizes biochemical cyst fluid analysis for carcinoembryonic antigen (CEA) and molecular testing (DNA content, KRAS mutation, LOH) for prediction of mucinous histology and likelihood of progression. PathfinderTG® (RedPath Integrated Pathology Inc, Pittsburgh, PA) recently implemented CEA and amylase testing in addition to its molecular testing on EUS-FNA cyst aspirates. We compare PathfinderTG® cyst CEA analysis to Massachusetts General Hospital (MGH) laboratory results.

Materials and Methods: Between 2008 and 2010, 14 patients with pancreatic cysts had EUS-FNA cyst fluid analysis with CEA and amylase measurement and molecular analysis (DNA content, KRAS mutations, LOH) through PathfinderTG®. The PathfinderTG® CEA levels were directly compared to CEA measured at MGH (Cohen κ coefficient). We also assessed whether the PathfinderTG® CEA level was predictive of mucinous etiology by comparing to either histology or a diagnosis based on a combination of cytology, clinical features, EUS findings, and cyst fluid analysis where MGH CEA ≥ 192 ng/mL and/or KRAS mutation, 2+ LOH mutations or greatly elevated DNA concentration defines a mucinous cyst.

Results: There were 14 patients who had CEA measurements performed by PathfinderTG®. There were 9 females (median age 77, range 52 - 86) and 5 males (median age 68, range 42 - 84). The median volume of cyst fluid aspirated was 5.0 mL (range 1 - 20 mL), including 7 cases with 5 mL or less. There were 6 cysts with both CEA and amylase measurements at PathfinderTG® and at MGH. Comparison of the two results demonstrated perfect agreement between the two laboratories (κ = 1.0, 95% CI 1.0), when CEA ≥ 192 ng/mL was considered positive for a mucinous cyst.

Three cases had PathfinderTG® CEA exceeding 192 ng/mL; all three were resected and demonstrated mucinous etiology. Eleven cases had CEA < 192 ng/mL: 2 resected, 6 diagnosed with cytology, EUS, and fluid chemistry, and 3 without cytology, diagnosed with EUS and fluid chemistry only. Of the 11 with PathfinderTG® CEA < 192ng/mL, 4 were considered mucinous cysts. PathfinderTG® CEA measurement has a specificity of 100% and a sensitivity of 43%.

Conclusions: Preliminary data demonstrate that CEA and amylase measurements from PathfinderTG® are comparable with analysis at MGH. Sensitivity of CEA measurement alone was low, but specificity for a mucinous cyst was 100% for levels > 192 ng/ml.

91

Inspissated Mucin in EUS-FNA Aspirates from Pancreatic Cysts: A Potential Diagnostic Pitfall

Ivan Chebib, MD, FRCPC, Martha Pitman, MD

Cytopathology Unit, James Homer Wright Pathology Laboratories, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts

Introduction: Aspirated thick and thin extracellular mucin (ECM) identified on cytology of pancreatic cyst fluid is characteristic of a mucinous cyst (intraductal papillary mucinous neoplasm (IPMN) or mucinous cystic neoplasm (MCN)). The presence of thick inspissated mucin is uncommon and can create a diagnostic pitfall. We report on this finding in our experience with EUS-FNA of pancreatic cysts.

Materials and Methods: The clinical, radiological, cytohistological and available fluid chemical analysis in 10 EUS-FNA specimens from pancreatic cysts with inspissated mucin are reported.

Results: There were 7 females (average age 64.6 years, range 45 - 82) and 3 males (average age 63.3, range 53 - 76). By EUS, the cysts averaged 2.0 cm (range 0.9 - 5.0 cm), and were located in the tail (5), body (4), and head (1). Three cysts had a thick cyst wall and two demonstrated a mural nodule. Pre-operative suspicion of a mucinous cyst was present in 6 of 9 cases where EUS was available.

The cytology demonstrated fan-shaped to geometric-shaped fragments of acellular material made up of fibrillary or needle shaped structures. They were blue-green to red-brown on Papanicolaou-stained, alcohol-fixed and liquid-based material. Alcian blue and/or mucicarmine stain was positive in 6 cases. Atypical epithelial cells (AEC) were noted in 1/10 cases. No sample was diagnostic of malignancy. Cytological interpretations were consistent with a mucinous cyst in 8/10 cases; however, in 2 cases the inspissated mucin was not recognized as mucin and was interpreted as precipitate or crystalline material. CEA was available in 5 cases, with only 2 cases measuring > 192 ng/mL, predictive of a mucinous neoplasm. There were 7 cases with amylase measurements, 6 of which exceeded 10,000 U/L. Four cases were resected: 3 MCN (adenoma) and 1 IPMN with carcinoma in-situ and a focus of invasive carcinoma, all alive without disease (median 70.0 months, range 12.2 - 163 months). In those patients without resected cysts, all patients were alive with follow-up between 8.1 - 49.7 months (median 12.2 months).

Conclusions: Inspissated mucin may not be recognized on cytology. It can be confused with crystals, proteinaceous debris, bacteria and other contaminants. EUS features, elevated CEA levels and AEC may be helpful to identify the nature of the cyst as mucinous.

92

Secondary Liver Tumors Diagnosed by Fine Needle Aspiration: Updating Old Data in the Proper Patient Cohort

Joshua Hanson, MD¹, Marie Dvorakova, MD², Sharon Bihlmeyer, MD¹, Liza Quintana, BS¹, Walid Khalbuss, MD², Shelly Naud, PhD³, Abdelmonen Elhosseiny, MD¹

¹Pathology, University of Vermont and Fletcher Allen Health Care, Burlington, Vermont; ²Pathology, University of Pittsburgh, Pittsburgh, Pennsylvania; ³Biostatistics, University of Vermont, Burlington, Vermont

Introduction: Secondary liver tumors (SLTs) are more common than primary tumors with a reported ratio of 40:1. It is widely believed that colon cancer is the most common SLT. This assertion is supported by three large autopsy studies published over twenty years ago. These studies contribute substantially to the medical community's knowledge regarding the relative proportions of SLTs, as they are referenced in popular publications such as the World Health Organization's Tumours of the Digestive System (2000), and the Armed Forces Institute of Pathology's Tumors of the Liver and Intrahepatic Bile Ducts (1999). However, the autopsy studies are limited from a cytopathologist's perspective as they reflect the prevalence of SLTs in deceased patients and predate the current era of widespread colon cancer screening. To our knowledge, there are no studies within the last decade that report the prevalence of SLTs in a cohort of patients presenting for fine needle aspiration (FNA), the patient population in which cytopathologists are primarily interested. The purpose of our study was to determine the relative proportions of SLTs based on the primary site of malignancy in a cohort of patients evaluated by FNA.

Materials and Methods: Five hundred seven patients (255 men, 252 women, mean age 66 years) who underwent FNA of SLTs at the University of Vermont (140 cases, 2004-2007) and the University of Pittsburgh (367 cases, 2005-2008) were retrospectively identified. All cases were reviewed by an experienced cytopathologist. Morphology and immunocytochemistry were correlated with clinical and radiologic findings to identify the origin of the SLTs.

Results: Please see Table 1 for complete results. The proportions of SLTs by primary site were as follows (rounded values): pancreas 23%, lung 16%, colorectal 13%, breast 7%, skin melanoma 5%, and stomach 2%. Esophagus, renal, bladder, and hematopoietic SLTs each showed rates of 3%. Ocular melanoma, ovary, small bowel and uterus each comprised 1%. All other sites showed rates of <1%. A primary site was not identified in 16% of the cases. The proportions of SLTs by histologic type strongly favored adenocarcinoma and neuroendocrine tumors. Adenocarcinoma comprised 61% of SLTs, of which 32% were pancreas (Figure 1), 20% colorectal (Figure 2), 16% unknown origin, 10% breast, and 6% lung. Neuroendocrine tumors comprised 16% of SLTs, of which 52% were lung (Figure 3), 17% pancreas, 17% unknown origin, and 6% small bowel. Poorly/undifferentiated carcinomas and malignant neoplasms not otherwise specified comprised 9% of SLTs, melanoma 6%, lymphoma 3%, urothelial 2%, squamous 1% (57% lung), mesenchymal 1%, and germ cell 0.4%. Metastases from pancreas cancer were more prevalent in our study than in the 3 prior autopsy studies combined (23% vs. 6%, p-value < .0001).

Conclusions: Pancreas and lung cancers are the most common SLTs in a cohort of patients who presented for FNA. Most notably, the proportion of SLTs resulting from metastatic pancreas cancer is significantly higher in these patients than the rates reported in the prior autopsy studies and may be the most common SLT diagnosed by cytopathologists. These results emphasize the importance of understanding prevalence rates in the patient population one works with and should help cytopathologists formulate a differential diagnosis when faced with a SLT of unknown origin.

Table 1. Secondary Liver Tumors by Primary Site and Histologic Type

Primary Site	Total	%(95% CI)	Adenocarcinoma n(%)	Neuroendocrine n(%)	Poorly DifferentiatedCarcinoma n(%)	Melanoma n(%)	Lymphoma n(%)	Urothelial n(%)	Malignant Neoplasmn(%)	Squamous Cell Carcinoma n(%)	Mesenchymal n(%)	Germ Cell n(%)	Nasopharyngeal carcinoman(%)
* 3 pancreatic adenocarcinoma cases were subclassified as adenosquamous carcinoma. † Gastrointestinal Stromal Tumor (GIST) ∧ leiomyosarcoma • Chordoma α Malignant Hemangiopericytoma.
Pancreas	117	23 (19, 27)	98 (32)*	14 (17)	5 (14)
Unknown	83	16 (13, 20)	49 (16)	14 (17)	10 (28)				9 (100)	1 (14)
Lung	79	16 (13, 19)	18 (6)	43 (52)	14 (39)					4 (57)
Colorectal	64	13 (10, 16)	63 (20)							1 (14)
Breast	33	7 (5, 9)	30 (10)	1 (1)	2 (6)
Skin	23	5 (3, 7)				23 (79)
Esophagus	16	3 (2, 5)	15 (5)		1 (3)
Kidney	14	3 (2, 5)	13 (4)		1 (3)
Bladder/Ureter	13	3 (1, 4)			2 (6)			11 (100)
Hematopoietic	13	3 (1, 4)					13 (100)
Stomach	10	2 (1, 4)	8 (3)								2 (29)†
Eye	6	1 (0.4, 3)				6 (21)
Ovary	6	1 (0.4, 3)	5 (2)	1 (1)
Small Bowel	5	1 (0.3, 2)		5 (6)
Uterus	5	1 (0.3, 2)	3 (1)								2 (29)∧
Prostate	4	0.8 (0.2, 2)	3 (1)		1 (3)
Ampulla of Vater	3	0.6 (0.1, 2)	3 (1)
Mesentery	2	0.4 (0.05, 1)		1 (1)							1 (14)†
Thyroid	1	0.2 (0, 1)		1 (1)
Cervix	1	0.2 (0, 1)								1 (14)
Testis	1	0.2 (0, 1)										1 (50)
Nasopharynx	1	0.2 (0, 1)											1 (100)
Abdominal Paraganglioma	1	0.2 (0, 1)		1 (1)
Gallbladder	1	0.2 (0, 1)	1 (0.3)
Buccal Mucosa	1	0.2 (0, 1)		1 (1)
Sacrum	1	0.2 (0, 1)									1 (14)•
Larynx	1	0.2 (0, 1)		1 (1)
Abdomen Soft Tissue	1	0.2 (0, 1)									1 (14)α
Germ Cell Tumor	1	0.2 (0, 1)										1 (50)
Total	507	100%	309 (61)	83 (16)	36 (7)	29 (6)	13 (3)	11 (2)	9 (2)	7 (1)	7 (1)	2 (0.4)	1 (0.2)

93

Cytomorphological Spectrum and Ancillary Studies in Liver Fine Needle Aspiration Biopsies of Non-Epithelial Neoplasms: A Study of 57 Cases

Walid Khalbuss, MD PhD, Sara Monaco, MD

Pathology, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania

Introduction: Non-epithelial lesions of the liver encompass a wide variety of benign and malignant conditions. However, due to their rarity, there are few reports describing the fine needle aspiration (FNA) biopsy findings in the literature. The current review emphasizes the cytomorphologic features, differential diagnosis, and potential pitfalls associated with these lesions.

Materials and Methods: We retrospectively reviewed US-guided liver FNAs with a positive diagnosis for malignancy or neoplasm over a 4-year period, along with available ancillary studies, pathological follow-up, and clinical data. For each case, we reviewed the cytomorphology, the final cytologic diagnoses, ancillary studies, and available corresponding histological material and clinical follow-up.

Results: A total of 576 ultrasound-guided liver FNAs with positive diagnosis of malignancy or neoplasm were retrieved from our records over a period of 48 months. Fifty-seven cases (1%) were diagnosed as non-epithelial lesions. The average age of the patients was 48 years (range: 30-86 years). Of these 57 FNAs, 26 (46%) were metastatic melanoma, 15 (26%) were mesenchymal tumors (7 metastatic sarcomas, 5 hemangiomas, and 3 metastatic gastrointestinal stromal tumors/GIST), 13 (23%) were lymphomas, 1 case each of metastatic meningioma, metastatic granulosa cell tumor, and bile duct hamartoma. A prior history of the neoplasm was present in 42 cases (74%). The cases with no prior history included: 5 hemangiomas, 5 lymphomas, 2 GIST, 1 melanoma, 1 leiomyosarcoma, and 1 bile duct hamartoma. Ancillary studies were utilized in 52 cases (91%) including immunohistochemical stains (IHC) in 42 cases, flow cytometry in 9 cases, and molecular testing in 5 cases (fluorescence in situ hybridization/FISH studies in 4 cases and PCR in one case). Histologic correlations were available in 52 cases with no discrepancies seen.

Conclusions: Non-epithelial neoplasms of the liver are rare and can be difficult to characterize in liver FNAs; however, our study shows the spectrum of lesions and the utility of ancillary studies in making these diagnoses. The most common entities include metastatic melanoma, mesenchymal tumors, and malignant lymphomas, and 26% of these cases occurred in patients with no previous history of tumor. By correlating clinical and radiological data, cytologic findings, and ancillary studies, a definitive specific diagnosis can be achieved with FNA cytology.

94

Improved Diagnostic Accuracy for Pancreatic Cystic Lesions: The OHSU Experience

Sarah Wachs, ASCP¹, Douglas Faigel, MD², Terry Morgan, MD, PhD¹

¹Pathology, OHSU, Portland, Oregon; ²Gastroenterology, OHSU, Portland, Oregon

Introduction: Endoscopic ultrasound (EUS)-guided FNA is becoming routine in many hospitals and cytologists are gaining experience reviewing previously uncommon preparations. Previous reports of EUS-guided pancreatic cyst FNA diagnoses suggested only moderate diagnostic accuracy (50%) distinguishing mucinous neoplasms from benign lesions. Therefore, many clinicians supplement cytology with cyst fluid CEA levels. Now that EUS-guided FNAs are more common and cytologists are more experienced with these lesions, we hypothesized that our cytology may yield significantly improved diagnostic accuracy of pancreatic cyst lesions.

Materials and Methods: We identified a retrospective series of 47 pancreatic cyst EUS-guided FNAs in our OHSU archives that had cytology, cyst fluid CEA levels, and surgical or long-term clinical follow-up. Aspirate cytology was blindly evaluated by a cytotechnologist (SW) and diagnosed as either “Negative”, “Cystic Mucinous Neoplasm +/- Atypia”, or “Adenocarcinoma.” These diagnoses were compared to the final cytologic diagnosis, cyst fluid CEA levels (ng/ml), and final surgical resection specimen. Sensitivity, specificity, PPV, NPV, and diagnostic accuracy of the cytologic diagnoses and cyst fluid CEA levels (>192 ng/ml) were calculated.

Results: See Table 1 - There was good agreement between the blinded cytotechnologist's review and the final pathologic diagnosis (kappa statistic = 0.7 (95% CI: 0.4-0.9). Final outcome diagnoses by surgery was available in 38/50 cases. The remainder had at least 5 years of negative follow-up. Thirty one of 50 cases were neoplastic (mucinous 23, adenocarcinoma 6, islet cell 1, pseudo-papillary 1) and 19 were non-neoplastic (16 pseudocyst, 3 microcystic adenomas). Our diagnostic accuracy was significantly greater than CEA levels and prior published reports.

Conclusions: Diagnostic experience appears to have significantly improved the accuracy of pancreatic cyst EUS-guide FNA diagnoses and questions the current utility of cyst fluid CEA levels.

Table 1. Diagnostic Accuracy

Outcome Metric	Cytotech	Final Diagnosis	CEA (>192 ng/ml)
Sensitivity (SN), Specificity (SP), Positive predictive value (PPV), Negative predictive value (NPV).
SN	80%	81%	38%
SP	88%	100%	78%
PPV	92%	100%	73%
NPV	71%	76%	44%
Accuracy	83%	88%	53%

95

Utility of Telecytopathology for Rapid Preliminary Diagnosis of Endoscopic Ultrasound-Guided Pancreatic Fine Needle Aspiration Biopsies

Rong Rong, MD, PhD¹, Brian Pavlovitz, MD¹, Ajoy Roy, MD², Kamal Khurana, MD¹

¹Pathology and Laboratory Medicine, SUNY Upstate Medical University, Syracuse, New York; ²Gastroenterology, SUNY Upstate Medical University, Syracuse, New York

Introduction: Telecytopathology is being increasingly used as a medium that can facilitate immediate assessment of cytology specimens from fine needle aspirations (FNAs). We present our institution's experience of using telecytology in rendering a preliminary diagnosis on endoscopic ultrasound-guided pancreatic FNA procedures.

Materials and Methods: All Diff-Quik slides that were prepared in the endoscopy suite for adequacy assessment were evaluated using a telepathology system consisting of an Olympus CX41 microscope and digital camera with NetCam software. Preliminary diagnoses were rendered on real-time images transmitted by cytotechnologists or fellows to the attending cytopathologist's PC over the hospital intranet; constant verbal communication with the attending was established via telephone. The concordance between preliminary and final diagnosis using telepathology was evaluated; diagnostic accuracy was based on the initial evaluation. The reasons for discrepancies were analyzed.

Results: A total of 40 endoscopic ultrasound-guided pancreatic FNAs were assessed for preliminary diagnosis. Of these, 22(55%) were classified as benign, 8(20%) positive for malignancy and 10(25%) suspicious for malignancy. All 8 cases classified as “positive for malignancy” at the time of preliminary diagnosis had the same final diagnosis. Nine of ten cases with an initial diagnosis of “suspicious for malignancy” were upgraded to a final diagnosis of “positive for malignancy;” the diagnosis of the remaining case was unchanged. Of the 22 cases that were initially classified as benign at the time of preliminary diagnosis, 3 were changed to a final diagnosis of “positive for malignancy”. For calculation of concordance rates, FNAs diagnosed as suspicious for malignancy at the time of preliminary diagnosis were considered to be concordant if the final diagnosis was the same or “positive for malignancy”. The overall concordance rate between preliminary and final diagnoses was 92.5%. Three discrepant cases that were initially diagnosed as benign showed bland tumor cells of acinic cell carcinoma (1) and neuroendocrine neoplasm (2) that on retrospective review were initially misinterpreted as benign acinar epithelium at the time of image transmission.

Conclusions: Telecytopathology is a reliable and effective method for rendering preliminary diagnoses during endoscopic ultrasound-guided pancreatic FNA procedures. It saves valuable time for the cytopathologist, and makes real-time consultation possible. Tumors with bland cytomorphology may prove to be diagnostically challenging at the time of initial review via telecytopathology.

96

Atypical Epithelial Cells in EUS-FNAB Samples of Pancreatic Mucinous Cysts Correlates with a High Risk of Malignancy

Martha Pitman, MD¹, Muriel Genevay, MD¹, Kurt Yaeger, BS¹, Ivan Chebib, MD¹, Brian Turner, MD², Mari Mino-Kenudson, MD¹, William Brugge, MD²

¹Pathology, Massachusetts General Hospital, Boston, Massachusetts; ²Medicine, Massachusetts General Hospital, Boston, Massachusetts

Introduction: Cytology of cyst fluid for the diagnosis and management of patients with mucinous cysts is controversial. We hypothesized that recognizing atypical epithelial cells (AEC) in addition to overtly malignant epithelial cells as an indication of malignancy would improve the detection of malignancy in these cysts and be an important risk factor for malignancy in the pre-operative assessment for surgical management.

Materials and Methods: We reviewed clinical, radiological and cytological data of 112 patients with histologically confirmed, currently classified mucinous cysts of the pancreas evaluated in a single institution tertiary medical center between 1993 and 2008 with pre-operative endoscopic ultrasound and fine needle aspiration biopsy. Cytology slides were blindly re-reviewed in all cases. Radiological parameters were retrieved from medical records. Neoplasms were grouped as benign (low grade and moderate dysplasia) and malignant (in-situ and invasive carcinoma) and categorical variables compared Fisher's Exact Test with significance at p=0.05.

Results: There were 92 IPMN and 20 MCN; 39 malignant [high grade dysplasia/carcinoma in-situ and malignant] and 73 benign [42 low grade dysplasia + 31 moderate dysplasia]. Only 28% (11/39) of the cysts were cytologically malignant. Atypical epithelial cells consisted of 3-dimensional aggregates of cells and small clusters or crowded, molded cells and single epithelial cells showing an increased nuclear to cytoplasmic ratio, irregular nuclear membranes or abnormal chromatin pattern with or without visible intracytoplasmic mucin, and were cells that were not consistent with low grade mucinous dysplasia or gastrointestinal epithelium. Recognizing AEC on cytology in addition to malignant cells improved the sensitivity of the detection of malignancy to 72% from 28% (p<0.0001); specificity decreased to 85% from 100%. If a cyst with moderate dysplasia was considered a true positive, sensitivity decreased to 49% without a significant gain in specificity (88%) although PPV improved from 72% to 87%. In the 101 cysts not malignant on cytology, AEC as a predictor of malignancy had a sensitivity of 61% and specificity of 85% (p<0.0001).

Conclusions: AEC has an 87% PPV for a mucinous cyst with at least moderate dysplasia and predicts malignancy with a specificity of 85%. AEC, even when scant, are important cells to recognize and report for proper risk assessment for malignancy and appropriate consideration for surgical management.

97

Onsite Adequacy Evaluation of Endoscopic Ultrasound-Guided Fine Needle Aspiration of Pancreatic Lesions: Does it Reduce the Rate of Non-diagnostic Results in the Hands of Experienced Operators?

Brant Wang, MD, PhD, Pedro deBrito, MD, Mary Sidawy, MD

Pathology, Georgetown University Hospital, Washington, District of Columbia

Introduction: Immediate onsite evaluation of endoscopic ultrasound-guided (EUS) fine needle aspiration (FNA) of pancreatic lesions by cytopathology personnel is a common practice aimed at reducing the rate of non-diagnostic samples. This study compares the results of EUS guided FNA of pancreas in two settings, one with and another without the benefit of onsite evaluation. We also assessed the additional value of onsite preparation of direct smears versus rinsing the entire sample in CytoLyt®. It is our experience that diagnostic cellular elements in direct smears tend to be partially obscured by gastrointestinal contaminants and blood.

Materials and Methods: We reviewed 381 consecutive EUS-guided FNA of pancreatic lesions performed by two experienced gastroenterologists (April 2004-September 2009). The cases were divided into groups with or without onsite adequacy evaluation. For the former group, cytopathology personnel were present to prepare and evaluate Diff-Quik-stained smears with the remaining material rinsed in CytoLyt®. In addition, the group without onsite adequacy evaluation was divided into two subgroups, one with the entire FNA sample rinsed in CytoLyt®, the other with one air-dried smear per pass prepared by the clinical team with the remaining material rinsed in CytoLyt®. Liquid-based smears and cell blocks were prepared from the CytoLyt® sample. The cytologic diagnoses were reviewed and the non-diagnostic rate of each group was calculated. Contingency tables were constructed and evaluated using chi-square analysis.

Results: Immediate onsite evaluation was provided for 167 cases and was not provided for 214 cases. Of the latter group, 46/214 FNA were entirely rinsed in CytoLyt®. The non-diagnostic rate of the group with onsite adequacy evaluation was 25.8% (43/167), whereas that without the benefit of onsite adequacy evaluation was 24.3% (52/214). The non-diagnostic rate of the subgroup with onsite air-dried smear preparation was 25.6% (43/168), whereas that with the entire sample rinsed in CytoLyt^®was 19.6% (9/46). There were no significant statistical differences in non-diagnostic rates among the different groups. Tables 1 & 2 summarize the results. Table 3 presents the diagnostic categories in both settings.

Conclusions: Our study showed that when experienced gastroenterologists performed the procedure, there was no significant difference in the non-diagnostic rate of EUS-guided pancreatic FNA whether onsite adequacy evaluation was provided or not. In addition, if the pathologist is comfortable interpreting liquid-based preparations, direct smears may be omitted with the entire FNA rinsed in liquid-based medium.

Table 1. Comparison of non-diagnostic rates of EUS-guided pancreatic FNA

Setting	Total	Diagnostic	Non-diagnostic
* Not statistically significant (p = 0.75).
With onsite evaluation	167	124 (74.2%)	43 (25.8%)*
Without onsite evaluation	214	162 (75.7%)	52 (24.3%)*

Table 2. Non-diagnostic rates of pancreatic FNA without onsite evaluation: Value of direct smears

Setting	Total	Diagnostic	Non-diagnostic
* Not statistically significant (p = 0.40).
Direct smears prepared  (remaining sample in CytoLyt)	168	125 (74.4%)	43 (25.6%)*
No direct smears (entire sample in  CytoLyt)	46	37 (80.4%)	9 (19.6%)*

Table 3. Diagnostic categories of pancreatic FNA with & without onsite evaluation

Setting	Total	Adeno- carcinoma	Suspicious for Neoplasm	Other Neoplasms*	Benign Lesions**
* Includes pancreatic endocrine, solid pseudopapillary, mucinous & metastatic neoplasms; squamous carcinoma; adenosquamous carcinoma; malignant fibrous histiocytoma. ** Includes chronic pancreatitis, abscess, benign cyst contents, benign pancreatic tissue, ectopic adrenal, accessory spleen & lymphoid tissue.
With onsite evaluation	124	46 (37.1%)	16 (12.9%)	23 (18.5%)	39 (31.5%)
Without onsite  evaluation	162	63 (38.9%)	19 (11.7%)	22 (13.6%)	58 (35.8%)

98

Diagnostic Accuracy of Endoscopic Ultrasound-Guided Fine Needle Aspiration of Pancreatic Lesions: A 7-Year Experience at One Institution

Haiying Zhang, MD¹, John Cunningham, MD², Achyut Bhattacharyya, MD¹

¹Department of Pathology, University of Arizona College of Medicine, Tucson, Arizona; ²Department of Medicine, University of Arizona College of Medicine, Tucson, Arizona

Introduction: Due to the retroperitoneal location of pancreas, endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) has been widely accepted in the past decade to establish diagnosis for pancreatic lesions and/or for staging. We summarize the utility of EUS-FNA in diagnosing pancreatic lesions, including cytology-histology correlation, over the past 7 years at our institution.

Materials and Methods: An electronic search of all pancreatic EUS-FNA cases at our institution (University Medical Center, Tucson, AZ) from January 2003 to early April 2010 was performed. Clinical and surgical follow-up data were collected. The cytology-histology correlation was evaluated in all cases with surgical follow-up.

Results: In the past 7 years, 136 EUS-FNAs were performed for cytopathological diagnosis of pancreatic lesions. Of the 136 aspirates, 65 (47.8%) were diagnosed as consistent with adenocarcinoma, 10 (7.4%) were neuroendocrine tumor, 1 (0.7%) was lymphoma, 9 (6.6%) were atypical or suspicious for malignancy, 49 (36.0%) were benign, and 2 (1.5%) were inadequate specimens. Forty-nine cases (36.0%) did not have follow-up. Among the 88 cases with clinical follow-up, 33 (24.3% of total) had unresectable carcinoma, 27 (19.9% of total) were clinically benign cases, and 27 cases had subsequent surgical resection or biopsy. Of those 27 cases with surgical follow-up, original cytologic diagnoses were: 15 (55.6%) consistent with adenocarcinoma, 4 (14.8%) neuroendocrine tumor, 6 (22.2%) benign, and 2 (7.4%) suspicious for malignancy. Corresponding histologic diagnoses were: 16 (59.3%) adenocarcinoma, 6 (22.2%) neuroendocrine tumor, 2 (7.4%) metastatic carcinoma, 1 (3.7%) intraductal papillary mucinous neoplasm (IPMN) and 2 (7.4%) other benign lesion. Cytology-histology correlation show sensitivity of cytologic diagnosis for malignant pancreatic lesion is 87.5%, specificity is 100%, and accuracy is 88.9%. A total of 3 false negative cytologic diagnoses were identified per cytology-histology correlation.

Conclusions: Our series demonstrates that EUS-FNA cytology has a very high specificity as well as high sensitivity and high accuracy in diagnosing pancreatic malignant lesions.

Quality Assurance

99

Adherence to Cervical Cancer Screening and HPV Guidelines: Investigation in Vermont

Jessica Bekker, MD, Tim St John, CT(ASCP), Carol Colasacco, MLIS, SCT(ASCP), Gladwyn Leiman, MD

Pathology, University of Vermont-Fletcher Allen Health Care, Burlington, Vermont

Introduction: While cytology and human papillomavirus (HPV) testing have been effective in reducing the incidence of cervical cancer, improper usage of these screening techniques may cause over-investigation and treatment-related morbidity. The American College of Obstetricians and Gynecologists (ACOG) and the American Society for Colposcopy and Cervical Pathology (ASCCP) have established guidelines for the proper use of cervical cytology and HPV testing to limit over-treatment and reduce cost. Updated recommendations from ACOG from December 2009 included advisories that cervical cancer screening begin at age 21, that HPV should not be performed under age 21 (due to the overwhelming rate of resolution of infections in this age group), and that over the age of 30, HPV testing in conjunction with Pap testing could reduce screening frequency. To facilitate better adherence to guidelines, this department circulated an educational bullin detailing current standards and recommendations to providers in Vermont. Despite general publication of the guidelines in the gynecologic and cytologic literature, and the Vermont educational bullin, trends were noted relating to continued improper use of screening methods, including over utilization in some age groups, and under utilization in others.

Materials and Methods: To evaluate the effectiveness of this educational bullin in changing practice, we analyzed cervical cytology and HPV ordering practices in women under 21 and over 30 years of age over two consecutive 3 month periods, one (1/1/2009 to 4/1/2009) before and one (1/1/2010 to 4/1/2010) after distribution of the educational bullin.

Results: The total number of cytological samples obtained from women under 21 over both time periods was analyzed; a 40% decrease after distribution of the educational bullin was noted. Regarding requests for HPV, the initial 3 month period showed 61% of Paps performed on women under age 21 were inappropriately accompanied by an HPV order. Thus, after the educational bullin, this number of inappropriate HPV requests remained essentially unchanged at 65%. In the cohort of women over 30, an appropriate order to perform HPV testing regardless of the cytologic diagnosis accompanied 32% of Paps before the educational bullin; again, there was no significant increase in this number, the order being placed with 36% of Paps after the bullin.

Conclusions: These preliminary analyses suggest that educational information distributed to individual providers reduces Pap tests in women under 21, but does not aid in increasing adherence to national guidelines regarding usage of HPV testing in women under age 21, or those over 30. Other strategies for adherence should be examined.

100

Onsite Adequacy Evaluations Performed by Cytotechnologists: Correlation with Final Cytopathology Reports of 5,793 Image-Guided Fine Needle Aspiration Biopsies

Oname Burlingame, MD, PhD¹, Edmund Cibas, MD¹, Stuart Silverman, MD²

¹Pathology, Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts; ²Radiology, Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts

Introduction: Practice patterns regarding on-site assessment of the adequacy of fine needle aspiration biopsies (FNABs) performed by radiologists vary among laboratories, but in many laboratories primary responsibility rests with the cytotechnologists (CT). Whether performed by CTs, cytopathologists (CPs), or pathology trainees, onsite evaluation provides feedback on the need for additional passes and facilitates triaging of the specimen for ancillary studies. Prior studies have suggested that CTs can perform these evaluations accurately, but the studies were limited by small size. The purpose of this study was to assess the accuracy of CTs on a large scale over a wide range of image-guided FNABs.

Materials and Methods: Ten ASCP-certified CTs provided onsite adequacy assessments of 5,793 image-guided FNABs over the 7-year period 2003-2009, inclusive. Adequacy was assessed on smears stained with a rapid Romanowsky-type stain. A CP provided consultation in a minority (13%) of cases. The accuracy of the adequacy assessment was established by correlation with the final cytologic diagnosis; accuracy was correlated with the body part and the imaging method (e.g., computed tomography, ultrasound). An “accurate” adequacy assessment was defined as an adequate reading on a case that was ultimately reported as satisfactory and unequivocally benign or malignant. An adequate reading on a case that was ultimately reported as unsatisfactory, atypical, or suspicious was deemed “inaccurate.”

Results: Of 5,793 FNABs performed, 3154 (51%) were read as adequate on site. 2956 (94%) of the adequate readings were accurate as judged by correlation with the final cytologic diagnosis, for an error rate of 6%. Of the common biopsy sites sampled, liver specimens demonstrated the highest accuracy of the adequacy assessments (97%) and kidney specimens showed the lowest (90%). Imaging method had no effect on accuracy.

Conclusions: Cytotechnologists are highly accurate in providing onsite accuracy assessment regardless of body part or the method of image guidance.

101

Correlation of Circulating Tumor Cell Enumeration between Cytotechnologists and Pathologists Using the CellSearch^®Assay

Michael Campion, BS, Angela Zieman, CT, MB(ASCP), Jesse Voss, CT, MB(ASCP), Carrie Zimmerman, BA, Jun Zhang, MD, Michael Henry, MD, Kevin Halling, MD, Amy Clayton, MD

Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota

Introduction: The CellSearch® System (Veridex, LLC. Raritan, NJ) is an FDA-approved assay that identifies and enumerates the number of circulating tumor cells (CTCs) in a blood sample. Studies suggest that the number of CTCs is associated with decreased progression-free and reduced overall survival in patients with metastatic breast, colorectal, and prostate cancer. The goal of this study was to determine the correlation, both qualitatively (positive/negative results) and quantitatively (number of CTCs), between cytotechnologists and pathologists when classifying CTCs using the CellSearch® System.

Materials and Methods: From September 2008 through January 2010, 722 blood samples were collected for CTC analysis using the CellSearch® System. Of the 722 samples, complete data was available for 687 (668 samples for metastatic breast and 19 samples for metastatic colorectal). Samples were processed on the CellSearch AutoPrep and images were gathered on the CellSearch Analyzer. Images were reviewed by a trained cytotechnologist and pathologist using the following criteria: 1) nuclear DAPI staining was used to distinguish viable cells from nonspecific debris; 2) circulating tumor cells were identified with phycoerythrin (PE) labeled antibodies to cytokeratins 8, 18, and 19; and 3)leukocytes were identified with allophycocyanin (APC) labeled antibodies to CD45. Cells identified as DAPI+/CD45-/CK+ were considered tumor cells and cells showing DAPI+ /CD45+/CK- were considered leukocytes. Each CTC was recorded and enumerated until the entire sample was analyzed or when 100 CTCs were identified. Specimens with ≥5 CTCs for metastatic breast cancer patients or ≥3 CTCs for metastatic colorectal cancer patients were diagnosed as positive (unfavorable prognosis).

Results: Of the 687 cases, there was a 99% (n=681) agreement in overall results among cytotechnologists and pathologist based on the appropriate cutoff. There was exact agreement in the number of CTCs assigned by the groups in 624 (91%) cases. Linear regression analysis also revealed a high level of agreement (R²=0.995). (Figure 1) The remaining 63 (9%) cases had a median difference of 1 CTC (range 1-8 CTCs) between the two analyses. Fifty-seven of those 63 (90%) still resulted in a similar overall result. However, the remaining 6 cases (<1% of all cases), had a difference in CTC counts large enough to change the overall result, thus potentially changing patient prognoses. In each of these 6 cases the pathologist classified a higher number of images as CTCs, resulting in a change from a negative (favorable prognosis) to positive (unfavorable prognosis) result based on the appropriate cutoff.

Conclusions: Data from this study shows that there is a very high correlation between cytotechnologists and pathologists when classifying CTC images using the CellSearch® assay. This suggests that cytotechnologists can provide accurate results with the CellSearch® assay and may be utilized to increase pathologist efficiency by pre-evaluating CellSearch® specimens.

102

Whole Slide Imaging in Gynecologic Cytology: A Feasibility Study Showing Potential for Use in Telecytology and Proficiency Testing

Ivan Chebib, MD, FRCPC, David Wilbur, MD

Cytopathology Unit, James Homer Wright Pathology Laboratories, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts

Introduction: Whole slide images (WSI) have the potential to recapitulate an entire glass slide in digital format which can facilitate remote interpretation. We assessed the feasibility and interpretability of internet access of WSI of cytology specimens as a potential method for telecytology and online proficiency testing (PT).

Materials and Methods: Ten liquid-based cervical cytology specimens were each scanned twice using a Mirax MIDI whole slide scanner with 20x scanning objective lens (Carl Zeiss, Jena, Germany): first as a single focal plane scan and again with integration of 8 focal planes into a single plane final image. The 10 slides were selected in a manner to recapitulate a standard CLIA-mandated PT event. The images were imported to a custom electronic medical database program (Corista, Concord, MA) where patient information was integrated with each WSI and stored on a server. The images and patient information were retrieved by internet remote access and the scans reviewed in a masked fashion by one author (IC). We assessed the ability to interpret the WSI scans, the comparative quality of the 2 scan modes, and the review/screening time. General observations about system features were gathered.

Results: The reference diagnoses for the 10 slides included negative for intraepithelial lesion (NILM) with repair, Candida spp, Trichomonas vaginalis, or Herpes simplex virus; low grade squamous intraepithelial lesion (LSIL); high grade squamous intraepithelial lesion (HSIL); endometrial adenocarcinoma (EMC); and unsatisfactory.

In slides scanned with 1 and 8-planes, 60% and 80% of cases, respectively were correctly classified compared to the reference diagnosis. There was more difficulty identifying microorganisms and assessing cell groups in 1-plane scans which were, in general, less well-focused. The 2 cases misinterpreted in the 8-plane slides include NILM with HSV, interpreted as atypical cells cannot exclude high grade (ASC-H), but the presence of HSV was identified; and a NILM-reactive case was interpreted as ASC-US. In the 1-plane scans, the EMC was called NILM, the HSIL was called ASC-US, an LSIL was called ASC-US, and the NILM with HSV was called ASC-H.

The average time for review of the cytology slides, was 12.2 min (range 3 - 23 min). There was a slight decrease in average review time for 8-plane slides (11.1 min) compared with 1-plane slides (13.2 min). There was also a trend of decreasing review time from the first to last cases, likely representing improved experience. Review time was dependent on cellularity and number of three dimensional groups.

Although scanning with 8-planes improved the cytologic details, especially in larger cell groups, there was still a lack of crisp detail at 40X objective equivalent magnification, even with 8-plane scanned slides. The 1-plane scans often contained out-of-focus diagnostic groups and in one scan, lacked large groups altogether.

Conclusions: This pilot study illustrates that remote interpretation internet telecytology with WSI of cervical cytology specimens is feasible and could be used for online proficiency testing. Multiplane scanning enhances focus which improves performance. Use of validated WSI without atypical responses would likely improve PT performance. Higher resolution initial scanning (at least 40X) is needed to improve high-power cytologic detail.

103

Screening High Grade Squamous Intraepithelial Lesions and Invasive Carcinoma of the Cervix Out of the Community: A Baseline Study to Permit Assessment of Newly Adopted Changes in Screening Policies

Carol Colasacco, MLIS, SCT(ASCP), CT(IAC), Gladwyn Leiman, MBBCh, FIAC, FRCPath, Tim St. John, BS, CT(ASCP)

Pathology and Laboratory Medicine, Fletcher Allen Health Care, Vermont

Introduction: The Papanicolaou (Pap) test has played an integral role in the declining incidence of cervical carcinoma. Efforts to further reduce this incidence have been supported by the development of liquid-based cytology, automation, HPV testing and an HPV vaccine. The American Congress of Obstetricians and Gynecologists' (ACOG) 2009 recommendations for less frequent Pap tests have raised understandable concerns within the cytopathology community. Information regarding the effectiveness of cervical cancer screening in the previously well-screened low-risk northern Vermont communities served by Fletcher Allen Health Care (FAHC) might satisfy a twofold purpose by both measuring the effectiveness of prior screening in this relatively stable population, and creating a baseline that will be useful in evaluating the impact of recent changes in cervical cancer screening and prevention protocols, including screening recommendation changes, hrHPV testing, and HPV vaccine utilization.

Materials and Methods: Cytopathology statistical records since 1995 were reviewed, and rates of LSIL, LSIL-H, HSIL, and malignant diagnoses were assessed. The annual numbers of malignant Pap test cases at FAHC since 1973 were also determined. The Anatomic Pathology CoPath database was searched for SNOMED codes M74006 (CIN I), M74007 (CIN II, LSIL-H), M74008 (CIN III, HSIL), and a collection of SNOMED codes used to code malignant cases in order to identify the numbers of pertinent biopsy cases (cervical biopsies, cervical LEEP, cervical cone, endocervical biopsy/curettage, and vaginal biopsies) coded with those diagnoses. Hysterectomy cases were not included, on the presumption of prior diagnostic biopsies. Ten years of biopsy results were available for review.

Results: Since 1995, the rates of HSIL and malignant diagnoses for Pap test cases submitted to this laboratory have exhibited a definite downward trend, while the rate of LSIL has trended strongly upward (Figure 1). LSIL-H diagnosis rates have remained fairly stable. In addition, malignant cases have decreased dramatically over the past 35+ years, from a high of 0.36% in 1975 to a low of only 0.01% in 2007 (Figure 2). Similar to Pap samples, surgical pathology cases with a malignant or CIN III diagnosis have exhibited a definite downward trend since 2000. Histologic diagnoses of CIN I and CIN II (LSIL-H) have shown an increasing trend (Figure 3).

Conclusions: The contrast between the upward trend of LSIL and the decrease in LSIL-H, HSIL, and malignant cases is noteworthy and points to the success of continued routine screening in this previously well-screened low-risk cervical cancer population in northern Vermont. Liquid-based preparations and imaging modalities may have contributed to the current situation in which high grade and invasive lesions are being progressively reduced in this population. In light of the recent changes in screening recommendations released by ACOG, baseline data such as compiled in this study will be essential to properly evaluate the impact which decreased frequency of cervical cancer screening, in combination with the utilization of the HPV vaccine and hrHPV testing, has on the incidence of cervical neoplasia in the future.

104

Establishing a Laboratory Baseline Atypical Glandular Cell Diagnostic Category Frequency Using Correlation and Error Root Cause

Heather Currens, MAED, Stephen Raab, MD

Cytopathology Section, University of Colorado Denver Department of Anatomic Pathology, Aurora, Colorado

Introduction: ASCCP guidelines recommend colposcopic examination and endocervical sampling (endometrial sampling for specific age groups) for the Bethesda System diagnosis of Atypical Glandular Cells (AGC). Published literature indicates that approximately 9-38% of women diagnosed with AGC have clinically significant histologic follow-up, although individual practices of cytotechnologists and cytopathologists may have considerably more variable clinically significant follow-up frequencies. Our goal was to determine if: overall laboratory and individual personnel had similar AGC histologic clinically significant follow-up frequencies; causes of over/under diagnosis of AGC; optimal AGC frequency considering implications of over/under diagnosis of AGC; and implementation strategies to improve the sensitivity and specificity of AGC.

Materials and Methods: A retrospective correlative review of all cases diagnosed as AGC over a 24 month period (145 total Pap tests) was performed. AGC cases were subclassified by Bethesda System diagnostic categories (AGC, AGC + high grade squamous intraepithelial lesion (HSIL), AGC + atypical squamous cells, cannot rule out HSIL, etc.). Frequency of AGC diagnostic subcategory by cytopathologist and cytotechnologist, patient age, menstrual status, parity, clinical history, histologic and/or Pap test follow-up, and the presence of endometrial cells were evaluated. Frequency of all Pap tests diagnosed as adenocarcinoma and negative Pap tests with adenocarcinoma follow-up during the same time period (Sams et al. Cancer Cytopathol. 2009;117(Supplement 1):416-417) was determined. A cytotechnologist and cytopathologist independently reviewed the slides, rendered a review diagnosis, and performed root cause analysis by assessing specimen quality, presence of glandular cells and 15 latent or active factors potentially contributing to error (reactive/reparative features, staining quality, familiarity with liquid based preparations, etc.). Based on contributing factors leading to over/under diagnosing AGC, improvement initiatives to render more accurate AGC diagnoses were designed.

Results: AGC frequency was 1.58% (range by pathologist, 1.0-4.5%) with 64.4% of AGC cases having histologic follow-up. Frequency of clinically significant histologic follow-up was 15.1% (range by pathologist, 0-20%). Root cause analysis showed main causes of over diagnosis of AGC being inexperience, reactive/reparative changes, lack of clinical history, presence of benign endometrial cells, and failure to use published criteria. Causes of under diagnosis were technical quality, inexperience, and failure to use published criteria. Improvement initiatives included reeducation in AGC criteria and double viewing all AGC cases prior to sign-out.

Conclusions: Our laboratory was not standard in the diagnosis of AGC, both over and under diagnosing AGC in specific circumstances. Main factors contributing to this phenomenon were personnel using different diagnostic criteria, lack of group consensus decision making, and interpreting specimens of poor quality. Benchmarking AGC diagnoses with follow-up allows laboratories to assess group and individual performance and target problems that lead to over and under diagnosis.

105

Unsatisfactory Pap Tests: A Performance Improvement Project

Megan DiFurio, MD¹, Michael Sundborg, MD²

¹Department of Pathology, Womack Army Medical Center, Fort Bragg, North Carolina; ²Department of Obstetrics and Gynecology, Womack Army Medical Center, Fort Bragg, North Carolina

Introduction: The importance of an adequate Pap test cannot be overstated. An inadequate test may delay diagnosis. Health care providers (HCPs) are expected to proficiently perform a Pap test upon completion of medical training, just as assessing adequacy of a Pap test is a skill cytotechnologists are expected to perform upon completion of their education. Liquid-based cytology (LBC) is associated with lower unsatisfactory rates versus conventional cytology due to the diminished effects of obscuring inflammation, blood and air-drying. Adjunctive methods, such as the use of glacial acetic acid, have been developed to further reduce the incidence of unsatisfactory Pap tests. While these pre-analytic variables are critical, having well-defined criteria for determining adequacy, such as The Bethesda System, is just as important for accurately determining Pap test adequacy. According to benchmark data, published by the College of American Pathologists, unsatisfactory reporting rates for Pap tests using ThinPrep® should range from 0.1-3.4%; laboratories that fall outside this range should investigate the reasons.

Materials and Methods: At the time of new directorship of a cytology laboratory located at a military hospital facility, it was noted that the Pap test unsatisfactory rate over the previous years (2002-2008) had increased from 1.6% to 10.7%. During this time period and currently, all the Pap tests interpreted by the laboratory were/are collected in ThinPrep® vials and processed using the ThinPrep®-2000 or -3000. Most of the strategy during these 5 years focused on retraining of the collecting HCPs. In June 2009, a targeted performance improvement plan to examine the Pap test unsatisfactory rate was implemented. This plan included the following: (1) the use of glacial acetic acid recovery for bloody specimens, (2) retraining of the HCPS at the five submitting clinics [conducted by representatives of Hologic LP (Marlborough, MA) annually from Sep-05 to Sep-09], (3) prospective review of all unsatisfactory Pap tests by the Cytology Medical Director (CMD) for a one month period and (4) determination of the criteria being used by the cytotechnologists for determining specimen adequacy and implementation of a standard method to be used by all cytotechnologists and pathologists.

Results: The annual unsatisfactory rates for the years 2002-2009 are shown in Table 1. Prior to June 2004 and after August 2009, the CMD was board certified in cytopathology. The monthly unsatisfactory rates from April 2009 - March 2010 are shown in Table 2. Glacial acetic acid recovery began in June 2009. The CMD review was conducted from 17 September to 17 October 2009. A standard method for determining adequacy was implemented in November 2009. Retraining of the HCPs occurred at various times during the time period indicated above, but no retraining occurred after September 2009.

Conclusions: While all the interventions may have played a significant role, it appears that the implementation of a standard method for determining adequacy had the greatest influence on improving the unsatisfactory rate. The use of glacial acetic acid and prospective review by the CMD also had considerable effects. It is unclear from these results if the clinician retraining played a significant role. This data also suggests that having a CMD board certified in cytopathology, versus a general anatomic pathologist, may be advantageous.

Table 1. Annual Unsatisfactory Rates (2002-2009)

2002	2003	2004	2005	2006	2007	2008	2009
1.6%	2.4%	3.1%	4.8%	5.1%	5.8%	10.7%	8.5%

Table 2. Monthly Unsatisfactory Rates (Apr-09 - Mar-10)

Apr-09	May-09	Jun-09	Jul-09	Aug-09	Sep-09	Oct-09	Nov-09	Dec-09	Jan-10	Feb-10	Mar-10
11.8%	8.0%	7.1%	8.9%	9.1%	8.4%	7.4%	3.4%	0.7%	1.4%	0.8%	1.4%

106

In the Blink of an Eye: Discrimination and Localization of Abnormalities in Cervical Cytology Screening from a Global Signal

Karla Evans, PhD¹, Jeremy Wolfe, PhD¹, Rosemary Tambouret, MD², David Wilbur, MD²

¹Brigham and Women's Hospital, Boston, Massachusetts; ²Massachusetts General Hospital, Boston, Massachusetts

Introduction: Understanding complex visual scenes involves rapid processing of global structural and statistical regularities of the image as well as scrutiny of the details. There is some evidence in the radiology literature showing that expert radiologists appear to use this “gestalt” mode of processing when interpreting mammograms (Kundel & Nodine, 1975; Kundel, Nodine, Conant & Weinstein, 2007). Here, we investigate if such an effect can be documented in cytology screening. Are cytologists able to detect subtle signs of cytologic abnormalities under brief visual exposure conditions? We test the hypothesis that a global signal, rather than chance detection of local features is the basis for this ability.

Materials and Methods: Field of view micrographs of cervical smears were briefly presented (250 or 1000 msec). We used 60 images considered to be normal and 60 images that have abnormal cells. Presence of abnormality was randomized across trials. Thirty-eight volunteer cytologists (cytopathologists and cytotechnologists) at a national meeting used a 0-100 analog rating scale to report their inclination to call back the case for further review. Immediately after rating a case, observers marked the location on the screen that they considered most likely to contain pathology.

Results: Performance of cytologists was above chance at both stimulus durations for detecting these subtle abnormalities. Performance was significantly better at 1000 msec (D' = 1.44) than at 250 msec (D' = 1). However, the ability to localize the abnormalities was quite poor with correct localization of only 12% of abnormal regions of interest. Localization was similar at both durations.

Conclusions: These findings are the first demonstration of an ability to discriminate normal and abnormal images in a single, brief glance. The failure to localize these abnormalities suggests that performance is based on a distributed/global signal and not on fortuitous attention on a target location. Of course, clinical screening requires slower scrutiny and cannot be based on this global signal alone. However, if this signal for abnormality could be identified, it could be very valuable in teaching and/or automated screening approaches.

107

Target Prevalence Influences Cytologists' Error Rates

Jeremy Wolfe, PhD¹, Karla Evans, PhD¹, Andrew Evered, MD³, Rosemary Tambouret, MD², David Wilbur, MD²

¹Brigham and Women's Hospital, Boston, Massachusetts; ²Massachusetts General Hospital, Boston, Massachusetts; ³Cervical Screening Wales, Cardiff, Wales, United Kingdom

Introduction: Some medical image-based screening tasks are difficult visual searches with low target prevalence. Studies in laboratory settings have shown that when targets are present in a low percentage of images (low “prevalence”), observers produce a high rate of false negative errors and a low rate of false positives compared to high prevalence. This prevalence effect represents a shift in response criterion and not a failure of attention to the tasks.

Materials and Methods: We measured the effect of prevalence in two groups of cytologists searching for evidence of cervical cancer in Pap smear stimuli. One group consisted of 10 cytologists from Boston, MA, USA. Another group of 12 was from Cardiff, Wales, UK. These experts evaluated images of cells at low prevalence of abnormality (2% - Boston or 5% - Wales) or higher prevalence (50%). In Boston, experts made normal/abnormal decisions using a 4-point rating scale. In Wales, experts used the mouse to localize apparent abnormalities in addition to giving a rating.

Results: Both groups produced clear evidence of a prevalence effect. Miss errors were 30% at low prevalence and 17% at higher prevalence in the simple rating condition (Boston). They were 42% at low prevalence and 27% at higher prevalence in the localization condition (Wales). Note that the groups cannot be compared to each other because different stimuli were used.

Conclusions: We have evidence that expert cytologists are not immune to the effects of disease prevalence. While we would not claim that prevalence is a cause of medical error, it may be important to consider the effects of prevalence on the performance of medical screening tasks.

108

Have We Met Before? How Good Is Cytologists' Recognition Memory For Microscopic Images?

Rosemary Tambouret, MD¹, Karla Evans, PhD², Jeremy Wolfe, PhD², David Wilbur, MD¹

¹Massachusetts General Hospital, Boston, Massachusetts; ²Brigham and Women's Hospital, Boston, Massachusetts

Introduction: Humans are remarkably good at recognizing if a scene or an object has been previously viewed. Excellent recognition memory is found with large sets of scenes and quite similar objects. Cytologists often claim that they can easily identify images that they have seen before. Is this claim true? Do expert cytologists develop robust visual long-term memory for their domain of expertise? If so, is this expertise specific to the trained image class, or do such experts possess generally superior visual memory?

Materials and Methods: We tested recognition memory of 10 cytologists, and controls with no medical experience for three classes of visual stimuli: isolated objects, scenes, and cervical microscopic images. The experiment was composed of study and test phases for each of three stimulus sets. In the study phase, each observer saw a randomly chosen subset of 72 images and was asked to try to memorize them. Each study image was presented on a computer display for 3 seconds. In the test phase, immediately after the study phase, observers saw a sequence of 72 images, 36 of which were old images from the study phase while the remaining 36 were new images. Half of the cervical microscope images showed some type of abnormality. Observers simply labeled each stimulus as old or new.

Results: Cytologists were modestly better than control observers at recognizing cervical microscope images (D' .66 vs .16), but their performance with these was much worse than their performance with scenes (D' 2.05) or objects (D' 3.48). The presence of abnormality did not influence memory. Furthermore, while experts did well with scenes and objects, they did no better than control observers.

Conclusions: Expertise with microscope images does not produce an impressive memory for those stimuli nor do these domain experts appear to have better visual recognition memory than the general public.

109

Immunohistochemistry Practices of Cytopathology Laboratories

Andrew Fischer, MD¹, Mary Schwartz, MD², Malcolm Schinstine, MD³, Ann Moriarty, MD⁴, Amy Clayton, MD⁵, David Wilbur, MD⁶, Rhona Souers, MS⁷, Lisa Fatheree, MS⁷

¹Pathology, University of Massachusetts, Worcester, Massachusetts; ²Pathology, The Methodist Hospital, Houston, Texas; ³Pathology, Hilo Medical Center, Hilo, Hawaii; ⁴Pathology, AmeriPath Indiana, Indianapolis, Indiana; ⁵Division Anatomic Pathology, Mayo Clinic, Rochester, Minnesota; ⁶Pathology, Massachusetts General Hospital, Boston, Massachusetts; ⁷College of American Pathologists, Northfield, Illinois

Introduction: Immunohistochemistry (IHC) is an important tool for cytologists, facilitating diagnosis and helping predict therapeutic response on minimally invasive biopsies. IHC on cytology samples can pose a challenge to laboratories because fixation and processing conditions may differ from the conditions used for IHC positive controls. There is little data on the relative numbers of various types of fixation and processing conditions (“platforms”) for cytology IHC, or whether recurring problems exist for certain platforms or antigens.

Materials and Methods: In November 2009, a voluntary survey of cytology IHC practices was sent to 1899 cytology laboratories participating in the College of American Pathologists Non-Gynecologic Cytopathology education program.

Results: A total of818 laboratories (43%) responded to the survey. Three hundred forty five (44%) of responding cytopathology laboratories perform IHC, with a total annual estimate of 146,376 IHC tests. Laboratories performing cytology IHC had a median of 50 antibodies available for testing. Seventeen different fixation and processing “platforms” prior to antibody reaction were reported. Formalin-fixed cell blocks was the most common principal platform used by labs (25.5%) followed by cell blocks of material fixed in ThinPrep media and post-fixed with formalin (19.1%) (Table 1). In 53.3% of labs, formalin is not included in their principal platform; 50.8% of labs do not use paraffin-embedded material as their principal platform. Cytology platforms differed at least sometimes from the platform of their positive controls in 59.2% of laboratories, yet 98.6% reported using the same antibody dilutions and 85.9% the same antigen retrieval method for both cytology and surgical pathology specimens. When asked to name antibodies for which the antibody dilution or antigen retrieval conditions differed between cytology and surgical samples, there were 18 responses distributed amongst 14 antibodies, with none of the 14 antibodies listed more than twice. Neither progesterone receptor nor Her2 was listed as requiring an adjustment for cytology platforms. Four laboratories reported performance of a Her2 validation of a non-formalin platform. Three of 3 laboratories successfully validated Her2 on alcohol-fixed samples and one laboratory was unsuccessful in validating Her2 on air-dried smears.

Conclusions: A wide variety of fixation and processing platforms are used for IHC on cytology samples. Most laboratories use a platform that differs from the platform for their positive controls, but antibody concentrations and antigen retrieval conditions are uncommonly adjusted for cytology samples. This finding suggests that the platform for cytology IHC does not affect IHC results, or laboratories have difficulty determining whether different cytology platforms affect IHC results. Except for the apparent unsuitability of air-dried smears for Her2 testing noted by one laboratory, the survey does not disclose evidence of a systematic problem with any one antibody or platform for cytology IHC.

Table 1. Principal fixation and processing platform used for cytology IHC

Platform	Number labs (%)
*PECB is abbreviation for paraffin-embedded cell block.
Formalin used as primary fixative PECB*	87 (25.5)
Fixed in ThinPrep media, post-fixed in formalin, PECB	65 (19)
Direct smears fixed with 95% alcohol	58 (17)
Liquid-based prep fixed with CytoLyt®/PreservCyt®	37 (10.9)
Direct smears, air-dried	26 (7.6)
Cytospin fixed with alcohol	22 (6.5)
Other	18 (5.3)
Liquid-based preparation fixed with CytoRich®  Red/SurePath fixative only	9 (2.6)
PECB from material primarily fixed with ThinPrep  media without any formalin	7 (2.1)
Cytorich® red followed by formalin PECB	7 (2.1)
Air-dried cytospin	5 (1.5)

110

Clinical Utility of FNA-Derived Cell Block When Compared to Simultaneously Collected Core Biopsy

Jessica VanDyke, BS¹, Sharon Davis-Devine, BS¹, Erica Mentock, BS¹, Travis Blunier, BS¹, Farah Gaudier, MD¹, Gregory Freund, MD²

¹Pathology and Laboratory Services, Carle Foundation Hospital, Urbana, Illinois; ²Pathology, University of Illinois, Urbana, Illinois

Introduction: Traditionally, the gold standard for the diagnoses of pathologic tissues lesions was excisional biopsy. More recently, core biopsies have been used as a less invasive method of tissue collection relative to excision. Currently, diagnostic flexibility has been enhanced with dual collection of core biopsies and fine needle aspiration biopsies (FNA). In this setting, FNA is often viewed as redundant in that FNA smears do not easily lend themselves to ancillary testing such as immunohistochemical (IHC) staining which is required for tumor typing/tissue of origin (TT/TON) information. Therefore, the goal of this study was to determine if FNA combined with FNA-derived cell block (FNA + cell block) was as effective as core biopsy in rendering TT/TON information.

Materials and Methods: FNA cases from January 2009 to December 2009 were reviewed (n=837). Study inclusion criteria was designated as all non-thyroid FNA cases (n=435). Final study inclusion was based on 1) existence of a core biopsy/FNA, core biopsy/FNA+ cell block or FNA + cell block alone, and 2) IHC staining on either the core biopsy and/or cell block. IHC was deemed required for TT/TON information (n=144). TT/TON-inclusive diagnoses were compared between core biopsies and FNA + cell block.

Results: In 11% (17/149) of cases the core biopsy was used for IHC because a cell block was not performed, 76% (13/17) of these cases possessed a TT/TON-inclusive diagnosis. In 40% (58/149) of cases the cell block was used for IHC because a core biopsy was not performed, 64% (37/58) of these cases possessed a TT/TON-inclusive diagnosis. In 50% (74/149) of cases both a core biopsy and cell block were collected. In 61% (45/74) of these cases the core biopsy was used for IHC. In 24% (18/74) of these cases the cell block was used for IHC. In 15% (11/74) of these cases both core biopsy and cell block were used for IHC. Finally, in cases where both core biopsy and cell block were available for IHC, a TT/TON-inclusive diagnosis was made 64% (29/45) on the core biopsy and 83% on the cell block (15/18).

Conclusions: Core biopsies are chosen more frequently than cell blocks for the purpose of IHC staining. However, when a cell block and a core biopsy are present, a TT/TON-inclusive diagnosis is more likely to result from a cell block.

111

Assessment of Variant Urine Fixatives for the UroVysion® Fluorescence In-Situ Hybridization Assay

Sarah Kerr, MD, Jesse Voss, CT, MB(ASCP), Michael Campion, BS, Carrie Zimmerman, BA, Angela Zieman, CT, MB(ASCP), Kevin Halling, MD, PhD

Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota

Introduction: The UroVysion® (Abbott Molecular, Inc., Des Plaines, IL) fluorescence in situ hybridization (FISH) assay for detection of bladder cancer in urine cytology specimens is often performed by high-volume reference laboratories. Although laboratories generally establish a preferred common urine fixative during assay validation, reference laboratories often receive urine in a variety of fixatives from their clients. Laboratories must decide to either reject specimens submitted with unvalidated fixatives, report results with a disclaimer statement, or validate the assay for a large number of variant fixatives. This study evaluates FISH quality measures for our most common fixative (PreservCyt®, approximately 75% of test volume) and nine other urine fixatives commonly received in our laboratory.

Materials and Methods: Five healthy volunteers donated voided urine; 100 mL of each sample was pooled into a common sample. The pool was then split into ten aliquots. Each 50 mL aliquot was mixed in a 1:1 ratio with fixative and stored at room temperature for 72 hours. Fixatives used were 95% ethanol, 50% ethanol, RPMI-1640 (Sigma Aldrich, St. Louis, MO), 2% Carbowax® (Sigma Aldrich, St. Louis, MO), CytoLyt® (Hologic, Inc., Bedford, MA), Surepath™ (BD Diagnostics - TriPath, Burlington, NC), CytoRich® Red (Thermo Scientific, Waltham, MA), Methanol, Isopropanol, and PreservCyt® (Hologic, Inc., Bedford, MA). Two slides were prepared from each fixative for UroVysion FISH and were evaluated by an experienced cytotechnologist as performed during clinical practice. Each slide was also scored on a five point scale for the following hybridization quality parameters: signal intensity (1=none/weak, 5=very strong), nuclear probe specificity/crispness (1=very poor/extensive non-specific binding, 5=excellent/clean), background signal (1=extensive non-specific probe hybridization/dirty, 5=none/clean), presence of bacteria (1=extensive/obscuring, 5=none), presence of fungus (1=extensive/obscuring, 5=none). An overall hybridization score was calculated by averaging all scores from the two slides for each fixative.

Results: All slides were diagnosed as negative (disomic). Hybridization quality parameter scores are shown in Table 1. All fixatives had an average overall score ranging from 4.2 - 4.4 out of 5, with the exception of RPMI (average 3.5). The inferior result with RPMI was largely attributed to obscuring bacteria (score=1). See Figures 1 and 2 for microscopic examples of RPMI and PreservCyt® fixed specimens.

Conclusions: Laboratories may be able to safely expand the range of urine fixatives accepted for the UroVysion assay; however, acceptance of variant fixatives requires additional validation experiments. Given average transportation times, RPMI does not seem to have appropriate antibacterial activity for urine-based FISH assays in the reference laboratory setting.

Table 1. Hybridization Quality Scores by Fixative for UroVysion FISH

Fixative	Signal Intensity	Nuclear Probe Specificity/ Crispness	Background Signal	Bacteria	Fungus	Overall Averagefor Fixative
Each hybridization score reflects the average score of duplicate slides for each fixative. Quality measures were scored on a scale from 1 (poor) to 5 (best) for each measure. See text for details.
95% Ethanol	4	4	4.5	4	5	4.3
50% Ethanol	4	4	5	4	5	4.4
RPMI	4	3	4.5	1	5	3.5
2% Carbowax®	4.5	4	4.5	4	5	4.4
CytoLyt®	4.5	3.5	3.5	3.5	5	4
Surepath	5	3.5	4	3.5	5	4.2
CytoRich® Red	4	4	5	4	5	4.4
Methanol	4	4.5	4.5	4	5	4.4
Isopropanol	5	4	4.5	3.5	5	4.4
PreservCyt®	5	3	4	4	5	4.2
Average for Quality Measure	4.4	3.75	4.4	3.55	5

Figure 1. UroVysion FISH field demonstrating specimen preserved in RPMI.

Figure 2. UroVysion FISH field demonstrating specimen fixed in PreservCyt.

112

Comparison of Rapid Prescreening and Automated FocalPoint™ Primary Screening for Abnormal Gyn Detection

Gail Mueller, MS, CT(ASCP)DLM¹, Heather Wilgenbusch, BA,CT,MP(ASCP)¹, Margaret Neal, MD¹, Andrew Renshaw, MD²

¹Ketchum, Wood & Burgert Pathology Associates, Tallahassee, Florida; ²Baptist Hospital of Miami, Miami, Florida

Introduction: We have previously shown that rapid prescreening (RPS) is just as sensitive as FocalPoint™ (BD Diagnostics) quality control rescreening in identifying abnormal liquid-based gynecologic cases. Studies comparing RPS to FocalPoint primary screening (FocPt) have not been performed.

Materials and Methods: A total of 5144 cases (SurePath™ Pap Test, BD Diagnostics) were subjected to rapid prescreening (RPS), FocalPoint™ primary screening (FocPt) and subsequent routine manual screening (FS). Three categories of cases diagnosed as normal by routine manual screen were rescreened to obtain a final diagnosis: cases abnormal by rapid prescreen (RPS), cases in the top 1.5 quintile ranking by FocalPoint™ (FocPt), and cases designated as high risk. The threshold for abnormal was ASCUS. A blinded re-review was conducted on ASCUS cases from the set of abnormal cases identified by RPS alone, FocPt alone and high risk rescreen alone. RPS and FocPt were blinded, but FS was performed with knowledge of the top 1.5 quintile FocPt ranking and patient history. Abnormal cases were recorded by diagnosis and FocPt quintile.

Results: A total of 490 abnormal cases were indentified by full screen (FS), 2 by RPS, 12 by FocPt, 8 by high risk rescreen, and 2 by RPS, FocPt and QC but not by FS. The sensitivity of full screening was 95%. The sensitivity of FocPt primary screening was significantly higher than RPS (58% versus 43%); however, the false positive rate was nearly 8 times greater than RPS. 55.3% of the abnormal cases were in the top 1.5 quintiles. The remainder of the abnormal cases (44.7%) fell into the lower quintiles, with seven cases occurring in the “No Further Review” category. No cases in that bottom category were archived. The blinded review showed a reproducibility rate of 30% to 41%.

Conclusions: FocalPoint™ primary screening is significantly more sensitive than rapid prescreening, but achieves this result by requiring full manual review of nearly 8 times as many negative cases. The individual slide cost of the FocalPoint™ instrumentation is an important consideration when evaluating quality assurance methods.

113

Comparison of Standardized Terminology and Quality Assessments on Thyroid Fine Needle Aspiration Accuracy

Sharon Sams, MD, MPH, Stephen Raab, MD

Pathology, University of Colorado Health Sciences Center, Aurora, Colorado

Introduction: Standardized diagnostic terminology (ST) and specimen quality assessments (SQA) have been reported to improve the accuracy of thyroid gland fine needle aspiration biopsy (FNA). Prior to this study, our practice was not standard and thyroid gland FNA specimens were interpreted variably, sometimes using ST and/or SQA. We examined how the standardization of both of these components affected the sensitivity and specificity of diagnosis in a single practice.

Materials and Methods: We performed a two year retrospective cytologic-histologic correlation of all complete and partial thyroidectomies performed at our institution that had one or more preceding thyroid gland FNA specimens with immediate interpretation. We classified each FNA into one of three groups: 1) no ST or SQA, 2) ST only, or 3) ST and SQA. We used a ST similar to the Bethesda System for Reporting Thyroid Cytopathology recommendations and our SQA consisted of a quantitative assessment based on sample cellularity, background quality, and cellular preservation. Cases without a ST generally had a descriptive diagnosis, and on review, we classified these into the most appropriate ST category. We calculated the sensitivity and specificity for each of the three groups and performed root cause analysis to determine the causes of false negative and false positive error. Benign conditions were considered disease negative and all neoplasms and malignancies were considered disease positive in the calculation of sensitivity and specificity. A diagnosis of benign, non-specific or unsatisfactory was considered diagnostically negative as the next clinically appropriate step would either be observation or repeat biopsy. The categories of follicular lesion, neoplasm, suspicious or malignant were considered diagnostically positive.

Results: The sensitivity and specificity of Groups 1 (no ST or SQA); 2 (ST only); and 3 (ST and SQA) were 62% and 50%; 92% and 27%; and 76% and 75%, respectively. The fewest and most indeterminate diagnoses were rendered in Group 3 and Group 1, respectively. More benign lesions were surgically excised if ST or SQA were not used. False negative errors in Group 3 cases generally were secondary to patient related factors (e.g., follicular neoplasms had cytomorphologic characteristics of benign lesions) and no definitive neoplastic diagnoses in Group 3 were false positive errors. False positive diagnoses in Groups 1 and 2 generally were related to poor specimen quality.

Conclusions: The use of both ST and SQA produced the best trade-off of thyroid gland FNA sensitivity and specificity. The use of SQA markedly reduced false positive errors in definitive diagnostic categories and shifted diagnoses from indeterminate to determinate categories. False positive errors occurred in poor quality specimens or if diagnostic terms were unclear and not standardized. A proportion of false negative errors are related to follicular neoplasms that resemble benign entities. Quality improvement initiatives that decrease the use of indeterminate diagnoses and improve quality would further improve thyroid gland FNA diagnostic accuracy.

114

Utilization of ASC-US/SIL Ratio and Percent Positive High-Risk HPV as Performance Measures

Marla Taylor, BS, Diane Carr, BS

Cytopathology, Scott & White Healthcare, Temple, Texas

Introduction: The performance measure for pathologists has traditionally relied on cytologic to histologic correlation. However, the diagnostic category of ASC-US presents a challenge in that it is a diagnosis of uncertainty. To help overcome this performance measure challenge, most labs utilize the ASC-US/SIL ratio, which recommends that the rate of ASC-US be no more than 3 times that of SIL. Recently it has been suggested that the percent of high-risk HPV (HR HPV) Positive ASC-US cases could be used as a performance measure for ASC-US interpretation. Based on the ASC-US/LSIL Triage Study (ALTS) data, ASC-US cases generally yield a 40 - 50% HR HPV Positive rate. In 2005, a CAP Q-Probe was performed which resulted in a mean of 43.74% HPV positive results. Our goal was to evaluate the data obtained for 6 pathologists over 6 years using both the ASC-US/SIL ratio and the percent of HR HPV positivity.

Materials and Methods: A computer search was performed to identify all Pap test cases with a diagnosis of ASC-US and Positive for HR HPV result and ASC-US/SIL ratio was calculated for each pathologist from January 2004-December 2009.

Results: The result for the percentage of HR HPV positivity for ASC-US by pathologist is displayed in Table 1. High risk HPV positivity ranged from 29% to 62%. The 6 year mean for each pathologist ranged from 36% to 57% with a combined mean of 45% and a SD of ± 8. Pathologist 1 had the highest 6 year percentage at 57% whereas pathologist 4 had the lowest at 36%. The result of the ASC-US/SIL ratio for each pathologist is displayed in Table 2. The ASC/SIL ratio for all 6 pathologist ranges from 0.3 to 1.3. The 6 year mean for each pathologist ranged from 0.4 to 0.9 with a combined 6 year mean of 0.7 and a SD of ± 0.2. Both pathologist number 2 and 4 had the highest ratio of 0.9 and pathologist 3 had the lowest at 0.4. Figure 1 is a scatter plot of the HR HPV positive ASC-US cases plotted against ASC-US/SIL ratio for each pathologist in the years 2004 - 2009.

Conclusions: The rate of HR HPV positivity for ASC-US cases was within acceptable ranges according to the ALTS data, except for pathologist 1 and 4. This could indicate that pathologist 1 with a 57% HR HPV positivity for ASC-US cases and an ASC-US/SIL ratio equal to the overall laboratory ASC-US/SIL ratio is under reporting ASC-US or undercalling SIL cases or both undercalling ASC-US and undercalling SIL at the same time. Therefore, some of the ASC-US cases were being reported as reactive and some SIL cases reported as ASC-US. Pathologist 4 had an ASC/SIL ratio within an acceptable range but a depressed HR HPV positivity rate for ASC-US cases. This could indicate that pathologist 4 is overcalling ASC-US or overcalling SIL or both overcalling ASC-US and SIL at the same time. This laboratory serves a low risk population which is reflected in our ASC/SIL ratio of 0.7 which has been relatively stable over the past six years. The ASC/SIL ratio for pathologist 3 fell slightly below the overall laboratory ASC/SIL ratio, however the HR HPV positivity for ASC-US cases fell within an acceptable range. This could indicate a tendency to interpret some ASC-US cases as SIL. We found no correlation between the ASC/SIL ratio and the percentage of positive HR HPV ASC-US cases which had a Pearson's correlation coefficient of -0.04. Although we found no correlation independently, these indicators provide useful feedback for the pathologist.

Table 1. Percentage of High-Risk HPV Positivity for ASC-US for each Pathologist

Pathologist	2004	2005	2006	2007	2008	2009	Average
1	64	49	55	57	62	53	57
2	44	42	36	40	52	52	44
3	50	39	32	29	49	35	39
4	37	29	31	34	45	41	36
5	44	42	48	56	62	55	51
6	43	39	34	42	50	33	40
Yearly Lab Average	47	40	39	43	53	45	45

Table 2. ASC/SIL Ratio for each Pathologist

Pathologist	2004	2005	2006	2007	2008	2009	Average
1	0.7	0.6	0.6	0.7	0.6	0.9	0.7
2	0.9	0.9	1.3	1	0.8	0.7	0.9
3	0.5	0.4	0.4	0.3	0.3	0.3	0.4
4	1	0.9	1	0.7	0.8	0.9	0.9
5	0.7	0.7	0.6	0.8	0.7	0.9	0.7
6	0.8	0.8	0.9	0.8	0.9	0.6	0.8
Yearly Lab Average	0.8	0.7	0.8	0.7	0.7	0.7	0.7

115

HPV Test Ordering Practices in a Large Community-Based Health Care System

Benjamin Witt, MD¹, Charnell Johnson, HTL², William Watkin, MD¹, Mark Dieterich, MD¹

¹Pathology, North Shore University Health System, Evanston, Illinois; ²Laboratory Information Systems, North Shore University Health System, Evanston, Illinois

Introduction: HPV testing, as an adjunct to the liquid-based Pap test, is an integral part of cervical cancer screening in many areas. While most young women experience transient HPV infections, those that are at risk for carcinoma are usually older and have persistent infections by high risk HPV subtypes. The American Society for Colposcopy and Cervical Pathology (ASCCP) has issued consensus guidelines (2006) for the use of HPV testing in cervical cancer screening; including age-related recommendations, use of reflex and co-testing, as well as recommending against testing for low risk HPV subtypes. Since excessive HPV testing wastes healthcare resources and potentially increases patient anxiety without significantly reducing cancer risk, we sought to better define the HPV test usage patterns at our institution, NorthShore University HealthSystem, by analyzing data from our LIS for 2009.

Materials and Methods: The pathology LIS was queried for all Pap tests (99+% ThinPrep type) and associated HPV testing ordered from 1/1/2009 to 12/31/2009. Our institution offers testing for high risk (HR) HPV types by Digene HC, with or without low risk (LR) types, which may be ordered as “reflex” for ASCUS only or as “automatic” for any Pap result. The patient data were stratified into clinically relevant age groups (less than 21, 21-29, 30 and older) for evaluation of the HPV test order type and incidence of positive results.

Results: A total of55,043/59,026 (93%) of the Pap tests performed at our institution in 2009 had HPV testing ordered, of which 63% were reflex and 37% were automatic. 22,648 HPV tests were actually performed, of which 2,677 (11.8%) were positive. Findings which appeared to go against the testing guidelines included: 11% (6,032) of all HPV test orders including LR types, 89% of women under 21 having HPV testing ordered, and 6% of all automatic HR HPV test orders occurring in women under 30. The data do not account for HPV testing done in the context of atypical glandular cells or as recommended follow-up of a prior abnormality, but these were not expected to represent a significant proportion of our patient population. As expected, the incidence of HPV positivity went down with age. See Table 1 for the complete data set.

Conclusions: While the majority of HPV testing ordered at our institution was in accord with the consensus guidelines, some appeared to fall outside of them. Requests for LR HPV testing, historically offered more as a STD test, were higher than expected. While co-testing is recommended for women 30 or older, automatic HPV testing was ordered in 6% of younger women. Perhaps most significantly, some form of testing was ordered in 89% of young women (under 21), a group for whom no HPV testing is recommended. Our data suggest that increased communication and education directed at practitioners and their office staff could substantially reduce inappropriate HPV testing with its associated costs and triggering of unnecessary workups.

Soft Tissue/Bone

Table 1. Complete Data Set for HPV Ordering and Test Performance at NSUHS 2009

	Age <21		Age 21-29		Age 30+		All
Total Pap tests	2,545		11,254		45,227		59,026
HPV Test Order	# Orders	# HPV Tests	# Orders	# HPV Tests	# Orders	# HPV Tests	# Orders	# HPV Tests
Reflex to High Risk	1,819	184	8,591	772	19,619	1,040	30,029	1,996
Reflex to High/Low Risk	199	17	712	54	3,665	143	4,576	214
Automatic High Risk	175	175	1,009	1,009	17,798	17,798	18,982	18,982
Automatic High/Low Risk	67	67	240	240	1,149	1,149	1,456	1,456
No HPV Testing	285		702		2,996		3,983
% HPV+		42% of tests		37% of tests		8.6% of tests		11.8% of tests
% Tested for Low Risk Strains		3.3%		2.6%		2.9%		2.8%

116

Utility of Fine Needle Aspiration in Managing Patients with Soft Tissue Lesions

Shefali Chopra Chopra, MD, Alexander Lazar, MD,PhD., Christina Fanning, MD

Department of Pathology, MD Anderson Cancer Center, Houston, Texas

Introduction: Fine needle aspiration (FNA) of soft tissue lesions can be challenging due to overlapping clinical, radiological and cytological features. The aim of this study is to review the diagnostic efficacy of benign and malignant lesions diagnosed by FNA and to highlight the diagnostic pitfalls and limitations of this approach.

Materials and Methods: A total of174 consecutive soft tissue lesions diagnosed by FNA with subsequent histologic assessment of the excision specimen were retrieved from January 1995 to December 1996. Lipomas were excluded from the study.

Results: Of the 174 cases reviewed there were 73 primaries, 71recurrences and 30 metastases. FNA was considered adequate for diagnosis in 96% of cases and was sufficiently accurate to correctly guide patient management in 92% of cases (all inaccurate diagnosis were false negative except one). Cutting needle biopsy was used in conjunction with FNA in 62 of the 73 primary cases.

A specific histologic classification was possible (often more precise than needed to guide patient management) in 49% of the primary cases. Of these, the primary examples in benign cases were neural tumors (5 cases), myxoma (4 cases), nodular fasciitis and ganglioneuroma (1 case each). Amongst malignancies, specific diagnosis was possible in undifferentiated pleomorphic sarcoma (6 cases), liposarcoma (3 cases), and 1 case each of Ewing sarcoma, synovial sarcoma, dermatofibrosarcoma protuberance, myxoid liposarcoma, leiomyosarcoma, rhabdomyosarcoma and epithelioid sarcoma. In 9 of the 17 sarcomas that were not specifically classifiable on FNA, the core biopsy and resection specimen did allow further classification.

Pitfalls included low cellularity lesions without distinctive cells (such as desmoid fibromatosis) and vascular lesions (such as hemangioma) where FNA failed to provide specific histologic classification in virtually all cases.

Diagnosis was possible in all cases of recurrence and metastasis.

Conclusions: FNA, when used in conjunction with radiology, clinical data and experienced hands, is a highly effective, rapid and relatively inexpensive method to triage and diagnose soft tissue lesions even in primary diagnosis. FNA is also very helpful in recurrence and metastasis and can spare a more invasive biopsy in the great majority of cases. Nonetheless, it is important to realize the pitfalls and limitations of this approach to ensure accurate diagnostic assessments.

117

Elastofibroma: Cytomorphologic, Histologic, and Radiologic Findings in 5 Cases

Jeremy Vincent, MD, Elizabeth Montgomery, MD, Zahra Maleki, MD

Pathology, Johns Hopkins University, Baltimore, Maryland

Introduction: Elastofibroma is a rare benign, solid, ill-defined and slow-growing soft tissue lesion that typically arises between the tip of scapula and chest wall of elderly. Fine needle aspiration (FNA) is a commonly used technique to definitively diagnose the condition. However, the collagenous nature of the mass often results in a paucicellular FNA smear, resulting in a false-negative report. Herein, we report cytomorphologic, histological and radiologic findings in 5 cases.

Materials and Methods: Five cases of elastofibroma were retrospectively retrieved from the cytopathology archives of a teaching hospital. The cytological material was obtained by CT-guided (n=2) and ultrasound-guided (n=3) FNA. Smears were stained with Diff-Quik and Papanicolaou stains. Elastic stain and mucicarmine were reviewed in one case. Radiologic images and medical records and available histology (2 cases) were reviewed.

Results: There were 4 patients (2=males, 2=females). Their ages ranged from 64 to 84 (mean=71.25 years). All tumors were subscapular (n=2 right side, n=1 left side, and one bilateral). Tumor sizes ranged from 3.3 cm to 7.5 cm in greatest dimension. Aspirated material was hypocellullar in all cases, consisting of scattered uniform spindle cells, mature adipocytes and collagen. Petaloid globules and serrated and braid-like linear elastic fibers were haphazardly scattered.

Conclusions: The characteristic petaloid globules typical of elastofibroma in aspirated material can be inadvertently mistaken or overlooked since the samples are hypocellular. Awareness of cytomorphologic features of elastofibroma and the typical clinical setting will permit an accurate diagnosis and eliminate the need for preoperative biopsy.

118

The Significance of Cytology in Evaluation of Pediatric Tumors Located Below the Diaphragm: A Five Years Institutional Experience

Arvind Rishi, MBBS, MD, Nora Morgenstern, MD, Patricia Wasserman, MD

Pathology, North Shore Long Island Jewish Health System, New Hyde Park, New York

Introduction: Fine needle aspiration cytology (FNAC) is a valuable tool in the diagnosis and management of pediatric tumors. Cytological smears made during intra-operative consult can be a very valuable resource for education, cytology histology correlation, and further molecular / cytogenetic analysis. Molecular studies done on intra-operative smears have significant utility in diagnosis, classification, treatment and prognostication of pediatric tumors. We present five years institutional data on the role of diagnostic and intra-operative cytology in pediatric tumors located below the diaphragm.

Materials and Methods: A retrospective five year analysis (year 2004-2009) was done on 187 pediatric tumors located below the diaphragm and diagnosed in the department of pathology at North Shore Long Island Jewish Health System. Fine needle aspiration cytology along with smears prepared during intra-operative consult were studied and correlated with the final histological diagnosis. Flow cytometry and molecular cytogenetics analysis was done on the intra-operative tissue procured in RPMI medium. EWSR1 gene rearrangement and n-myc status were studied in Ewing's sarcoma and neuroblastoma respectively. Detailed clinical features were obtained from the hospital records.

Results: See Table 1 - Overall age distribution ranged from 6 months to 19 years. Overall sex distribution showed male predominance in all the group of lesions. Distribution of malignant lesions included lymphoma (28.8%), PNET/ Ewing's sarcoma (20.3%), Wilms' tumor (12.8%), neuroblastoma (8.5%), rhabdomyosarcoma (6.9%), hepatoblastoma (4.2%), desmoplastic small round cell tumor (1.6%), testicular lesions (2.6%), renal cell carcinoma (1.5%), sarcoma (1.6%), and adrenocortical carcinoma (0.5%). Benign lesions included ganglioneuroma (2.1%), neurofibroma (1.1%), mature cystic teratoma (1.1%), angiomatoid fibrous histiocytoma (0.5%), congenital adenomatous malformation (0.5%) and bronchogenic cyst (0.5%). Cytology in the form of either intra-operative smear or FNAC was available in 74 cases and most of them were malignant lesions. Concordance with final histopathology was subsequently obtained in 92% (n = 68) cases. Discordance in 8% of cases was predominantly partial discordance because of the sampling issues or varied differentiation in the tumors. Most commonly encountered within the partial discordance group were Wilms' tumor with blastema predominance and small blue round cell tumors with spindle cell morphology. N-myc analysis in nine neuroblastoma cases showed amplification in 43% cases. Complex cytogenetic abnormalities in the form of t(14;15)(q10;q10), trisomy 12, trisomy 7, t(1;16) and del 16 were seen in Wilms' tumor. EWSR1 gene rearrangement was also seen in one case of angiomatoid fibrous histiocytoma. Cytogenetics studied in one case of synovial sarcoma showed characteristic t(X; 18).

Conclusions: Fine needle aspiration cytology and smears prepared during intra-operative consult is a highly specific diagnostic modality and valuable resource in the management of pediatric tumors. Triage of cytological specimens during the adequacy evaluation and smear preparation during intra-operative consult is of help in omitting the need for invasive diagnostic procedures. The material obtained by using sterile precautions can be further utilized for molecular and genetics studies thereby adding to the value of integrated management and prognostication.

Special/New Techniques

Table 1. Distribution of type of pediatric tumors below the diaphragm

Lesion	Number	%age
Lymphoma	54	28.8
Ewing's Sarcoma	38	20.3
Wilms' Tumor	24	12.8
Neuroblastoma	16	8.5
Rhabdomyosarcoma	13	6.9
Hepatoblastoma	8	4.2
Ganglioneuroma	4	2.1
Desmoplastic small round cell tumor	3	1.6
Neurofibroma	2	1.1
Testicular lesions - Embryonal carcinoma - Monodermal  teratoma - Seminoma - Burkett's Lymphoma	2 1 1 1	1.1 0.5 0.5 0.5
Renal cell carcinoma - Papillary renal cell  carcinoma - Medullary carcinoma - Translocation  associated RCC	1 1 1	0.5 0.5 0.5
Pheochromocytoma	4	2.1
Adrenocortical carcinoma	1	0.5
Synovial sarcoma	2	1.1
Small cell sarcoma	1	0.5
Cystic renal dysplasia	5	2.6
Mature cystic teratoma	2	1.1
Angiomatoid fibrous histiocytoma	1	0.5
Congenital cystic adenomatous malformation	1	0.5
Bronchogenic cyst	1	0.5

119

Unexpected New Primary Neoplasia Diagnosed During a Staging Procedure Using Endoscopic Ultrasound-Guided Fine Needle Aspiration

Sherrie White, MD¹, Andrew Pellecchia, MD², Maria Abadi, MD¹

¹Pathology, Jacobi Medical Center, Bronx, New York; ²Medicine, Jacobi Medical Center, Bronx, New York

Introduction: Endoscopic ultrasound-guided fine needle aspiration (EUS guided-FNA) has proven to be a cost-effective and relatively safe method in the diagnosis and staging of malignancy. If diagnostic material is obtained, it reduces hospitalization time and the need for more invasive procedures such as exploratory laparoscopy or laparotomy. We present 4 cases in which unexpected new primaries were diagnosed with cytologic material obtained from endoscopic ultrasound-guided aspirations intended for staging the patients' original malignancies

Materials and Methods: Four patients (ages 65-79) underwent EUS-guided FNAs. The patients presented with multiple abdominal, thyroid and lung masses. Smears and cell blocks were prepared in all cases and immunohistochemical studies were performed when necessary.

Results: As seen in Table 1.

Conclusions: In our small series, four patients were diagnosed with a new unexpected primary during a EUS guided-FNA staging procedure. As a result, the patients' stage and treatment changed. EUS guided-FNA is a useful and effective method in the diagnosis and clinical management of oncologic patients.

Table 1. 

Case #	Original Neoplasia	New Primary Diagnosis
1	Rectal Adenocarcinoma	Hepatocellular Carcinoma
2	Cholangiocarcinoma	Thyroid Carcinoma
3	Rectal Adenocarcinoma	Lung non-small cell Carcinoma
4	Gastric Adenocarcinoma	Follicular Lymphoma

120

The Utility of Abdominal Fat Pad Aspiration for Diagnosing Systemic Amyloidosis

Stephen Baker, MD, Carmela Tan, MD, Christine Booth, MD

Anatomic Pathology, Cleveland Clinic, Cleveland, Ohio

Introduction: Abdominal fat pad aspiration is considered the “gold standard” for diagnosing systemic amyloidosis. At our institution, the fine needle aspiration (FNA) service routinely performs this procedure on patients being evaluated for amyloidosis. The purpose of this study was to evaluate the utility of abdominal fat pad FNA in the diagnosis of systemic amyloidosis.

Materials and Methods: An electronic search of the cytology files was performed to find all patients with an abdominal fat pad FNA from 1/1/2004 to 5/31/09. The patients' electronic medical records and the pathology database were used to determine presenting diagnoses and subsequent findings related to amyloidosis. All abdominal fat pad aspirations were performed using either a 21 or 23 gauge needle. All aspirations were stained with Congo red.

Results: One hundred two fat pad FNAs were performed on 100 patients. The indications for the aspirations included amyloidosis (14), multiple myeloma (27), MGUS (41), cardiac abnormalities (10), renal abnormalities (4), GI/hepatic abnormalities (2), and neuropathy (2). The results were as follows: Congo red negative 91, Congo red positive 5, equivocal 1, and unsatisfactory 5. Four of the 5 patients with positive FNA had evidence of serum and/or urine monoclonal gammopathy. Of the 91 negative aspirations, 25 patients had tissue diagnosis of amyloidosis either prior or subsequent to the aspiration. In this study, fat pad FNA had a sensitivity of 13.7% and specificity of 98.5%.

Conclusions: The diagnostic utility of abdominal fat pad biopsy for the evaluation of systemic amyloidosis in our study is low. Its utility as a routine screening method for amyloidosis is limited by its low sensitivity, nonspecific nature of Congo red staining, and inadequate material available for amyloid typing.

121

Hybrid Capture 2 HPV Testing in Head and Neck Fine Needle Aspirations

Ghada Aramouni, CT (ASCP)¹, Jennifer Brainard, MD¹, Belinda Yen-Lieberman, PhD², Aaron Hoschar, MD¹, Deborah Chute, MD¹

¹Anatomic Pathology, Cleveland Clinic, Cleveland, Ohio; ²Clinical Pathology, Cleveland Clinic, Cleveland, Ohio

Introduction: A subset of head and neck squamous cell carcinomas (SCCs) are related to high risk human papilloma virus (HPV) infection. These have unique epidemiologic associations and important treatment differences from non-HPV related SCC. HPV-related SCCs most commonly occur in the oropharynx, and are associated with an improved overall survival after treatment at 5 years. As many HPV-related SCCs of the head and neck present with neck metastasis, they are frequently initially diagnosed by fine needle aspiration (FNA) biopsy. Detecting HPV in cytology specimens has the potential to direct patient care while avoiding additional procedures. HPV detection in surgical resection tissue has included polymerase chain reaction for HPV, in situ hybridization (ISH) for high risk HPV and p16 immunohistochemical staining (a surrogate marker for HPV). The most commonly used test for detecting high risk HPV in uterine cervix samples acquired during Pap smears is the Digene Hybrid Capture 2 test (HC2). We propose to evaluate the utility of HC2 in metastatic SCC FNA sample rinses to detect clinically relevant high risk HPV, as well as HPV ISH in FNA cell block material.

Materials and Methods: Head and neck FNAs with a diagnosis of squamous cell carcinoma or suspicious for squamous cell carcinoma and sufficient material for testing were included. Cases without surgical tissue confirmation were excluded. FNA samples stored in a methanol-based fixative (CytoLyt® or Preservcyt®) and were tested using the HC2 test according to the manufacturer's directions for cervical cytology testing. HC2 tests were considered positive when the relative light threshold ratio was greater than 1.0. In addition, cell blocks were made using the Cellient™ Automated Cell Block System when sufficient material was available. Cell block sections and subsequent formalin-fixed surgical specimens were stained with the INFORM HPV III Family 16 Probes chromogenic ISH (CISH) for high risk HPV. Samples were considered positive when discrete, dot-like signals were localized within tumor nuclei.

Results: Over 4 months, 11 patient FNA samples were collected which had surgically confirmed head and neck squamous cell carcinoma. Of these, 5 had sufficient material for both cell block and HC2 testing, 3 had sufficient material for HC2 testing only, and 3 had cell blocks without HC2 testing. The primary tumor location was tonsil in 6 cases, larynx in 2 cases, pyriform sinus in 1 case, and unknown in 2 cases (confirmation testing performed on surgical biopsy of metastatic tumor). Surgical CISH HPV was positive in 6 cases (all tonsillar location) and negative in 5 cases. Correlation of the surgical CISH, cell block CISH, and HC2 testing is seen in Table 1. CISH and HC2 showed the same sensitivity (75%) and specificity (100%). One HPV positive tumor was negative on the HC2 test; this case had insufficient material for cell block, and may not have had tumor cells in the vial for HC2 testing. One HPV positive tumor was negative on cell block CISH testing; this case was positive by HC2, but the cell block showed almost all necrotic debris.

Conclusions: The Hybrid Capture 2 test shows similar sensitivity and specificity to high risk HPV CISH for determining HPV status in squamous cell carcinoma metastasis. This test may be particularly useful in those cases with insufficient material for cell block.

Table 1. Comparison of HC2 with Cell Block HPV CISH and Surgical HPV CISH

Surgical HPV CISH	Cell block HPV CISH		Hybrid Capture 2
	Positive	Negative	Positive	Negative
HPV: Human Papilloma Virus, CISH: Chromogenic in situ hybridization.
POSITIVE	3	1	3	1
NEGATIVE	0	4	0	4

122

Circulating ALDH1 Positive Cells with Chromosome10/10q22 and EGFR Abnormalities are Novel Biomarkers for Monitoring Cancer Stem Cells and Correlate with ALDH1 Positive Cells of Non Small Cell Lung Cancer: An Immunocytochemistry - FISH based Study

Abha Khanna, MS¹, Tanweer Zaidi, MD¹, Ricardo Fernandez, BS¹, Weigong He, MD¹, Reza Mehran, MD², David Blowers, PhD³, Ruth Katz, MD¹

¹Pathology, MD Anderson Cancer Center, Houston, Texas; ²Thoracic Surgery, MD Anderson Cancer Center, Houston, Texas; ³AstraZeneca, Macclesfield, Cheshire, United Kingdom

Introduction: Non-small cell lung cancer (NSCLC) contains a population of cancer stem cells (CSC) that are responsible for its maintenance and metastases. Circulating CSCs from patients contain genetic abnormalities indicative of the biological behavior of the primary tumor and provide attractive targets for innovative drug therapy. We have recently demonstrated that the circulating tumor cells (CTCs) with deletions of 10/10q22, containing the surfactant protein gene (SP-A) were strongly associated with relapse. Lung cancers expressing SP-A are regulated by TTF-1, and have been shown to comprise a distinct and important subpopulation of adenocarcinomas known as TRU-type, which are associated with female gender, non-smoking and EGFR-mutations. We showed that in NSCLC there was a positive correlation between ALDH1 expression, stage, grade of tumor, loss of SP-A and poor prognosis. We wished to evaluate if 1) circulating ALDH1+ CSCs bear the identical cytogenetic phenotype of the ALDH1+ cells in the tumors by FISH and 2) if PBMCs (peripheral blood mononuclear cells) from patients with lung cancer have higher levels of circulating CSCs compared to controls.

Materials and Methods: Using a Bioview Duet scanner, PBMCs from patients with early (6) and advanced (4) NSCLC, small cell lung cancer (SCLC) (3) and controls (6) with no evidence of lung cancer were isolated by Ficoll-Hypaque gradient separation from peripheral blood and stained for ALDH1. At least 300 ALDH1+ cells were selected from three thousand PBMCs and stored as “target” cells and subsequently evaluated for loss or gain of chromosomes10/10q22 and EGFR as detected by FISH. In three primary NSCLCs, ALDH1+/Cytokeratin negative tumor cells were analyzed for the same 3 color FISH markers as the PBMCs.

Results: There were highly significant differences between the PBMCs of NSCLC cases and controls for all FISH biomarkers analyzed. Also notable was a significant difference between the numbers of deletions of chromosome 10 for SCLC versus NSCLC (p<0.028). Chromosome abnormalities within ALDH1+ cells in the primary tumor had a different genotype from the surrounding cytokeratin (CK)+ /ALDH1- cells. ALDH1+ tumor and ALDH1+ CTCs displayed monosomy 10, and disomy of 10q22 and EGFR, while CK+, ALDH1- tumor cells showed amplification of 10q22 and EGFR genes. See Figures 1, 2, and 3.

Conclusions: Our preliminary results suggest that both EGFR and 10q22 genomic abnormalities and monosomy of chromosome 10 may serve as important lung cancer stem cell markers. In particular we have identified that the ALDH1+ cells within the tumor contain unique 10/10q22 subpopulations which are the metastasizing population, and can be followed over time in the peripheral blood of patients.

Our results from our study provide compelling evidence for using 10/10q22/EGFR genes as novel biomarkers for lung cancer diagnosis and for monitoring of disease in the peripheral blood.

Figure 1. ALDH+/Cytokeratin negative tumor cells (green fluorescence) showing disomy chromosome 10q22 (SP-A) (Red), disomy EGFR (Green), and monosomy chromosome 10 (Aqua), while cytokeratin+ (red fluorescence) ALDH- tumor cells show amplification of EGFR and 10q22 (multiple green and red signals coalesced). (FISH and Immunocytochemistry performed on the same section from formalin fixed paraffin embedded tissue).

Figure 2. Non-small cell carcinoma cells, showing positive ALDH staining by immunocytochemistry (left image) and corresponding FISH on the same section showing monosomy of chromosome 10/10q22 (right image) (Section taken from paraffin embedded tissue).

Figure 3. ALDH+ cells showing monosomy of chromosomes 10 (single green signal)/10q22 (one or two red signals) in peripheral blood.

123

Remote Adequacy Assessment of Fine-Needle Aspiration Biopsy Specimens via Telepathology: A Review of the Success and Inter-observer Variability of 50 Consecutive Cases

Brian Smola, CT, Xin Jing, MD, Rodolfo Rasche, MD, Michael Roh, MD, PhD, Claire Michael, MD, Robertson Davenport, MD, Stewart Knoepp, MD, PhD

Pathology, University of Michigan, Ann Arbor, Michigan

Introduction: The geographic demands of cytologists performing immediate adequacy assessments on fine needle aspiration biopsy (FNAB) specimens may be great, especially at a large academic institution or for cytopathologists serving multiple remote clinical locations. As the technology of telepathology rapidly advances, it provides excellent opportunities for the transmission of high-quality pathologic images at a distance and the alleviation of increasing geographic constraints imposed on cytopathologists performing immediate assessments. To date, very few studies have validated the technique of real-time transmission of cytopathology specimens.

Materials and Methods: Fifty routine sequential FNAB cases were remotely re-screened by four staff cytopathologists. A senior cytotechnologist screened the cases while simultaneously transmitting real-time images to the remote staff cytopathologist. Constant communication between the cytotechnologist and staff cytopathologist was maintained via telephone. Images were captured utilizing a Nikon Digital Sight DS-Fi1 camera set at 1280 x 960 pixels (maximum resolution 2560 x 1920 pixels) attached to a Nikon Eclipse E400 microscope. Captured digital images were sent over the internet via Nikon Digital Sight software. Digital images were available for viewing by the sender via either attached Digital Sight DS-L2 monitor or Dell 22-inch monitor, and were available to the reviewer on Dell flat-screen 24-inch monitors attached to internet-capable computers. The cytopathologist was aware of any clinical history available at the time of the initial adequacy assessment. The initial screening was limited to Diff-Quik-stained slides; final correlation was made with permanent Pap-stained slides and surgical resection, if available.

Results: The time taken to assess adequacy and render a preliminary diagnosis ranged from a few seconds to eleven minutes, with an average time of three minutes fifteen seconds and median time of three minutes. Diagnostic concordance between preliminary diagnoses made via telepathology and the original preliminary diagnosis ranged from 98% to 100%. The final diagnostic accuracy ranged from 94% to 96%, with two cases (4%) not showing a malignant cell population on immediate adequacy (Diff-Quik) preparations but showing malignant cells on Pap-stained slides. Two cases that showed final results of low-grade lymphomas were suggested for evaluation with flow cytometry by all cytopathologists in the study.

Conclusions: Utilizing a system for dynamic, real-time, remote immediate-adequacy assessments on FNAB specimens is an efficient and highly accurate method for properly triaging FNAB cases and rendering a preliminary diagnosis. When the cytologist on the sending end is experienced and diagnostic material is present on the slides, our diagnostic accuracy approached 100%, showing that this method safely and efficiently increases the effective range of practicing cytopathologists.  

Table  . Presentations by Author

Author	Presentation Number
Abadi, Maria	119
Adams, Christina	78
Akalin, Ali	39
Akhtar, Israh	25
Akinfolarin, Josephine	51, 70
Alarcon, Jason	34
Alexander, Rhonda	25
Ali, Syed	17
Allen, Lisa	44
Almeida, Margarida	9
Alzohaili, Opada	51
Amsden, Erin	66
Anagnostou, Valsamo	69
Anisa, Kanbour	3
Apichairuk, Wipapat	89
Aramouni, Ghada	121
Askin, Frederic	78
Auger, Manon	6
Austin, Marshall	29
Aydin, Hakan	PP #1
Azuma, Arata	72
Bailey, Denis	85
Bajaj, Jaya	PP #10
Baker, Stephen	47, 120
Balachandran, Indra	7, 8
Balassanian, Ronald	87
Baliga, Mithra	25
Bandyopadhyay, Sudeshna	51, 70
Barkan, Guliz	PP #9
Barnes, Michael	13
Barroeta, Julieta	36
Baumal, Caroline	12
Bedrossian, Kristi	38
Bekker, Jessica	99
Belinson, Jerome	45
Belinson, Susane	45
Bendre, Manali	4
Benson, Jenna	26
Bentz, Joel	PP #5
Bergeron, Christine	PP #2
Bhattacharyya, Achyut	98
Bihlmeyer, Sharon	92
Billah, Shahreen	PP #7
Bird, Benjamin	38
Biscotti, Charles	47
Blond, Barbara	PP #11
Blowers, David	122
Blunier, Travis	110
Boerner, Scott	85
Bonfiglio, Thomas	26
Booth, Christine	6, 120
Bose, Shikha	37, 61
Bowlin, Ric	25
Bowman, Melissa	27
Brainard, Jennifer	47, 67, 121
Brandt, Suzanne	71, 82
Bridgeman, Andrew	65, 66
Brooks, John	36
Brugge, William	90, 96
Burke, Stephanie	39
Burlingame, Oname	100
Burroughs, Frances	22, 73, 74
Burstein, David	PP #8
Cahill, Leigh Ann	28
Campion, Michael	101, 111
Carr, Diane	114
Carvalho, Jason	19
Cary, Donna	27
Cash, Beverly	47
Chadwick, Barbara	18
Chakraborty, Subhendu	30
Chebib, Ivan	90, 91, 96, 102
Chen, Beiyun	5
Chen, Hua	PP #8
Chen, Longwen	48
Chen, Shaoxiong	10
Chhieng, David	PP #4
Chivukula, Mamatha	29
Choi, Hyun	63
Chopra, Shefali Chopra	116
Churchwell, Stacie	30
Chute, Deborah	PP #1, 121
Cibas, Edmund	100
Cimino-Mathews, Ashley	17
Clark, Douglas	22, 56, 74, 78
Clark, Jennifer	49
Clayton, Amy	PP #5, PP #6, 5, 24, 65, 101, 109
Coffey, Donna	30
Cohen, Cynthia	2
Colaizzi, Amy	3
Colasacco, Carol	99, 103
Colborn, Lisa	PP #6, 65
Collier, Michelle	68
Collins, Brian	18, 40
Covell, Jamie	PP #3
Cramer, Harvey	81, 86
Croyle, Maureen	47
Cunningham, John	98
Currens, Heather	104
Darrow, Mary	2
Davenport, Robertson	123
David, Odile	31, 89
Davis-Devine, Sharon	110
Dawlett, Marilyn	32
deBrito, Pedro	97
Deeds, David	47
del Rosario, Kristina	66
Deshpande, Vikram	21
Dhurandhar, Bakula	30
Diaz De Vivar, Andrea	32
Diem, Max	38
Dieterich, Mark	115
Dietrich, Ruth	31
DiFurio, Megan	105
Dimou, Anastasios	69
Dolar, Sandra	67
Doobay, Hema	44
Du, Hui	45
Duker, Jay	12
Duwe, Alexander	29
Dvorakova, Marie	92
Edelman, Morris	75
Eilers, Stan	49
Elhosseiny, Abdelmonen	92
Elsheikh, Tarik	PP #4, 62
Epperson, Keri	50
Erroll, Maria	35
Etzel, Carol	52
Evans, Karla	106, 107, 108
Evered, Andrew	107
Facik, Michael	49
Faigel, Douglas	94
Fairley, Todd	30, 41
Fanning, Christina	116
Fargahi, Sima	14
Fatheree, Lisa	109
Felgar, R.	83
Feller-Kopman, David	74
Fenderson, Allison	34
Feng, J.	70
Feng, Jie	52
Feng, Jining	51
Fernandez, Ricardo	122
Filomena, Carol	PP #3
Finley, David	71
Fischer, Andrew	PP #5, 109
Fischer, Timothy	44
Flowers, Rhyne	25
Frable, William	28
Freidman, Teresa	29
Freund, Gregory	110
Friedlander, Maria	7, 71
Gabrielson, Edward	78
Ganjei-Azar, Parvin	88
Gao, Faye	33
Gattuso, Paolo	68
Gaudier, Farah	110
Geddie, William	85
Geisinger, Kim	80
Genevay, Muriel	96
Ghofrani, Mohiedean	50
Giampoli, Ellen	15
Gidwani, Raja	51
Gillihan, Elizabeth	50
Giorgadze, Tamar	51, 70
Gong, Yun	PP #7
Gopez, Evelyn	18
Gore, Marquetta	44
Gorman, Blythe	1
Goulart, Robert	PP #3
Gu, Mai	14
Guo, Ming	PP #7, 10, 32, 52
Halline, Allen	89
Halling, Kevin	PP #6, 24, 65, 101, 111
Hanley, Krisztina	2
Hansing, Kara	66
Hanson, Joshua	92
Harigopal, Malini	PP #4, 69
He, Qin	44
He, Weigong	122
Henry, Michael	PP #5, PP #6, PP #11, 24, 57, 65, 66, 101
Hirata, Yukiyo	72
Hirschowitz, Sharon	58
Hoda, Rana	35
Hong, Waun-ki	52
Hoschar, Aaron	48, 121
Hudson, Roxanne	44
Hunt, Bryan	49
Husain, M.	70
Husain, Mujtaba	51
Ibrahim, Ahmad	3
Ihrke, Johnita	PP #6, 65, 66
Ikenberg, Hans	PP #2
Illei, Peter	74
Inagaki, Hirofumi	72
Izquierdo, Roberto	55
Jack, Annie	32
Jarboe, Elke	40
Jax, Jeanne	27
Jenkins, Robert	84
Jiang, Xiaoyin	46
Jing, Xin	123
Johnson, Charnell	115
Johnson, Michael	73
Jones, Bruce	PP #11
Jung, Han	63
Kamat, S.	83
Kanbour-Shakir, Amal	3
Kane, Lindsey	66
Kang, Jin	63
Kang, Suk	63
Kapur, Umesh	PP #9, 49
Karnes, R. Jeffery	24
Katz, Ruth	122
Kawada, Tomoyuki	72
Kerr, Sarah	84, 111
Khalbuss, Walid	11, 83, 92, 93
Khanna, Abha	122
Khurana, Kamal	55, 95
Kindelberger, David	21
Kirby, John	74
Kluskens, Larry	68
Knoepp, Stewart	19, 87, 123
Ko, Hyangmi	85
Koreishi, Aashiyana	82
Kosarac, Ognjen	1
Krishnamurthy, Savitri	4, 64
Kudoh, Shoji	72
Kumar, Lalita	58
Kuntzman, Tomi	27
LaPolice, Paula	49
Lappinga, Paul	57
Lastra, Ricardo	36
Laucirica, Rodolfo	6
Laver, Nora	12, 38
Lazar, Alexander	116
Lee, Hak	63
Lee, John	14
Leiman, Gladwyn	99, 103
Lemon, Larry	49
Levi, Angelique	PP #4
Li, Qing Kay	56, 73, 74, 78
Li, Ying-Ji	72
Lian, Qian	11
Lin, Fritz	14
Lin, Oscar	82
Llaudes, Maximo	75
Lobko, Igor	20
Lodhi, Ashutosh	64
Loghavi, Sanam	37
Lonardo, F.	70
Lu, Shaolei	53
Lwin, Toe	5
Ma, Gui-Ying	35
Major-Nicolas, Maud	25
Maleki, Zahra	76, 117
Malinowski, Douglas	44
Manglik, Niti	20
Mariani, Rachel	31
Marks, Rebecca	81, 86
Marshall, Carrie	77
Martin, Matthew	54
Matsko, Jonee	29
Mauser, Nancy	29
Mawhorter, Michael	67
Maygarden, Susan	23
McGarrah, Kathryn	31
Mead, Kristen	55
Mehran, Reza	122
Mehrotra, Swati	PP #9
Mentock, Erica	110
Metzler, Rhonda	PP #3
Meyer, Kevin	84
Michael, Claire	123
Miljković, Miloš	38
Mino-Kenudson, Mari	96
Minot, Douglas	5
Mody, Dina	1, 30
Monaco, Sara	3, 11, 83, 93
Montgomery, Elizabeth	117
Montone, Kathleen	36
Moreira, Andre	71, 80
Morgan, Terry	94
Morgenstern, Nora	PP #10, 20, 75, 118
Moriarty, Ann	PP #5, 6, 109
Mueller, Gail	112
Mukunyadzi, Perkins	49
Munfus-McCray, Delicia	78
Narra, Umadevi	64
Nassar, Aziza	5
Naud, Shelly	92
Nayak, Anupma	20
Nayar, Ritu	PP #5
Neal, Margaret	112
Nelson, Ramona	44
Nelson, Sally	PP #6, 65, 66
Nugent, Summer	56
Ohori, N. Paul	79
Ono, Jill	21, 87
Orsulak, Jessica	5
Otell, Susan	74
Papadimitrakopoulou, Vali	52
Parashette, Kalyan Ray	89
Patel, Shobha	52, 64
Pavlovitz, Brian	95
Payne, Linda	64
Pellecchia, Andrew	119
Perkins, Sherrie	40
Pestana, Ines	9
Pisharodi, Latha	53
Pitman, Martha	90, 91, 96
Plavnicky, John	15
Policarpio-Nicolas, Maria	54
Powers, Celeste	28
Preisel-Simmons, Barbara	34
Prpic, Nikki	44
Quintana, Liza	92
Raab, Stephen	104, 113
Rademacher, Susan	5
Ramirez, Rosa	64
Rasche, Rudolfo	123
Ray, Nancy	48
Reichel, Elias	12
Reising, Kathleen	50
Rekhtman, Natasha	71, 80
Renshaw, Andrew	112
Reuschenbach, Miriam	42
Reynolds, Jordan	57
Rezaei, M. Katayoon	59
Ridder, Ruediger	PP #2, 29
Rimm, David	69
Rishi, Arvind	118
Rizk, Nabil	71
Roh, Michael	123
Romain-Delva, Edele	35
Rong, Rong	95
Root, Renee	PP #6, 65, 66
Rose, Crystal	62
Rosenthal, Barbara	56
Rosenthal, Dorothy	22, 34, 74
Rosso, Peter	67
Roy, Ajoy	95
Rudomina, Dorota	80
Ruiz, Phillip	88
Russell, Donna	15, 26
Sabo, Debbie	47
Saccomanno, Gina	66
Saleh, Husain	49
Salhadar, Alia	PP #9
Samantray, Julie	51
Sampson, Susan	34, 44
Sams, Sharon	113
Sangle, Nikhil	18
Santos, Gilda	85
Scanga, Lori	23
Schaaf, Michael	PP #12
Schinstine, Malcolm	109
Schmidt, Dietmar	PP #2, 42
Schofield, Kevin	PP #4
Schubert, Jennifer	38
Schwartz, Mary	1, 6, 109
Seiz, Mirjam	42
Selbst, Megan	69
Sethi, S.	70
Shah, Ajay	49
Shaw, Gregory	58
Shi, Dongping	51, 70
Shimizu, Takako	72
Shinkai, Yusuke	72
Shroyer, Kenneth	39
Si, Qiusheng	10
Sidawy, Mary	97
Siddiqi, Anwer	25
Siddiqui, Momin	2
Siegel, David	20
Sigel, Carlie	71, 80
Silverman, Stuart	100
Singh, Meenakshi	39
Smith, Debora	30, 41
Smith, G. Denice	18, 40
Smola, Brian	123
Sneige, Nour	52
Solky, Ana	12
Sorenson, Angela	PP #6, 65, 66
Souers, Rhona	PP #5, PP #11, 109
Spencer, Jeremy	59
St. John, Tim	99, 103
Staerkel, Gregg	PP #7, 64
Steele, Cindy	55
Stephenson, Betty	34
Stewart, John	PP #7, 77
Stoll, Lisa	73
Straseski, Joely	56
Sturgis, Charles	PP #12
Sturgis, Erich	52
Sugrue, Chiara	PP #10, 20
Sullivan, Peggy	58
Sundborg, Michael	105
Sweeney, Brenda	16
Syrigos, Konstantinos	69
Tabatabai, Laura	13
Tabbara, Sana	59
Takeda, Ken	72
Takizawa, Hajime	72
Tambouret, Rosemary	16, 106, 107, 108
Tan, Carmela	120
Taylor, Adriann	44
Taylor, Marla	114
Teruya-Feldstein, Julie	82
Thomas, Nicole	6
Thomas, Rhona	6
Thrall, Michael	41
Tickman, Ronald	PP #12
Tochigi, Naobumi	79
Tolentino, Carrie	35
Tranchida, Paul	51
Travis, William	80
Tripp, Sheryl	40
Tsao, Anne	52
Turner, Brian	90, 96
Tuve, Delores	66
Tworek, Joseph	PP #11
Underwood, Dawn	67
Vakil, Behzad	35
Valente, Philip	54
VanDyke, Jessica	110
Varghese, Sharlin	15, 26
Villarmarzo, Guillermo	60
Vincent, Jeremy	117
Vinokurova, Svetlana	42, 43
Virella-Santana, Wilma	60
von Knebel Doeberitz, Christina	42
von Knebel Doeberitz, Magnus	42, 43
Voss, Jesse	5, 24, 101, 111
Wachs, Sarah	94
Wafai, Razan	PP #9
Wakely, Jr., Paul	49
Walker, Jr., Joe	8
Wallace, Susan	44
Walts, Ann	37, 61
Wang, Brant	97
Wang, Chun	45
Wang, Jessica	16
Wang, Jian-Ping	32
Wang, Jun	45
Wasserman, Jason	PP #10
Wasserman, Patricia	PP #5, PP #10, 20, 75, 118
Watkin, William	115
Weiss, Gary	20
Wendel Spiczka, Amy	49
White, Sherrie	119
Whitehead, Clark	44
Wilbur, David	16, 102, 106, 107, 108, 109
Wilgenbusch, Heather	112
Witt, Benjamin	115
Wojcik, Eva	PP #9
Wolfe, Jeremy	106, 107, 108
Wright, Moisha	64
Wu, Howard Her-Juing	62, 81
Wu, Maoxin	PP #8
Wu, Ruifang	45
Yaeger, Kurt	96
Yang, Bin	PP #1, 45, 67
Yang, Huaitao	11
Yang, Wei	4
Yeh, I-Tien	54
Yen-Lieberman, Belinda	121
Yoo, Young	63
Youens, Kenneth	46
Young, Kevin	50
Yu, Gengsheng	53
Yung, Rex	74
Zaidi, Tanweer	122
Zakowski, Maureen	71
Zhang, Haiying	98
Zhang, Jun	24, 101
Zhang, Yaxia	88
Zhao, Chengquan	33
Zheng, XiaoYong	PP #8
Zieman, Angela	66, 101, 111
Zimmerman, Carrie	101, 111

Presentations by Category

Breast

1, 2, 3, 4, 5, 6

Education/Training/Current Trends

7, 8

Fluids/CSF

9, 10, 11, 12, 13

Gastrointestinal Tract

14, 15, 16

Genoitourinary (Kidney/Bladder)

17, 18, 19, 20, 21, 22, 23, 24

Gynecologic

25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46

Head/Neck

47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63

Laboratory Operations

64, 65, 66, 67

Lung and Mediastinum

68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81

Lymph Node/Hematopoietic

82, 83, 84, 85, 86, 87, 88

Pancreas/Liver

89, 90, 91, 92, 93, 94, 95, 96, 97, 98

Quality Assurance

99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115

Soft Tissue/Bone

116, 117, 118

Special/New Techniques

119, 120, 121, 122, 123