Fax: (765) 747-4466
Comparison of ThinPrep and cytospin preparations in the evaluation of exfoliative cytology specimens
Article first published online: 20 MAR 2006
Copyright © 2006 American Cancer Society
Volume 108, Issue 3, pages 144–149, 25 June 2006
How to Cite
Elsheikh, T. M., Kirkpatrick, J. L. and Wu, H. H. (2006), Comparison of ThinPrep and cytospin preparations in the evaluation of exfoliative cytology specimens. Cancer, 108: 144–149. doi: 10.1002/cncr.21841
- Issue published online: 12 JUN 2006
- Article first published online: 20 MAR 2006
- Manuscript Accepted: 26 SEP 2005
- Manuscript Revised: 29 AUG 2005
- Manuscript Received: 22 JUN 2005
- nongynecologic cytology;
- exfoliative cytology
There exists limited literature comparing ThinPrep (TP) with conventional cytospins (CS) in nongynecologic specimens.
The differences between TP and CS were evaluated for a variety of parameters including cellularity, cytologic morphology, specimen preparation, screening time, laboratory cost effectiveness, cytologist preference, and impact on final diagnosis. Eighty-eight cases including 38 urine, 13 respiratory, and 37 body fluids were prepared simultaneously.
TP and CS demonstrated similar cellular yield in the majority of cases. Cytologists preferred TP in 63 (71.6%) and CS in 6 (6.8%) cases; whereas they indicated no preference in 19 (21.6%) cases. Of 14 abnormal cytologies, a more definitive diagnosis of malignancy was rendered by TP in 6 (42.9%) and by CS in 2 (14.3%) cases. TP demonstrated better nuclear chromatin morphology and more uniform distribution of cells. CS showed larger-sized clusters with better preservation of their architecture compared with smaller-sized clusters and significant shrinkage of cell size in TP.
TP was preferred over CS in the majority of cases by both cytotechnologists and pathologists. Cellularity, screening time, and specimen preparation were comparable, although the latter was easier to standardize in TP. In abnormal cases, TP was found to be 3 times more helpful than CS in rendering a definitive diagnosis of malignancy. TP, however, was associated with certain artifacts that cytologists must become familiar with when examining such preparations. Although TP was superior to CS in most cases, the application of both methods may be helpful in selected cases in which the TP diagnosis is not conclusive. Finally, TP was found to be more cost effective than CS. Cancer (Cancer Cytopathol) 2006. © 2006 American Cancer Society.
Numerous studies have compared ThinPrep® (TP) (Cytyc Corporation, Marlborough, MA) with conventional preparations in gynecologic cytology1–4 and, more recently, fine-needle aspiration (FNA) specimens.5–8 However, to our knowledge, only limited literature exist contrasting TP to conventional cytospins (CS) in the examination of nongynecologic (non-FNA) specimens.9–19 The purpose of the current study was to evaluate the differences between TP and CS in a variety of parameters including cellularity, specimen preparation, screening time, laboratory cost effectiveness, cytologist preference, and impact on final diagnosis.
MATERIALS AND METHODS
Eighty-eight nongynecologic specimens were evaluated, including 38 urine, 13 respiratory, and 37 body cavity fluids. Two fixed cytospins were routinely prepared from respiratory and urine specimens, whereas 2 fixed and 1 air-dried CS were prepared from body cavity fluids. CS were prepared utilizing the Shandon 3 Cytospin® cytocentrifuge (Thermo Electron Corporation, Waltham, MA). The samples were centrifuged at 750 revolutions per minute (rpm) for 5 minutes. Only 1 slide was prepared from all specimen types using the TP, according to the manufacturer's instructions.20 One to two additional concentrated pull smears (CPS) were prepared if initial CS were unsuitable for evaluation due to excessive thickness (ie, highly concentrated or compromised by excessive blood or inflammation). This more diluted CPS allowed for better observation of cellular elements. CPS was prepared by centrifuging the sample at 1700 rpm for 10 minutes and decanting the supernatant fluid. Using a disposable pipette, a drop of the specimen was placed onto a fully frosted glass slide. A plain glass slide was placed on top of the frosted slide and a CPS made. Average times for specimen preparation and screening were calculated from 70 cases. Specimen preparation times did not include computer entry, staining, or coverslipping. Specimens were processed in batch mode. Five cytotechnologists and 1 pathologist (T.M.E.) independently examined paired CS and TP specimens and recorded cellularity and their preferences for each of the cases. Cellularity was semiquantitatively graded as low, moderate, and high (Tables 1, 2). Cellularity and the preferences of the cytotechnologists were determined by tallying the majority response per individual case. Abnormal TP and CS cases were also evaluated by 1 pathologist (T.M.E.) whose observations are recorded in Table 3. Costs of TP and CS (Tables 4, 5) were calculated as follows: 1) specimen preparation labor = hourly preparation technician rate × 0.25 (15 min); 2) screening labor = hourly cytotech rate × 0.083 (5 min); and 3) supply costs = total costs of individual supplies used in processing for each technique.
|Specimen type||No.of cases||Conventional CS cellularity||TP cellularity||Cytologist preference|
|Total||88||36||40||12||36||48||4||6 (6.8%)||63 (71.6%)||19 (21.6%)|
|Specimen type||No. of cases||Better cellularity||Cytologist preference|
|Total||27||18 (66.7%)||9 (33.3%)||4 (14.8%)||19 (70.4%)||4 (14.8%)|
|Case no.||Specimen type||Age, y||Sex||Conventional CS||TP||Pathologist preference||Significant change in diagnosis||Follow-up|
|1||Pleural fluid||90||F||Suspicious, favor adenocarcinoma||Moderate||Atypical, favor reactive||Moderate||CS||Yes||None|
|2||Pleural fluid||70||F||Small cell carcinoma||Moderate||Favor small cell carcinoma||Moderate||CS||Yes||History of small cell carcinoma. Diff-Quik stain was superior|
|3||Pleural fluid||58||F||Favor small cell carcinoma||Moderate||Small cell carcinoma||Moderate||TP||Yes||None|
|4||Pleural fluid||69||F||Adenocarcinoma, rare cells||Low||Adenocarcinoma||Moderate||TP||No||Previous history of adenocarcinoma|
|5||Abdominal fluid||82||F||Suspicious for adenocarcinoma||Low||Adenocarcinoma||Low||TP||Yes||None|
|6||Bronchial wash||77||F||Atypical, favor reactive||Moderate||Suspicious for carcinoma||Moderate||TP||Yes||Biopsy nonsmall cell carcinoma|
|7||Bronchial wash||70||M||Adenocarcinoma||Moderate||Adenocarcinoma||High||TP||No||Biopsy: nonsmall cell carcinoma|
|8||Bronchial wash||65||M||Suspicious for squamous carcinoma||Moderate||Suspicious for squamous carcinoma||Low||CS||No||Biopsy: squamous carcinoma|
|9||Urine||50||F||Atypical, rule out LGUC||Low||HGUC vs. polyoma||High||TP||Yes||None|
|10||Urine||39||F||Nondiagnostic, no urothelial cells||High||HPV change, nondiagnostic||High||TP||No||None|
|13||Urine||82||M||Suspicious for HGUC||Low||HGUC, therapy effect||Moderate||TP||Yes||None|
|14||Urine||78||F||Degenerated atypical cells||Moderate||Reactive changes||Moderate||TP||Yes||None|
|ThinPrep||Cytospin (CPT 88108)*||Cytospin (CPT 88162)†|
|Screening time||5||5||8 (20, if CPS)|
|ThinPrep (CPT 88112)||Cytospin (CPT 88108)*||Cytospin (CPT 88162)†|
|Specimen preparation labor||3.50||2.52||2.52|
|Screening labor||2.50||2.50||3.00 (10.00 if CPS)|
|Total cost||13.30||11.57||12.12 (34.10 if CPS)|
|Revenue (not including pathologist expense)||100.32||52.22||52.84 (30.86 if CPS)|
Cytologists (cytotechnologists and pathologists) preferred TP over CS in the majority of cases (Table 1). The cytologist preference for TP was demonstrated in all types of specimens and was shown to be statistically significant (P = .00, Student t-test [SPSS software for Windows, version 11.5; SPSS Inc., Chicago, IL]). Sixty-one of 88 (69.3%) cases had similar cellularity, whereas 27 cases showed variable cellularity (better cellularity by either CS or TP). However, better cellularity by either methodology did not appear to influence the cytologists' preference (Table 2). There was an approximate 5-fold increased preference for TP (19 cases) over CS (4 cases), despite better cellularity demonstrated by CS in twice as many specimens (18 cases vs. 9 cases) (Table 2). Abnormal findings (atypical, suspicious, or malignant diagnoses) were reported by CS or TP in 14 of 88 (15.9%) cases (Table 3). The pathologist reviewer preferred TP over CS in 9 of the abnormal cases, and preferred CS in 3 cases. A more definitive, clinically relevant diagnosis was rendered by TP in 6 abnormal cases, compared with only 2 CS cases (Table 3, Cases 1 and 2). Overall, 5 of 88 (5.7%) cases were diagnosed as atypical/suspicious by TP, but a more definitive diagnosis was rendered in 2 of those cases after examination of their counterpart CS slides. In the latter cases, the combined cytomorphologic features of CS and TP, not differences in cellularity, allowed for better classification of the abnormal cells.
Cytomorphologically, compared with CS, TP demonstrated more uniform distribution of cells with less cellular overlapping and background debris. TP also demonstrated superior nuclear chromatin morphology in urothelial carcinoma (Fig. 1) and adenocarcinoma (Fig. 2), but showed more significant shrinkage of cells. Conversely, CS showed larger-sized clusters with better preservation of their architecture. Cluster formation was retained in TP, but clusters tended to be smaller in size (Fig. 2). In small cell carcinoma, nuclear chromatin details were better observed by TP; however, there was loss of the spindling and the prominent nuclear molding typically observed in CS (Fig. 3).
TP and CS specimen preparation times were comparable (Table 4). However, there was significant intertechnician variability noted in specimen processing times and quality of preparations using CS, compared with less variability associated with TP. Screening times were also comparable using TP or CS; however; they were found to be significantly increased (up to a 4-fold increase) when additional CPS were prepared (approximately 20% of CS cases required additional CPS) (Table 4). Although TP is slightly more costly than CS, the potential revenue/profit for the laboratory is approximately doubled with TP (Table 5).
Although our laboratory has utilized the TP methodology for the examination of gynecologic specimens for over 5 years, we have routinely employed CS in the preparation of nongynecologic exfoliative cytology specimens. More recently, however, we experienced high turnover rates in the technical staff in charge of specimen processing. This appeared to produce great variability in the quality of cytospins prepared and, on many occasions, led to the preparation of additional slides. Because this high staff turnover did not appear to affect the quality of gynecologic specimens prepared by TP, we considered implementation of TP in the routine processing of nongynecologic exfoliative specimens. However, in evaluating the literature, we found numerous studies comparing TP with conventional preparations in gynecologic cytology and FNA specimens, but there was limited literature comparing TP with CS in nongynecologic, non-FNA specimens.9–19 Therefore, the current study was undertaken to evaluate the differences between TP and CS for a variety of parameters, including specimen preparation, cellularity, cellular morphology, screening time, laboratory cost effectiveness, cytologist preference, and impact on final diagnosis.
The results of the current study demonstrated that cytotechnologists and pathologists preferred TP over CS in the majority of normal and abnormal cases, despite the better cellularity noted with CS in twice as many cases (Tables 1, 2). Specimen preparation and screening times were comparable, but specimen processing was easier to standardize in TP and appeared to be less operator-dependent (Table 4). Wright and Halford18 and Rana et al.16 also reported that their laboratory staff found TP specimen processing simpler and less labor intensive than CS. Because the amount of cell loss and cellular fixation may be attributed to the skill and experience of preparation staff, the simplicity and uniformity of TP technology appears to reduce this dependence on operator skill significantly, with minimal training required.18
Similar to other previous reports,9, 10, 13, 14, 16–18 the results of the current study also demonstrated that TP is superior or equal to CS for the microscopic evaluation of nongynecologic non-FNA specimens. TP demonstrated more uniform distribution of cells, with superior nuclear chromatin morphology and less cellular overlapping and background debris. However, TP was associated with certain artifacts and cytomorphologic changes that are different from CS and, although subtle, these changes need to be recognized by cytologists. TP was associated with significant shrinkage of cells, resulting in smaller-sized cells compared with CS. Cluster formation in adenocarcinoma was retained in TP, but tended to be smaller in size, compared with larger-sized clusters with better-preserved architecture in CS. In small cell carcinoma, nuclear chromatin details were better observed by TP, but there was loss of the spindling and prominent nuclear molding typically noted in CS. These cytomorphologic differences also have been observed by others.12, 16, 19
We found TP and CS to have comparable diagnostic sensitivity, similar to other studies.9, 10, 14, 17 However, more definitive, clinically relevant diagnoses were rendered in the current study by TP in 6 of 14 abnormal cases, compared with more definitive diagnoses rendered by CS in 2 of those cases (Table 3). Therefore, CS processing may be helpful in selected cases when the TP diagnosis is not conclusive. The latter approach takes advantage of the combined cytomorphologic features of CS and TP in difficult cases, without adding significant burden to the laboratory workload (approximately 6% of our cases). Finally, despite the slight increase in cost, TP is associated with higher reimbursement and profit for the laboratory (Table 5).
- 20Cytyc Corporation. Non-gynecological sample preparation. In: ThinPrep 2000 operators manual. Boxborough, MA: Cytyc Corporation; 1995: 1.1–1.29.