Value of Molecular Tests in Cytologically Indeterminate Lesions of Thyroid.
These lesions bear remarkably different risks of malignancy, varying from 5% to 75% and, thus, pose a dilemma for patient management. Follow-up surgical excision of these cytologically indeterminate cases may yield a diagnosis of nodular hyperplasia (Figure 5), follicular adenoma, follicular carcinoma (FC; Figure 6), Hurthle cell adenoma (HcA, Figure 7), Hurthle cell carcinoma (HcC), or papillary thyroid carcinoma (PTC), either classic or follicular variant (FV-PTC, Figure 8). Many factors contribute to the uncertainty of diagnosis, including technical instability, the nature of the lesion, the quality of the specimen, and interobserver variability in morphology interpretation. A recent meta-analysis by Bongiovanni et al (3) found that the reported malignancy rates of AUS/FLUS varied from 3% to 27.2%, FN/SFN from 1.2% to 25.3%, and suspicious for malignancy from 1% to 6.3%. The reported wide ranges cannot be explained by different patient populations at different institutions but at least partially reflects the subjective difficulty in applying the diagnostic criteria at the practicing level. Such indeterminate FNA results have led to a treatment dilemma, causing many patients to undergo diagnostic surgery even though most cases turn out to be benign. (4-6) This has contributed to a recent trend of an increased number of surgeries for thyroid nodules. How to further triage the indeterminate lesions is an important challenge for cytopathologists.
Morphologic interpretation of cytologic specimens is limited by diagnostic variability. Multiple molecular tests have recently been explored in an attempt to improve the diagnostic accuracy of thyroid FNA and to reduce unnecessary surgeries. Some of these tests are available commercially. This review focuses on the advantages and limitations of the following molecular tests: immunocytochemistry (ICC) gene mutation analysis (GMA), next-generation sequencing (NGS), gene expression analysis (GEA), and microRNA analysis.
Although different opinions exist, ICC is considered by some scholars as a molecular test that detects cellular changes at molecular levels (7-9) and is included in this review for completeness. Immunocytochemistry is available in most pathology laboratories, is less time consuming, is less expensive, and practicing pathologists are confident in interpreting its results. Multiple ICC markers, including galectin-3 (GAL-3), HBME-1, CK19, CXCR4, CD44, HMGA2, MRC2, SFN, and cyclins D1 and D3, have been investigated on cell block materials individually, and all have demonstrated some value in distinguishing benign from malignant lesions. (10-14) Although the low sensitivity, specificity, positive predictive value (PPV), or negative predictive value (NPV) has rendered the use of individual markers suboptimal for practical utility in indeterminate FNA specimens, combinations of markers have yielded significantly more-reliable results. Using GAL-3, HBME-1, TPO, CK19, and KS, Saggiorato et al (15) compared the diagnostic usefulness of each individual marker, sequential staining of 2 markers with one on a single section, and staining of 2 or 3 markers on a single section (multiplex panel). The cell block sections were from a cohort of 125 thyroid FNA cases suspicious for malignancy with follow-up surgery that identified 50 follicular adenomas, 33 FCs, and 42 FV-PTCs. The study found that the 2-marker panel of GAL-3 and HBME-1 significantly improved the NPV over that of single markers (96% versus 89% and 76%, respectively), with the highest accuracy of 94%. The 3-marker panel of GAL-3, HBME-1, and CK19 had lower specificity (82% versus 90%) and lower PPV (79% versus 94%) compared with the 2-marker panel. The authors concluded that using GAL-3 and HBME-1 sequentially (testing HBME-1 as a second marker whenever GAL-3 staining is negative) yielded the best results with the highest combination of sensitivity (97%) and specificity (95%). The authors also analyzed 24 of the 125 cases (19%) that exhibited oncocytic features. The 2-marker panel of GAL-3 and CK19 was discovered to be the best at identifying oncocytic malignancy with 100% sensitivity, specificity, PPV, NPV, and highest accuracy, whereas the same panel demonstrated significantly lower accuracy in nononcocytic tumors. The study suggests that the Hurthle cell tumors have unique molecular markers and may be considered a separate group of FNs. In addition, Troncome et al (16) found that the panel of cyclins D1 and D3 could reliably distinguish benign from malignant HcNs in a cohort of 51 cases with a cytologic diagnosis of "suspicious for malignancy.* Furthermore, Cochand-Priollet et al (17) tested the panel of CK19 and HBME-1 in a cohort of 48 FNA-indeterminate lesions, consisting of a mixture of AUS/ FLUS, FN/SFN, and suspicious for malignancy, and found that the panel could separate benign from malignant tumors with 100% sensitivity and 100% NPV when the cutoff for positivity was set at more than 70% of tumor cells being stained positive. Such a panel, therefore, may identify cases with low malignant potential and help reduce unnecessary aggressive management. However, the specificity (85.2%) and PPV (86.2%) of the CK19 and HBME-1 panel were only moderately acceptable. In summary, the GAL-3 and HBME1 sequential stain protocol (reflex to HBME-1 when results for GAL-3 are negative) is worth recommending for its high sensitivity and specificity.
A major concern of the published ICC studies are the few cases in each cohort, and further validation has not been performed in large-scale prospective studies. In addition, intrinsic impediments of ICC are poor reproducibility in staining and interobserver variability in interpretation, which have prevented its widespread use in differentiating benign from malignant lesions in the daily practice of cytopathology. However, specific ICC markers are available to confirm a diagnosis of uncommon thyroid tumors, such as calcitonin for medullary carcinoma, parathyroid hormone for parathyroid tumor, and lymphoid markers for lymphoma and to work up metastatic carcinomas in the thyroid.
GENE MUTATION ANALYSIS
A number of specific gene mutations have been identified in thyroid cancers, including point mutations of B-Raf proto-oncogene, serine/threonine kinase (BRAF) and RAS viral (r-ras) oncogne homolog (RAS), and gene rearrangements of ret proto-oncogene (RET/PTC) and paired box 8/ peroxisome proliferator activated receptor G (PAX8/PPARG). BRAF and RET/PTC mutations are found in up to 69% and 20% of classic PTC, respectively, and RAS and PAX8/PPARG mutations are found in up to 53% and 63% of FC. (18,19) However, PAX8/PPARG and RAS mutations are also found in up to 31% and 48% of follicular adenomas. (18,20) Some of these mutations, such as BRAF, are highly specific for malignant thyroid tumors and have very high PPV in identifying malignancy, whereas others, such as RAS, have much lower specificity. The value of these mutations in distinguishing benign from malignant thyroid tumors has been explored. Although single GMA suffers from low sensitivity, assays employing multiple GMAs (gene classifier) have demonstrated greater potential. (21-24) Nikiforov et al (25) studied a 4-gene panel consisting of BRAF V600E, neuroblastoma RAS viral (v-ras) oncogene homlog (NRAS), Harvey rat sarcomal viral oncogene homolog (HRAS), and Kirsten rat sarcoma viral oncogene homolog (KRAS) point mutations, RET/PTC gene rearrangements, and PAX8/ PPARG gene rearrangement in 513 indeterminate aspirates with surgical follow-up. This gene panel had high specificity and PPV in distinguishing benign from malignant tumors in all subgroups of 247 AUS/FLUS, 214 FN/SFN, and 52 suspicious for malignancy cases. However, the sensitivity (57%-68%) and NPV (72%-94%) were particularly low in the FN/SFN and suspicious for malignancy subgroups. Nevertheless, the authors proposed that any positive result in this panel should be considered an indication for total thyroidectomy in all categories of indeterminate cytology. Even if gene mutations were not identified, patients with FN/SFN or diagnoses as suspicious for malignancy were still recommended for lobectomy, and those with AUS/FLUS might go on for lobectomy or might be watched with or without repeat FNA. This 4-gene test, therefore, does not significantly reduce the number of surgeries. The high specificity and PPV make the 4-gene assay more suitable for confirming malignancy (rule-in test), whereas its ability to exclude a malignancy is still questionable.
Asuragen, Inc (Austin, Texas), has developed a commercial product named miRInform Thyroid, which employs the same 4-gene classifier studied by Nikiforov and colleagues. (25) It detects 17 known thyroid cancer gene mutations and translocations. Patients with an indeterminate cytology diagnosis would be referred for total thyroidectomy if any of the gene abnormalities is identified, whereas those with a negative result can be managed conservatively by clinical observation, repeat of the FNA procedure, or lobectomy. An additional FNA pass is required and the sample has to be shipped to the company. It costs more than $2000 and has not gained US Food and Drug Administration (FDA) approval yet. It may be reasonable to consider this assay for cases in which a patient doubts the cytology diagnosis of FN/SFN or suspicious for malignancy, hesitates to go for surgery, or desires confirmation of cytology results before surgery.
Next-generation sequencing technology is part of the emerging field of innovative and personalized health care. Clinical implementation of the technology is rapidly evolving for the diagnosis of genetic diseases and in management of advanced cancers. Next-generation sequencing provides simultaneous analysis of large regions of the genome with high sensitivity for detection of mutations and quantitative assessment of mutant alleles. Interpretation of results and data reporting are improving. (26) Nikiforova et al (27) tested the accuracy of their NGS custom panel in 228 thyroid neoplastic and nonneoplastic samples, including 177 excisional specimens and 51 FNA samples, with follow-up surgical diagnosis. The panel was designed to target 12 cancer genes with 284 mutational hot spots in thyroid cancer, and the analytic accuracy for mutation detection was 100%, with a sensitivity of 3% to 5% for the mutant allele. The assay successfully identified mutations in 70% of classic PTC, 83% of FV-PTC, 78% of FC, and 39% of HcC, which can be contrasted with the 6% mutation-detection rate in benign thyroid nodules. However, neither the percentage nor the case number of the FNA specimens with an indeterminate cytology diagnosis was specified in the report, and the FNA samples were not analyzed separately.
This NGS panel has been converted into a commercial product under the name of ThyroSeq marked by CBLPath (Rye Brook, New York). It requires only 1 to 2 drops of FNA aspirate or 10 ng of DNA. Additionally, RNA is also used for reverse transcriptase-polymerase chain reaction analysis of RET/PTC and PAX8/PPARG gene rearrangements. Nikiforov et al (28) have recently reported an assessment of 143 FN/SFN cytology specimens collected retrospectively and prospectively, using ThyroSeq (version 2) NGS, which simultaneously tests for point mutations in 13 genes and 42 types of gene fusions identified in thyroid cancer. The results demonstrated remarkably better sensitivity (90% versus 57%), NPV (96% versus 86%), and accuracy (92% versus 86%) when compared with the results of the 4-gene classifier analysis. It would be interesting to investigate whether and how the assay improves sensitivity, NPV, and accuracy in the AUS/FLUS and suspicious for malignancy groups. So far, there is no guideline provided by the vendor or professional societies regarding its indication in cytology samples. Because of its relatively high costs, the main use of the assay may be restricted to the indeterminate group of cytology cases. (29)
GENE EXPRESSION ANALYSIS
Although expression of single genes, such as HMGA2 and UbcH10, has been evaluated to determine the potential to separate benign from malignant tumors in thyroid FNA aspirates, the application of GEA has been limited by low sensitivity and specificity. (30,31) However, an analysis of multiple gene expressions has provided a means to improve diagnostic accuracy. (11,32) One example is the Afirma thyroid FNA analysis (Veracyte, Inc, South San Francisco, California). The Gene Expression Classifier (GEC; Veracyte) measures the expression of 167 genes (including 25 genes filtering for rare thyroid tumors) and applies a multidimensional algorithm to reclassify a nodule with cytologic indeterminate diagnosis as benign or suspicious. The GEC was validated by a multicenter study (33) in a cohort of 265 indeterminate thyroid nodules and demonstrated high sensitivity (92%) and NPV (93%) but low specificity (52%) and PPV (47%). The assay appears valuable for excluding malignancy (as a rule-out test), particularly in the AUS/ FLUS group, which bears low risk of malignancy (5%-15%) and carries the largest portion of the cytologically indeterminate lesions. In fact, the company recommends that the assay should be limited to the indeterminate lesions of AUS/ FLUS and FN/SFN, including HcN, lesions 1 cm or larger, and patients 21 years or older. Lesions suspicious for malignancy were not included in the recommendation. The test is not yet approved by the FDA and costs more than $3000 per specimen. A practical question is whether repeat FNA or GEA as a reflex test should be recommended for indeterminate lesions. In the abovementioned validation study, the assay accurately subclassified 55 of 129 (43%) of the AUS/FLUS cases as benign, a percentage similar to the result of a study (34) on repeated FNA, which yielded 49% benign cases, and only 19% remained as AUS/FLUS in a cohort of 3589 thyroid aspirates. It, therefore, seems reasonable to use GEA for FN/SFN cases and for AUS/ FLUS lesions diagnosed on repeat FNA materials. A recent multicenter analysis (35) of 339 cytologically indeterminate nodules with the Afirma GEC confirmed its substantial effect on clinical care recommendations.
Both Afirma GEC and miRInform Thyroid have their own difficulties in successfully subclassifying thyroid indeterminate lesions. The information they provide is complimentary to each other, and the use of both assays seems to improve diagnostic accuracy. Kloos et al (36) recently added BRAF V600E detection into the Afirma GEC and tested it in a cohort of 208 thyroid indeterminate aspirates that included 95 cases of AUS/FLUS, 70 cases of FN/SFN, and 43 cases of suspicious for malignancy. The addition of BRAF GMA did not improve the sensitivity and specificity of the GEC test. One of the reasons may be that the BRAF mutation was only identified in 10% of the cases, which were all classified as suspicious by Afirma GEC alone. It is probably also due to the low BRAF V600E prevalence in FV-PTC, FC, and HcC, which constitutes the major portion of malignancy in the cytologically indeterminate group and are likely to be missed by GEC. (25,33) Therefore, addition of BRAF V600E alone may not increase the sensitivity of the assay. However, few cases were studied and only the BRAF mutation was included in the evaluation. Investigation to further explore the efficiency of combined GEC with the GEC is still justified.
MicroRNA is a short, single-stranded, noncoding RNA, consisting of 19 to 23 nucleotides. It regulates messenger RNA by binding to the 3' nontranslated region and can function as tumor suppressor or oncogene. As a small molecule, it is relatively stable in FNA specimens and in formalin-fixed, paraffin-embedded cell-block materials and is, therefore, amenable to detection. Specific microRNA profiles (panels) have been identified in a variety of tumors, and their diagnostic utility has also been explored in thyroid neoplasms, including FNA indeterminate lesions (37-41) (Table). The panels are usually small, consisting of 2 to 4 microRNAs, and their sensitivity, specificity, and accuracy are improving as different studies on microRNAs become available. Keutgen et al (40) found that a panel of 4 microRNAs consisting of miR-222, miR-328, miR-197, and miR-21 classified thyroid indeterminate lesions into benign and malignant with 100% sensitivity, 86% specificity, and 90% accuracy in a cohort of 72 cases. Further analysis indicated that the panel was suboptimal for HcN, and when 13 of those cases were excluded from the previous study, the sensitivity and accuracy were increased to 100% and 95%, respectively. Dettmer et al (41) identified a 3 microRNA classifier consisting of miR-885-5p, miR-221, and miR574-3p that successfully classified 19 indeterminate aspirates with 100% accuracy. It is worth mentioning that no gene mutation was identified in any case of this cohort, including 7 Hurthle cell carcinomas, 1 follicular carcinoma, and 11 hyperplastic nodules. In addition, miR-885-5p was found to be a specific marker for Hurthle cell carcinoma in tissue materials, as its upregulation was identified in 15 of 17 (88%) of the Hurthle cell carcinomas, but in only 1 of 21 (4.8%) of the follicular carcinomas.
In a separate study of excised tumors, the same group found that highly upregulated MiR-375 could be a novel marker of PTC, and that upregulation of miR-181a-2-3p and miR-99-3p was associated with relapse-free survival in patients with FV-PTC. (42) MicroRNA profiling may, therefore, have the potential to identify the subtypes of thyroid cancer in tissue sections and may serve as prognostic indicators for patients. It would be interesting to see whether the same microRNA profile works in cytology specimens.
MicroRNA has great potential, but further study is required to validate the findings before it is used for clinical application. To date, there is no commercially available microRNA test.
Indeterminate thyroid FNA results present a challenge to the clinical management of thyroid nodular lesions. Molecular methods have the potential to provide a means for risk stratification of these lesions. However, the advantages and limitations of each test must be enumerated and an appropriate assay chosen to meet the needs of individual patients. Immunocytochemistry is simple and fast but is flawed intrinsically by technical inconsistency and interobserver variability. Its value is best appreciated in identifying uncommon thyroid and parathyroid tumors and in ruling out metastatic carcinomas. Gene mutation analysis, such as miRInform Thyroid, intended to confirm a malignancy, is more helpful for patients with FN/SFN and lesions that are suspicious for malignancy. The NGS technology seems to have the power to advance cytopathology to a new level, but its application in daily practice remains to be further clarified. The GEA, such as Afirma GEC, is designed to exclude malignancy and to reduce unnecessary surgeries and, therefore, would be more helpful for patients with AUS/FLUS, FN/SFN, and HcN/ SHcN. However, justification for making Afirma GEC a reflex test is still questionable and further validation is required. Although, so far, there are no available guidelines on selecting the most appropriate molecular test for cytologically indeterminate lesions, and a decision has to be made on a case-by-case basis in the best interest of the patient and cost efficiency, my recommendation is to consider the miRInform Thyroid assay for cases in which a patient doubts the cytology diagnosis of FN/SFN or suspicious for malignancy, hesitates to go for surgery, or desires confirmation of cytology results before surgery, whereas Afirma GEC should be reserved for those patients who have repeated AUS/FLUS lesions. MicroRNA panels have demonstrated great potential in classifying thyroid indeterminate lesions into management-related subgroups and even in serving as prognostic indictors, but further study and validation are required before they can be used in the laboratory. Although more questions have to be answered before the molecular assays become widely accepted, this direction seems the right one for our patients.
The author is grateful to Gordon Pringle, DDS, at Temple University School of Medicine, Pennsylvania, and Lanjing Zhang, MD, at the University Medical Center of Princeton at Plainsboro and Rutgers University, New Jersey, for their critical review and improvement of the manuscript.
Please Note: Illustration(s) are not available due to copyright restrictions.
(1.) Gharib H, Goellner. Fine-needle aspiration biopsy of the thyroid. Ann Intern Med. 1993; 118(4):282-289.
(2.) Syed AZ, Cibas ES. The Bethesda system for reporting thyroid cytopathol ogy. Am J Clin Pathol. 2009; 132(5):658-665.
(3.) Bongiovanni M, Spitale A, Faquin WC, Mazzucchelli L, Balock ZW. The Bethesda system for reporting thyroid cytopathology: a meta-analysis. Acta Cytol. 2012; 56(4):333-339.
(4.) Ghanem M, Levy Y, Mazeh H. Preoperative diagnosis of benign thyroid nodules with intermediate cytology. Cland Surg. 2012; 1(2):89-91.
(5.) Calo PG, Medas F, Santa Cruz R, et al. Follicular nodules (Thy3) of the thyroid: is total thyroidectomy the best option? BMC Surg. 2014:14:12. doi:10. 1186/1471-2482-14-12.
(6.) Chung YS, Yoo C, Jung JH, Choi HJ, Suh YJ. Review of atypical cytology of thyroid nodules according to the Bethesda system and its beneficial effect in the surgical treatment of papillary carcinoma. J Korean Surg Soc. 2011; 81(2):75-84.
(7.) Yip L. Molecular diagnostic testing and the indeterminate thyroid nodule. Curr Opin Oncol. 2014; 26(1):8-13.
(8.) Keutgen XM, Filicori F, Fahey TJ III. Molecular diagnosis for indeterminate thyroid nodules on fine needle aspiration. Expert Rev Mol Diagn. 2013; 13(6): 613-623.
(9.) Kouniavsky G, Zeiger MA. The quest for diagnostic molecular markers for thyroid nodules with indeterminate or suspicious cytology. J Surg Oncol. 2012; 105(5):438-443.
(10.) Das DK, Al-Waheeb SK, George SS, Haji BI, Mallik MK. Contribution of immunocytochemical stainings for galectin-3, CD44, and HBME1 to fine-needle aspiration cytology diagnosis of papillary thyroid carcinoma. Diagn Cytopathol. 2014; 42(6):498-505.
(11.) Prasad N, Kowalski J, Tsai HL, et al. Three-gene molecular diagnostic model for thyroid cancer. Thyroid. 2012; 22(3):275-284.
(12.) Torregrossa L, Faviana P, Filice ME, et al. CXC chemokine receptor 4 immunodetection in the follicular variant of papillary thyroid carcinoma: comparison to galectin-3 and Hector Battifora mesothelial cell-1. Thyroid. 2010; 20(5):495-504.
(13.) Bartolazzi A, Orlandi F, Saggiorato E, et al; Italian Thyroid Cancer Study Group (ITCSG). Galectin-3-expression analysis in the surgical selection of follicular thyroid nodules with indeterminate fine-needle aspiration cytology: a prospective multicentre study. Lancet Oncol. 2008; 9(6):543-549.
(14.) Khurana KK, Truong LD, LiVolsi VA, Baloch ZW. Cytokeratin 19 immunolocalization in cell block preparation of thyroid aspirates: an adjunct to fine-needle aspiration diagnosis of papillary thyroid carcinoma. Arch Pathol Lab Med. 2003; 127(5):579-583.
(15.) Saggiorato E, De Pompa R, Volante M, et al. Characterization of thyroid 'follicular neoplasms' in fine-needle aspiration cytological specimens using a panel of immunohistochemical markers: a proposal for clinical application. Endocr Relat Cancer. 2005; 12(2):305-317.
(16.) Troncone G, Volante M, Iaccarino A, et al. Cyclin D1 and D3 overexpression predicts malignant behavior in thyroid fine-needle aspirates suspicious for Hurthle cell neoplasms. Cancer. 2009; 117(6):522-529.
(17.) Cochand-Priollet B, Dahan H, Laloi-Michelin M, et al. immunocytochemistry with cytokeratin 19 and anti-human mesothelial cell antibody (HBME-1) increases the diagnostic accuracy of thyroid fine-needle aspirations: preliminary report of 150 liquid-based fine-needle aspirations with histological control. Thyroid. 2011; 21(10):1067-1073.
(18.) Kondo T, Ezzat S, Asa SL. Pathogenetic mechanisms in thyroid follicularcell neoplasia. Nat Rev Cancer. 2006; 6(4):292-306.
(19.) Nikiforov YE. Molecular diagnostics of thyroid tumors. Arch Pathol Lab Med. 2011:135(5):569-577.
(20.) Nikiforova MN, Lynch RA, Biddinger PW, et al. RAS point mutations and PAX8-PPAR[gamma] rearrangement in thyroid tumors: evidence for distinct molecular pathways in thyroid follicular carcinoma. J Clin Endocrinol Metab. 2003; 88(5): 2318-2326.
(21.) Kleiman DA, Sporn MJ, Beninato T, et al. Preoperative BRAF (V600E) mutation screening is unlikely to alter initial surgical treatment of patients with indeterminate thyroid nodules: a prospective case series of 960 cases. Cancer. 2013; 119(8):1495-1502.
(22.) Suh I, Kebebew E. The biology of thyroid oncogenesis. Cancer Treat Res. 2010; 153:3-21.
(23.) Marques AR, Espadinha C, Catarino AL, et al. Expression of PAX8-PPAR gamma 1 rearrangements in both follicular thyroid carcinomas and adenomas. J Clin Endocrinol Metab. 2002; 87(8):3947-3952.
(24.) Gomberawalla A, Elaraj DM. Howto use molecular testing results to guide surgery: a surgeon's perspective. Curr Opin Oncol. 2014; 26(2):14-21.
(25.) Nikiforov YE, Ohori NP, Hodak SP, et al. Impact of mutational testing on the diagnosis and management of patients with cytologically indeterminate thyroid nodules: a prospective analysis of 1056 FNA samples. Clin Endocrinol Metab. 2011; 96(11):3390-3397.
(26.) Roy S, Durso MB, Wald A, Nikiforov YE, Nikiforova MN. SeqReorter: automating next-generation sequencing result interpretation and reporting workflow in a clinical laboratory. J Mol Diagn. 2014; 16(1):11-22.
(27.) Nikiforova ME, Wald AI, Roy S, Durso MB, Nikiforov YE. Targeted next-generation sequencing panel (ThyroSeq) for detection of mutations in thyroid cancer. J Clin Endocrinol Metab. 2013; 98(11):E1852-E1860. doi:10.1210/jc. 2013-2292.
(28.) Nikiforov YE, Carty SE, Chiosea SI, et al. Highly accurate diagnosis of cancer in thyroid nodules with follicular neoplasm/suspicious for a follicular neoplasm cytology by ThyroSeq v2 next-generative sequencing assay. Cancer. 2014; 120(23):3627-3634.
(29.) Hershman JM. Next-generation sequencing has identified new oncogenic mutations in thyroid nodules. Clin Thyroidal. 2013; 25(12):288-289.
(30.) Jin L, Lloyd RV, Nassar A, et al. HMGA2 expression analysis in cytological and paraffin-embedded tissue specimens of thyroid tumors by relative quantitative RT-PCR. Diagn MolPathol. 2011; 20(2):71-80.
(31.) Guerriero E, Ferraro A, Desiderio D, et al. UbcH10 expression on thyroid fine-needle aspirates. Cancer Cytopathol. 2010; 118(3):157-165.
(32.) Chudova D, Wilde JI, Wang ET, et al. Molecular classification of thyroid nodules using high-dimensionality genomic data. J Clin Endocrinol Metab. 2010; 95(12):5296-5304.
(33.) Alexander EK, Kennedy GC, Baloch ZW, et al. Preoperative diagnosis of benign thyroid nodules with indeterminate cytology. N Engl J Med. 2012; 367(8): 705-715.
(34.) Yassa L, Cibas ES, Benson CB, et al. Long-term assessment of a multidisciplinary approach to thyroid nodule diagnostic evaluation. Cancer. 2007; 111 (6):508-516.
(35.) Alexander EK, Schorr M, Klopper J, et al. Multicenter clinical experience with the Afirma gene expression classifier. J Clin Endocrinol Metab. 2014; 99(1): 119-125.
(36.) Kloos RT, Reynolds JD, Walsh PS, et al. Does addition of BRAF V600E mutation testing modify sensitivity or specificity of the Afirma gene expression classifier in cytologically indeterminate thyroid nodules? J Clin Endocrinol Metab. 2013; 98(4):E761-E768.
(37.) Kitano M, Rahbari R, Patterson EE, et al. Evaluation of candidate diagnostic microRNAs in thyroid fine-needle aspiration biopsy samples. Thyroid. 2012;
(38.) Agretti P, Ferrarini E, Rago T, et al. MicroRNA expression profile helps to distinguish benign nodules from papillary thyroid carcinomas starting from cells of fine-needle aspiration. Eur J Endocrinol. 2012; 167(3):393-400.
(39.) Shen R, Liyanarachchi S, Li W, et al. MicroRNA signature in thyroid fine needle aspiration cytology applied to *atypia of undetermined significance* cases. Thyroid. 2012; 22(1):9-16.
(40.) Keutgen XM, Fillicori F, Crowley MJ, et al. A panel of four miRNAs accurately differentiates malignant from benign indeterminate thyroid lesions on fine needle aspiration. Clin Cancer Res. 2012; 18(7):2032-2038.
(41.) Dettmer M, Vogestseder A, Durso MB, et al. MicroRNA expression array identified novel diagnostic markers for conventional and oncocytic follicular thyroid carcinomas. J Clin Endocrinol Metab. 2013; 98(1):E1-E7.
(42.) Dettmer M, Perren A, Moch H, Komminoth P, Nikiforov YE, Nikiforova MN. Comprehensive microRNA expression profiling identified novel markers in follicular variant of papillary thyroid carcinoma. Thyroid. 2013; 23(11):1383-1389.
Xinmin Zhang, MD
Accepted for publication January 26, 2015.
From the Department of Pathology and Laboratory Medicine, Temple University School of Medicine, Philadelphia, Pennsylvania.
The author has no relevant financial interest in the products or companies described in this article.
Presented at The Princeton Integrated Pathology Symposium: Head and Neck Pathology; February 9, 2014; Princeton, New Jersey.
Reprints: Xinmin Zhang, MD, Department of Pathology and Laboratory Medicine, Temple University School of Medicine, B243 Outpatient Building, 3401 N Broad St, Philadelphia, PA 19140 (e-mail: Xinmin.Zhang@tuhs.temple.edu).
Caption: Figure 1. Atypia of undetermined significance/follicular lesion of undetermined significance. Hypocellular specimen with a few loose clusters of microfollicles, showing mild nuclear enlargement and scanty colloid (Papanicolaou, original magnification x200).
Caption: Figure 2. Suspicious for follicular neoplasm. Hypercellular specimen with almost pure microfollicles and scanty colloid (Diff-Quick, original magnification X200).
Caption: Figure 3. Suspicious for Hurthle cell neoplasm. Cellular specimen with almost pure Hurthle cells and minimal cytologic atypia (Papanicolaou, original magnification X200).
Caption: Figure 4. Suspicious for papillary thyroid carcinoma. Rare clusters of follicular cells with nuclear crowding, enlargement, ground-glass appearance, and grooving. No intranuclear inclusions were identified (Papanicolaou, original magnification X400).
Caption: Figure 5. Follow-up thyroid excision of the case in Figure 1, showing thyroid nodular hyperplasia with a predominant microfollicular growth pattern (hematoxylin-eosin, original magnification X100).
Caption: Figure 6. Follow-up thyroid excision of the case in Figure 2, showing minimally invasive follicular carcinoma with focal capsule invasion (hematoxylin-eosin, original magnification X100).
Caption: Figure 7. Follow-up thyroid excision of the case in Figure 3 showing Hurthle cell adenoma with intact and uniformly thickened capsule (hematoxylin-eosin, original magnification X100).
Caption: Figure 8. Follow-up thyroid excision of the case in Figure 4 showing the follicular variant in papillary thyroid carcinoma (hematoxylin-eosin, original magnification X200).
MicroRNA Profiling in Indeterminate Thyroid Cytology Samples Study, FNA (a) Cases, SN, SP, PPV, No. MicroRNA Panel No. % % % 1 MiR-7, MiR-126, 52 100 20 25 MiR-374a, Let-7g 2 MiR-146b, MiR-155, 53 60 58 36 MiR-1221 3 MiR-146b, MiR-221, 30 66 79 64 MiR-187, MiR-30d 4a MiR-222, MiR-328, 72 100 86 NA MiR-197, MiR-21 4b Same panel as 4a 59 100 97 NA 5 Mir-885-5p, MiR-221, 19 NA NA NA MiR-574-3p Study, NPV, Accuracy, No. % % Source, y 1 100 37 Kitano et al, (37) 2012 2 79 58 Agretti et al, (38) 2012 3 79 73 Shen et al, (39) 2012 4a NA 90 Keutgen et al, (40) 2012 4b NA 95 Keutgen et al, (40) 2012 5 NA 100 Dettmer et al, (41) 2012 Study, No. Comments 1 Calculation based on miR-7 2 Calculation per table 2 data in Agretti et al (38) 3 Calculated on AUS cases in the validation set 4a Including 13 Hurthle cell lesions 4b Non-Hurthle cell follicular lesions 5 Surgical follow-up: 11 HNs, 7 HcCs, 1 FC Abbreviations: AUS, atypia of undetermined significance; FC, follicular carcinoma; FNA, fine-needle aspiration; HcC, Hurthle cell carcinoma; HN, hyperplastic nodules; NA, not available; NPV, negative predictive value; PPV, positive predictive value; SN, sensitivity; SP, specificity. (a) Number of FNA indeterminate cases.