SEARCH

SEARCH BY CITATION

Keywords:

  • thyroid;
  • fine needle aspiration;
  • FLUS;
  • AUS;
  • BRAF;
  • RAS;
  • PAX8/PPARγ

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

BACKGROUND:

“Follicular lesion of undetermined significance/atypia of undetermined significance” is a heterogeneous category of cases that cannot be classified into 1 of the other established categories. The use of ancillary molecular studies has not been widely explored for this diagnosis.

METHODS:

All thyroid cytology cases diagnosed as follicular lesion of undetermined significance/atypia of undetermined significance were retrieved from April 2007 to December 2008. During this time period, samples were collected routinely at the time of aspiration for cytologic and molecular studies. Analysis for BRAF and RAS gene mutations and RET/PTC and PAX8/PPARγ gene rearrangements were performed and correlated with the cytologic features and surgical pathology outcome.

RESULTS:

From a total of 513 follicular lesion of undetermined significance/atypia of undetermined significance cases identified, 455 had adequate molecular results. Of these, 117 cases had cytologic-histologic correlation. In this group, 35 (29.9%) cases had a neoplastic outcome and 20 (17.1%) cases from 19 patients were carcinoma. Positive molecular results were found in 12 cases, all of which were papillary carcinoma. There were no false-positive molecular results. In correlating the molecular results with surgical pathology outcome, we found that the cancer probability for follicular lesion of undetermined significance/atypia of undetermined significance cases with molecular alteration was 100%, while the probability for follicular lesion of undetermined significance/atypia of undetermined significance cases without molecular alteration was 7.6% (P < .001).

CONCLUSIONS:

By cytomorphology alone, follicular lesion of undetermined significance/atypia of undetermined significance specimens represent cases that are intermediate in risk between the benign and “suspicious for follicular neoplasm” categories. Although not all papillary carcinoma cases are detected by molecular testing, a positive molecular test result is very helpful in refining follicular lesion of undetermined significance/atypia of undetermined significance cases into high-risk and low-risk categories. Cancer (Cancer Cytopathol) 2010. © 2010 American Cancer Society.

The new Bethesda 2007 Thyroid Cytology Classification defines “follicular lesion of undetermined significance/atypia of undetermined significance” as a heterogeneous category that cannot be classified as benign, suspicious for follicular neoplasm, or suspicious for malignancy because of the identification of minimally atypical cells, architectural atypia, or compromising factors including low cellularity, poor fixation, or obscuring elements.1 The clinical significance of this group is uncertain; however, the risk of malignancy has been estimated at 5%-10% and falls between benign (1%) and follicular neoplasm (20%-30%) diagnoses. Therefore, a repeat fine-needle aspiration (FNA) procedure is generally recommended. Although a repeat FNA may provide clarification, additional modalities that can streamline the clinical management are potentially helpful. The use of ancillary testing has not been widely explored for the follicular lesion of undetermined significance/atypia of undetermined significance or other indeterminate diagnoses.2 In recent years, molecular tests have been shown to be useful in the diagnosis of thyroid neoplasms. Point mutations in BRAF and RAS genes and gene rearrangements involving RET/PTC and PAX8/PPARγ have been found in over 70% of thyroid malignancies.3 Molecular testing is well-suited as an adjunct to thyroid FNA diagnosis because it requires a relatively small amount of tissue. The objective of this study is to compare the cytologic findings of our follicular lesion of undetermined significance/atypia of undetermined significance cases with the results of our molecular studies and surgical pathology outcome.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

The study was reviewed and approved by the Institutional Review Board of the University of Pittsburgh. All thyroid FNA specimens were collected under radiologic ultrasound guidance using a 23- to 27-gauge needle by a radiologist or an endocrinologist. On average, 3 to 4 FNA passes were performed for each thyroid nodule. The aspirated samples were used for making direct smears (wet-fixed for Papanicolaou staining and air-dried for DiffQuik staining) and monolayered slides by ThinPrep processing. Residual FNA material was placed directly into the nucleic acid preservative solution for molecular analysis. Cellular material collected in the nucleic acid preservative solution was frozen at −20°C and transported to the Molecular Anatomic Pathology Laboratory for analysis.

Cytology specimens with the features of follicular lesion of undetermined significance/atypia of undetermined significance were obtained from our cytology files dating from April 2007 to December 2008. Because the transition to the Bethesda 2007 thyroid classification system took place in September 2008, each cytology diagnosis from the earlier period of the study (until September 2008) was reviewed by 1 of the authors (NPO). In general, prior cases that were designated “less than optimal” (mostly because of low cellularity, mild cellular atypia, architectural atypia, poor fixation, or obscuring elements) were placed in the follicular lesion of undetermined significance/atypia of undetermined significance category. For cases in which the translation was equivocal, the original slides were reviewed by 1 of the authors (NPO) to determine whether to include them in the study.

Mutational analysis was performed on all cytology specimens for BRAF (V600E), NRAS codon 61, HRAS codon 61, and KRAS codons 12 and 13 by the LightCycler real time polymerase chain reaction (PCR) and post-PCR melting curve analysis as previously described.4 Single-step reverse transcriptase (RT)-PCR was performed to amplify the fusion points of RET/PTC1, RET/PTC3, and PAX8/PPARγ gene rearrangements. The cytologic findings of the follicular lesion of undetermined significance/atypia of undetermined significance cases were compared with the molecular results and surgical pathology outcome. Although the results of the molecular studies from the cytology samples were available to the surgical pathologist, the diagnoses on the thyroid resection specimens were based primarily on the histopathologic features. Statistical analysis was performed using Fisher exact test.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

During the 21-month period of our study, there were 2507 ultrasound-guided FNA specimens at our institution. On the basis of the Bethesda 2007 thyroid classification, 513 (20.5%) follicular lesion of undetermined significance/atypia of undetermined significance cases were identified. Seven cytopathologists were involved in making the diagnoses during this time period. The rate of follicular lesion of undetermined significance/atypia of undetermined significance diagnosis among the cytopathologists ranged from 14 to 27%, and the most common reason for placing the cases in the follicular lesion of undetermined significance/atypia of undetermined significance category was because of low cellularity (79% of cases). Other factors included architectural atypia, cellular atypia, poor preservation, and obscuring blood (Figs. 1-5). Onsite evaluation was performed on a subset of thyroid FNA cases (39%). Molecular testing was performed on all follicular lesion of undetermined significance/atypia of undetermined significance cases; however, the molecular test result was uninformative because of scant sampling in 58 cases (11.3%). Of the 455 follicular lesion of undetermined significance/atypia of undetermined significance cases with adequate molecular results, corresponding surgical pathology resection cases were available for a core group of 100 patients with 117 follicular lesion of undetermined significance/atypia of undetermined significance cases (Table 1). Outside of the core group, there were 2 molecular-positive follicular lesion of undetermined significance/atypia of undetermined significance cases that received subsequent care elsewhere (no corresponding surgical pathology resection available) and were not included in subsequent analysis. In the core group, 20 cases from 19 patients had carcinoma upon resection and 18 cases from 15 patients corresponded to an adenoma on final histopathology. The specific types of neoplasms included follicular adenoma -9, oncocytic adenoma -6, follicular variant papillary thyroid carcinoma (PTC) -14, classic type PTC -4, and tall cell variant PTC -1.

thumbnail image

Figure 1. A follicular lesion of undetermined significance/atypia of undetermined significance-low cellularity is shown. (A) Low magnification image shows a rare cluster of epithelial cells. (B) The rare cluster is composed of oncocytic cells with binucleation and prominent nucleoli. Molecular studies were negative and the resection specimen showed an oncocytic adenoma (A, DiffQuik stain; original magnification, ×40; B, DiffQuik stain; original magnification, ×400).

Download figure to PowerPoint

thumbnail image

Figure 2. A follicular lesion of undetermined significance/atypia of undetermined significance-architectural atypia is shown. A group of epithelial cells demonstrates abortive microfollicle formation. Molecular studies revealed an HRAS61 mutation and the resection specimen showed follicular variant papillary thyroid carcinoma (Papanicolaou stain; original magnification, ×400).

Download figure to PowerPoint

thumbnail image

Figure 3. A follicular lesion of undetermined significance/atypia of undetermined significance-cellular atypia is shown. A rare cluster shows occasional epithelial cells with nuclear elongation and irregular nuclear membranes. This case demonstrated a NRAS61 mutation by molecular studies and was shown to be follicular variant papillary thyroid carcinoma upon resection (Papanicolaou stain; original magnification, ×400).

Download figure to PowerPoint

thumbnail image

Figure 4. A follicular lesion of undetermined significance/atypia of undetermined significance-poor preservation is shown. Poorly preserved epithelial cells are entrapped in fibrin. Molecular studies were negative but the resection specimen showed a classic papillary thyroid carcinoma (Papanicolaou stain; original magnification, ×400).

Download figure to PowerPoint

thumbnail image

Figure 5. A follicular lesion of undetermined significance/atypia of undetermined significance-obscuring blood is shown. Numerous epithelial cells are present but obscured by blood. Molecular studies were negative and the resection specimen showed an oncocytic adenoma. (Papanicolaou stain; original magnification, ×100).

Download figure to PowerPoint

Table 1. Follicular Lesion of Undetermined Significance/Atypia of Undetermined Significance Cases and Molecular Results
Follicular Lesion of Undetermined Significance/Atypia of Undetermined Significance CasesAll Molecular TestsUninformative Molecular ResultsAdequate Molecular ResultsPositive Molecular ResultsNegative Molecular Results
All cytology follicular lesion of undetermined significance/atypia of undetermined significance cases5135845514441
Follicular lesion of undetermined significance/atypia of undetermined significance cases with surgical pathology correlation124711712105

A molecular alteration was found in 12 of 117 cases from 12 patients and all of these cases resulted in PTC, yielding a positive predictive value of 100% for molecular alteration. Of the remaining 88 patients with 105 negative molecular results, 7 patients with 8 negative molecular results had an outcome of PTC. Conversely, 97 cases from 81 patients with negative molecular results had a benign outcome yielding a negative predictive value of 92.7%. The difference in the proportion of cancer cases between the positive and negative molecular test groups (100% vs 7.6%) was statistically significant (P < .001). Furthermore, molecular results from 18 cases corresponding to 15 benign neoplasms (ie, follicular adenoma or oncocytic adenoma) were all negative (Table 2). Analysis of the 12 cases with molecular alterations on the cytology specimens demonstrated the surgical pathology outcome correlation summarized in Table 3. As expected, BRAF positive cases were either classic type PTC or tall cell variant PTC. All NRAS, HRAS, and PAX8/PPARγ gene-positive cases were follicular variant PTC. There were no KRAS, RET/PTC1, or RET/PTC3 gene-positive results.

Table 2. Correlation of Follicular Lesion of Undetermined Significance/Atypia of Undetermined Significance Diagnoses With Molecular Results and Surgical Pathology Outcome
CategorySurgical Pathology: NonneoplasticSurgical Pathology: AdenomaSurgical Pathology: Papillary Carcinoma
  • a

    The difference between these 2 groups was statistically significant (P < .001).

All follicular lesion of undetermined significance/atypia of undetermined significance cases, n = 11779/117 (70.1%)18/117 (12.8%)20/117 (17.1%)
Follicular lesion of undetermined significance/atypia of undetermined significance cases with positive molecular result, n = 120/12 (0%)0/12 (0%)12/12 (100%)a
Follicular lesion of undetermined significance/atypia of undetermined significance cases with negative molecular result, n = 10582/105 (78.1%)15/105 (14.3%)8/105 (7.6%)a
Table 3. Comparison of Genetic Alteration and Type of Neoplasm in the 12 Molecular Positive PTC Casesa
MutationPTC Follicular VariantPTC ClassicPTC Tall Cell
  • a

    There were no KRAS, RET/PTC1, or RET/PTC3 positive results.

BRAF021
NRAS61700
HRAS61100
PAX8/PPARgamma100

The value of molecular studies was compared with diagnostic value of repeat FNAs. From the core group of 100 patients with 117 follicular lesion of undetermined significance/atypia of undetermined significance FNAs, 51 patients went to resection without a repeat FNA and 49 patients had repeat FNAs (Table 4). Molecular studies alone identified 12 patients with PTC (12%), while repeat FNAs resulting in a second follicular lesion of undetermined significance/atypia of undetermined significance, follicular neoplasm, or suspicious for malignancy diagnosis led to the detection of 8 patients with PTC (8%). Molecular studies alone detected 7 PTCs from patients that did not have a repeat FNA. Conversely, repeat FNAs resulting in a second follicular lesion of undetermined significance/atypia of undetermined significance, follicular neoplasm, or suspicious for malignancy diagnosis detected 3 PTC cases that were negative by molecular studies.

Table 4. Diagnostic Value of Repeat FNA and Molecular Results Based on 100 Patients
OutcomeSingle Follicular Lesion of Undetermined Significance/Atypia of Undetermined Significance FNA Diagnosis, n = 51 patientsFollicular Lesion of Undetermined Significance/Atypia of Undetermined Significance Diagnosis Followed by Repeat Negative Diagnosis, n = 1 patientFollicular Lesion of Undetermined Significance/Atypia of Undetermined Significance Diagnosis Followed by Repeat Follicular Lesion of Undetermined Significance/Atypia of Undetermined Significance Diagnosis, n = 33 patientsFollicular Lesion of Undetermined Significance Diagnosis Followed by Repeat Follicular Neoplasm or Suspicious for Malignancy Diagnosis, n = 15 patients
  • FNA indicates fine-needle aspiration.

  • a

    All molecular test results were negative for mutations.

Nonneoplastic34a1a22a9a
Adenoma6a05a4a
Papillary carcinoma11062
molecular+: 7; molecular−: 4molecular+: 4; molecular−: 2molecular+: 1; molecular−: 1

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

Indeterminate thyroid cytology categories such as follicular neoplasm and suspicious for malignancy are well established in the literature.5 The cytologic diagnosis of follicular neoplasm is given to cellular specimens with a predominance of microfollicles and absent to minimal amount of colloid. The follicular neoplasm cytologic diagnosis implies a differential diagnosis, including hyperplastic nodule (in benign processes such as goiter or lymphocytic thyroiditis), follicular adenoma, follicular carcinoma, and follicular variant PTC. The expected risk of malignancy is 20%-30%.1 Likewise, the suspicious for malignancy diagnosis is given to cases with abnormal cytologic features concerning for a malignancy (usually PTC). The probability for the outcome of carcinoma for the suspicious for malignancy diagnosis is 50%-75%. Usually, surgical excision is considered for patients with the follicular neoplasm and suspicious for malignancy diagnoses.

In addition, there remain other indeterminate cases that cannot be placed in follicular neoplasm or suspicious for malignancy diagnostic categories. The Bethesda 2007 Thyroid Cytology Classification places these cases in the follicular lesion of undetermined significance/atypia of undetermined significance category. The new follicular lesion of undetermined significance/atypia of undetermined significance category is least understood among cytopathologists and clinicians. The authors of the Bethesda Classification estimated the probability for malignancy to be 5%-10% and projected the incidence of this diagnosis to be no greater than 7% of the thyroid FNA cases. In our study, we encountered follicular lesion of undetermined significance/atypia of undetermined significance in 20.5% of cases. Although this figure exceeds the 7% limit, in a recent multi-institutional study by Layfield et al, the average frequency of follicular lesion of undetermined significance/atypia of undetermined significance diagnoses rendered by institutions ranged from 3.3% to 14.9%, and the frequency of follicular lesion of undetermined significance/atypia of undetermined significance diagnoses given by individual cytopathologists ranged from 2.5% to 28.6%.6 Likewise, Nayar and Ivanovic showed that indeterminate cases equivalent to follicular lesion of undetermined significance/atypia of undetermined significance cases comprised 18% of total thyroid FNA cases.7 In their study, immediate evaluation was performed on 63% of cases and the majority of follicular lesion of undetermined significance/atypia of undetermined significance cases were attributed to “morphologic” issues pertaining to hyperplasia versus neoplasia or lymphocytic thyroiditis with concern for concurrent neoplasia. Our figure of 20.5% was higher but comparable to those quoted and likely because of several factors. The majority (79%) of our follicular lesion of undetermined significance/atypia of undetermined significance cases was because of low cellularity, and the remaining cases were because of architectural atypia, cellular atypia, and obscuring elements. “Adequacy related” (low cellularity) issues can be attributed to the procedure and the finding that onsite evaluations were performed on a minority of cases (39%). The remaining issues related to architectural and cellular atypia could be attributed to the individual cytopathologist's interpretation of “atypia”. Layfield et al studied the variability in the frequency of the reporting of follicular lesion of undetermined significance/atypia of undetermined significance cases, and he found that the rate of follicular lesion of undetermined significance/atypia of undetermined significance ranged from 2.5% to 28.6% by individual cytopathologists. In our study, there were 7 cytopathologists involved in making the diagnoses and the individual cytopathologist's rate of making the follicular lesion of undetermined significance/atypia of undetermined significance diagnosis ranged from 14% to 27%. For improved diagnostic conformity, a reference “study set” of diagnostic images would have been valuable. However, during the time of this study, the Bethesda atlas detailing the cytomorphology of the diagnostic categories had not yet been published (due in fall of 2009). Although our follicular lesion of undetermined significance/atypia of undetermined significance rate of 20.5% was relatively high, our study showed that adequate molecular testing results were achieved on the vast majority of cases, many of which were low in cellularity.

The calculation of the probability of carcinoma requires cytologic-histologic correlation. Because the majority of follicular neoplasm and suspicious for malignancy cases go to surgery, a realistic estimate of the probability of carcinoma can be made by cytologic-histologic correlation with the denominator representing all operated cases. The recent reports by Layfield et al and Nayar and Ivanovic demonstrated that the frequency of resection in follicular lesion of undetermined significance/atypia of undetermined significance cases ranged from 18.9% to 46.5%.6, 7 In our experience, 25.7% of cases underwent surgical resection, and of these, 20 cases (17%) from 19 patients had carcinoma. This figure is higher than that from Nayar and Ivanovic's study (6%) but lower than that of Layfield's study (28.3%).

The main focus of our study was to examine our follicular lesion of undetermined significance/atypia of undetermined significance cases for comparison with molecular results (BRAF and RAS gene mutations and RET/PTC and PAX8/PPARγ gene rearrangements) and surgical pathology outcome. We found that the molecular test results were very specific; there were no false-positive cases (all 12 molecular-positive cases were true positives). Previous reports showed that BRAF gene mutation and RET/PTC gene rearrangement were extremely specific for PTC (to date, only 1 false-positive case has been reported in an “atypical nodular hyperplasia”).8, 9, 10, 11RAS gene mutation and PAX8/PPARγ gene rearrangement were associated with follicular-patterned lesions, and the majority of lesions harboring these genetic alterations were cancer. However, RAS gene mutation was seen occasionally in benign cases such as an adenoma.4, 12 In our study, all 15 patients with adenoma failed to demonstrate molecular alteration by the markers studied. However, in a larger series of follicular lesion of undetermined significance/atypia of undetermined significance cases, some RAS-positive cytology cases would be expected to occur with follicular adenoma. Our focused study on the follicular lesion of undetermined significance/atypia of undetermined significance group showed that molecular testing had high specificity but the sensitivity was not quite as high (carcinoma from 7 of 19 patients did not show any positivity by these molecular markers). Investigation into the clinical indication for surgery in the 8 molecular negative cases (from 7 patients) was beyond the scope of the current study. Common genetic alterations in BRAF, HRAS, NRAS, KRAS, PAX8/PPARγ, RET/PTC1, and RET/PTC3 genes have been found in approximately 70% of thyroid neoplasia.3 This figure is comparable to our experience of finding a molecular alteration in 12 of 19 (63%) carcinoma patients. Given the results reported here, we can state that follicular lesion of undetermined significance/atypia of undetermined significance cases with positive molecular results are expected to have a high probability of cancer outcome, because all patients with positive molecular markers demonstrated carcinoma in the corresponding resection specimen. Conversely, follicular lesion of undetermined significance/atypia of undetermined significance cases with negative molecular results had a cancer outcome in only 7.3% of patients.

Although molecular studies alone detected more PTCs than repeat FNAs, there were 3 PTC patients that were identified by repeat FNAs and were negative by molecular methods. Overall, molecular studies were negative in 7 of 19 (37%) PTC patients, and this is in keeping with the proportion of PTCs that have been reported to be negative by the panel of common molecular markers, including BRAF, RAS, RET/PTC, and PAX8/PPARγ.3 Therefore, the combination of molecular studies and repeat FNAs increases the sensitivity of detection of PTCs. Larger studies from other institutions are necessary to verify these data and to determine whether they have therapeutic implications.

Regarding the technical aspects of the molecular procedure, we showed that molecular testing of cytology samples is feasible and provides high yield. Of the 513 follicular lesion of undetermined significance/atypia of undetermined significance cases identified, 58 cases had indeterminate molecular test results. Systematic specimen collection methodology is important in delivering consistent results. Given the proper specimen-handling modalities, our study demonstrated that the application of a panel of common molecular markers to follicular lesion of undetermined significance/atypia of undetermined significance cases was helpful in separating cases into a very high-risk group and a low-risk group.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES

The authors thank Marcia Kurs-Lasky for assistance with statistical analysis.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. CONFLICT OF INTEREST DISCLOSURES
  8. REFERENCES