Thyroid nodules are currently found in up to 67% of the

Transcription

1 Original Research Endocrine Surgery Diagnostic Value of RAS Mutations in Indeterminate Thyroid Nodules: Systematic Review and Meta-analysis Otolaryngology Head and Neck Surgery 2017, Vol. 156(3) Ó American Academy of Otolaryngology Head and Neck Surgery Foundation 2017 Reprints and permission: sagepub.com/journalspermissions.nav DOI: / William Clinkscales 1, Adrian Ong, MD 1, Shaun Nguyen, MD, MA 1, Elizabeth Emily Harruff 1, and Marion Boyd Gillespie, MD, MSc 1 Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article. Abstract Objectives. To determine the diagnostic value of HRAS, KRAS, and NRAS mutations in fine-needle aspiration biopsies of thyroid nodules that are nondiagnostic on cytology. Data Sources. PubMed, Scopus, Embase, CINAHL. Review Methods. Two authors independently searched the data sources. To be included, studies reported the RAS mutational status and postoperative histopathologic diagnosis of nodules that exhibited indeterminate cytology after fineneedle aspiration biopsy. Data were extracted to calculate sensitivity, specificity, and positive/negative predictive values of any HRAS, KRAS, or NRAS mutation. A meta-analysis was performed to generate pooled values for each parameter. Results. A total of 7 studies with a combined 1025 patients met inclusion criteria. The pooled sensitivity of a RAS mutation for detecting cancer was (95% confidence interval [95% CI], ), while the pooled specificity was (95% CI, ). The weighted averages for positive predictive value and negative predictive value were 78.0% and 64.0%, respectively, with 68.0% accuracy. The positive likelihood ratio was (95% CI, ), and the negative likelihood ratio was (95% CI, ). Conclusion. Our data suggest that testing for any RAS mutation is unlikely to change the clinical management of thyroid nodules that have indeterminate cytology. While a RAS mutation may rule in malignancy, the sensitivity of testing is low enough to merit further mutational analysis, repeat fineneedle aspiration, or surgical excision, even in the presence of a negative test. Keywords thyroid, carcinoma, RAS, fine-needle aspiration, biopsy Received August 20, 2016; revised November 17, 2016; accepted December 2, Thyroid nodules are currently found in up to 67% of the general population. 1 While the majority of these are benign, the remaining malignancies represent a significant diagnostic challenge. For nodules with ultrasound features concerning for malignancy, the mainstay evaluation is fineneedle aspiration (FNA) and subsequent cytologic evaluation. Unfortunately, 20% to 30% of FNA cytology cases are categorized as 1 of 3 indeterminate categories according to the Bethesda System for Reporting Cytopathology. 2 These classes include the following: class III atypia of undetermined significance/follicular lesion of undetermined significance; class IV follicular neoplasm/suspicious for follicular neoplasm, follicular neoplasm of Hurthle cell type; and class V suspicious for malignancy. 3 These classes denote an ambiguous group of nodules that may or may not harbor malignancy and currently require surgical resection for definitive diagnosis. However, indeterminate nodules without atypia are considered to have only a 20% to 30% risk of malignancy, 4,5 and only 5% to 15% of atypia of undetermined significance/follicular lesion of undetermined significance nodules are found to be malignant postoperatively. 6,7 Furthermore, 70% to 85% of cytologically indeterminate thyroid nodules that are excised are found to be benign and thus, in hindsight, place an unwarranted risk and cost burden on the patient Consequently, the hunt for reliable preoperative molecular tumor markers has become paramount. Up to 70% of thyroid cancers harbor at least 1 known genetic mutation. 11 Ideally, detection of genomic alterations could be used to characterize indeterminate nodules and prevent or limit unnecessary surgical management. Point mutations in 1 of the 3 RAS genes, NRAS, KRAS, and HRAS, are currently the second-most common genetic alteration in 1 Department of Otolaryngology Head and Neck Surgery, Medical University of South Carolina, Charleston, South Carolina, USA This article was presented at the 2016 AAO-HNSF Annual Meeting & OTO EXPO; September 18-21, 2016; San Diego, California. Corresponding Author: William Clinkscales, Department of Otolaryngology Head and Neck Surgery, Medical University of South Carolina, 135 Rutledge Avenue, MSC 550, Charleston, SC 29425, USA. clinksc@musc.edu

2 Clinkscales et al 473 thyroid cancer 12,13 the first being mutations of B-RAF. Furthermore, the increase in the incidence of thyroid cancer in the last 40 years has been attributed in part to the rise in the percentage of RAS-mutated tumors. 14 For this reason, a great effort has been made to determine how these mutations can be used to guide management when the clinical picture and cytology remain equivocal. A number of studies have sought to define the diagnostic utility of RAS mutations, both in isolation and as part of larger panels of molecular tests. In the present study, we sought to systematically search the literature and identify these studies to reliably summarize these findings. A similarly designed study concluded that B-RAF V600E testing is a potentially useful adjunct in evaluating indeterminate nodules. 15 To our knowledge, this is the first meta-analysis on the diagnostic value of RAS mutations in cytologically indeterminate thyroid nodules. Methods Two authors (W.C., A.O.) independently searched 5 databases (PubMed, Scopus, Embase, Cochrane, CINAHL) in January 2016 using the following strategy: thyroid AND RAS AND (biopsy OR aspiration). Results were restricted to the English language. Articles were initially screened by title and abstract, and the full text of reviewed articles was selected for inclusion according to predetermined criteria. To be included, studies must have reported the mutational status of at least 1 nucleotide point mutation within each of the 3 RAS genes (HRAS, NRAS, and KRAS), in addition to the postoperative histopathologic diagnosis of nodules that exhibited indeterminate cytology after FNA biopsy. For the purposes of this study, any FNA cytology was considered indeterminate if, according to the Bethesda system, it was classified as class III atypia of undetermined significance/ follicular lesion of undetermined significance, class IV follicular neoplasm, or class V suspicious for malignancy. 3 Manuscripts were excluded if the indeterminate nodules studied were not from the thyroid gland, they only represented 1 or 2 cytologic classes, they only represented confirmed carcinoma cases, or the mutational status of HRAS, NRAS, and KRAS was not investigated. Case reports, animal studies, and reviews were excluded as well. No date restrictions were used during search of each database. Differences in the included articles were settled by discussion among the authors. Details of the literature search can be seen in the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) review diagram (Figure 1). Data from the included studies were extracted independently with a standardized form. The primary outcomes of interest were sensitivity, specificity, and positive/negative likelihood ratios of any RAS mutation (including HRAS, KRAS, and NRAS) in indeterminate thyroid nodules in the diagnosis of thyroid carcinoma. When appropriate data were available, we calculated the rate of true positives, true negatives, false positives, and false negatives for each study. Biostatistics All analyses and graphs were performed with statistical software (Sigma Stat 3.0 and Sigma Plot 8.0; SPSS, Chicago, Illinois) and MedCalc (MedCalc Software bvba, Mariakerke, Belgium). The main outcome measures included pooled estimation of sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and accuracy. Between-study heterogeneity was evaluated by Cochran s Q test (P \.05 or I 2. 50%). Additionally, the Sterne and Egger tests were performed for further assessment of risk of publication bias. 16 In addition, a meta-analysis of proportions was done for sensitivity, specificity, positive predictive value, negative predictive value, accuracy, and mutation incidence. Data were presented as weighted proportions with corresponding 95% confidence intervals (95% CIs). The pooled proportion (with 95% CI) is given for the random effects model. MedCalc used a Freeman-Tukey transformation to calculate the weighted summary proportion under the fixed and random effects models. 17,18 A P value of.05 was considered significant for all statistical tests. A meta-analysis of area under the receiver operating characteristic curve was not possible due to lack of data. Results Study Identification A total of 7 studies with 1025 patients met the inclusion criteria. Details of the individual studies can be seen in Table 1. These studies presented low risk of bias and were analyzed further. However, Sterne and Egger testing (P \.0001) suggested a relationship between the sample size of these studies and their effect sizes. The I 2 statistic of 93.11% (95% CI: 88.28%-95.95%) also suggested significant study heterogeneity (Figure 2). Thus, meta-analysis of these 7 studies was performed with a random effects model. A number of studies with similar aims were not included, for the following reasons: included FNA samples with only 1 or 2 Bethesda classes, tested for mutations in \3 oftheras genes, 11,30-36 and evaluated limited subtypes of cancer. 32,37-44 In the included studies, the following codons were probed for point mutations: KRAS codons 12, 13, 19, 61; NRAS codons 12, 13, 61; HRAS codons 12, 13, 61. There was some heterogeneity among the studies in the combination of codons that were tested. All 7 studies tested for mutations in codon 61 of HRAS and NRAS. Five studies tested for mutations in KRAS codons 12 and 13 in addition to codon 61 of HRAS and NRAS. Diagnostic Accuracy RAS testing for indeterminate thyroid nodules had a relatively low pooled sensitivity (0.343, 95% CI: ) and high pooled specificity (0.935, 95% CI: ). The weighted averages for positive predictive value and negative predictive value were 78.0% and 64.0%, respectively, with 67.9% accuracy. Finally, the pooled positive likelihood ratio and negative likelihood ratio of RAS testing

3 474 Otolaryngology Head and Neck Surgery 156(3) Figure 1. PRISMA review diagram. Table 1. Included Studies with RAS Diagnostic Parameters. Study n TP TN FP FN SE SP PPV NPV INC An (2015) Beaudenon-Huibregtse (2014) Cantara (2010) Eszlinger (2015) Nikiforov (2011) Nikiforov (2009) Stence (2015) Abbreviations: FN, false negative; FP, false positive; INC, mutation incidence; NPV, negative predictive value; PPV, positive predictive value; SE, sensitivity; SP, specificity; TN, true negative; TP, true positive. were (95% CI: ) and (95% CI: ), respectively (Figures 3-6). Discussion Current guidelines offer only weak recommendations (moderate evidence) to supplement management of indeterminate nodules with molecular markers. 9 This weakness most likely stems from the heterogeneous validity of such tests reported in the literature. The sensitivity of 7-gene mutational panel tests, which include RAS mutations, ranges from 44% to 100% After BRAF mutations, RAS mutations are the next-most common genetic alteration noted in thyroid carcinomas. 45 In this study, we sought to assess the diagnostic accuracy of detecting RAS mutations in indeterminate thyroid nodules by systematically searching the literature and combining relevant data in 1 meta-analysis.

4 Clinkscales et al 475 Figure 2. Funnel plot, P \.0001, I 2 (inconsistency) = 93.11% (95% confidence interval: 88.28%-95.95%). Figure 4. Forest plot, pooled specificity = 93.5% (95% confidence interval: 88.2%-97.3%). Figure 3. Forest plot, pooled sensitivity = 34.3% (95% confidence interval: 19.8%-50.6%). Figure 5. Forest plot, pooled positive predictive value = 78.0% (95% confidence interval: 65.0%-88.7%). After pooling the extractable data from 7 studies, we found that RAS mutations carried an overall sensitivity of 34.3% and a specificity of 93.5% for detecting malignancy among indeterminate FNA samples. Subsequently, the positive likelihood ratio and negative likelihood ratio were found to be and 0.775, respectively. Therefore, while a positive result would moderately increase the risk of malignancy, a negative result would only slightly reduce that risk. The low sensitivity is likely due to the fact that RAS mutations are associated predominantly with the subset of thyroid neoplasms exhibiting follicular growth patterns (follicular adenoma, follicular thyroid carcinoma, follicular variant of papillary thyroid carcinoma, etc). 45 While this subset represents a significant portion of thyroid malignancies, the remaining subtypes likely account for the relatively large number of malignancies not harboring a RAS mutation. Our data affirm the results of a previous study not limited to indeterminate nodules, which found that RAS had the lowest sensitivity of 36 markers. 46 Interestingly, our data indicate that RAS mutations are highly specific for malignancy despite mutations being present in 20% to 40% of thyroid adenomas. 13 This apparent discrepancy could be explained by assuming that benign lesions carrying a RAS mutation are more likely to have benign cytology and are therefore excluded from our data. Furthermore, RAS-positive carcinomas are more likely to have indeterminate cytology and lack suspicious ultrasound features. 13 Thus, in indeterminate nodules, the specificity and positive likelihood ratio suggest that the presence of a RAS mutation is strongly predictive of malignancy. The ideal test would have a positive predictive value and negative predictive value comparable to that of malignant (class VI) and benign (class 2) cytology (98.6% and 96.3%, respectively). 47 Because our data suggest that RAS mutations are specific but not sensitive for carcinoma and carry only a positive predictive value of 78.0% and a negative predictive value of 64.0%, we deduce that testing cytologically indeterminate nodules for RAS mutations in isolation

5 476 Otolaryngology Head and Neck Surgery 156(3) Figure 6. Forest plot, pooled negative predictive value = 64.0% (95% confidence interval: 50.0%-76.9%). is unlikely to change management and is therefore unnecessary. However, we acknowledge that RAS mutations represent 1 subset out of many molecular markers currently being researched and must be considered in that context. A growing body of literature is assessing the value of casting a broader net by simultaneously testing multiple markers with the aim of increasing sensitivity. Nikiforov et al found that a panel including point mutations in BRAF and RAS, as well as gene rearrangements such as RET/PTC1 or RET/PTC3 and PAX8/PPARg, provided test sensitivity and specificity of 61% and 98%, respectively. 19 While the specificity of this panel is similar to ours, the sensitivity is nearly twice that of detecting RAS mutations alone. Furthermore, Yip et al demonstrated that utilizing molecular testing of 9 markers to guide surgical management increased the likelihood that a total thyroidectomy was performed for carcinomas, while a lobectomy was more likely performed for what turned out to be benign neoplasms. 48 This finding implies that the panel was helpful in limiting the extent of surgery. Conversely, Aragon Han et al retrospectively found that whole-panel testing did not alter surgical management in a group of 87 patients. In the 9 cases for which management was altered, 6 were deemed inappropriate. 49 The apparent contradiction between these studies may be due to differing study designs, thus highlighting the need for further investigation. By taking differing molecular approaches, current commercially available panels may be the solution to the problem of simultaneously maximizing sensitivity and specificity to avoid or limit surgery. 50 One approach tests for point mutations and gene rearrangements directly, as previously discussed. Examples are Asurgen s mirinform and ThyroSeq 2 from CBLPath. Similar to testing for RAS in isolation, these have been shown to have relatively high specificities and the subsequent ability to rule in malignancy. The mirinform panel was at one point offered for a retail price of $2250; however, many insurance companies will not pay for it despite evidence suggesting its cost efficacy. 51 Another methodology, used by the Veracyte Afirma test, assesses the expression of mrna from 167 genes. This approach has been shown to have a higher sensitivity (92% in 1 study 7 ) and the ability to rule out malignancy. A projected cost-effectiveness model (assuming a retail price of $3200) found the Afirma test to save $1453 on total cost of care. 52 Thus, it would seem promising to use such tests in conjunction, for diagnostic and economic reasons, but more prospective research is needed to elucidate the actual effect on cost efficiency and patient outcomes now that these tests are increasingly available. Regardless, some research implies that RAS mutations do not contribute to the utility of these panels. One study found that simultaneously testing for RAS and BRAF mutations did not significantly increase the diagnostic sensitivity of cytology toward thyroid cancer as compared with testing for BRAF mutations alone. However, that study included only indeterminate FNA samples from nodules exhibiting clinical or ultrasonographic findings concerning for malignancy. 53 This could present selection bias due to the likelihood of RAS-positive nodules lacking suspicious ultrasound features. Another study found that adding RAS mutations to a panel decreased the specificity from 100% to 98%. 54 The slight diminishment likely results from follicular adenomas harboring RAS mutations. This study, however, did not directly test for markers but instead applied epidemiologically determined mutation rates to a cohort, to project what the diagnostic parameters would have been. Based on these studies, it could be recommended, albeit weakly, to forgo inclusion of RAS testing in such panels altogether. But given the relative limitations of these studies and the robust positive likelihood ratio and specificity determined in the current study, inclusion would be more likely to help than harm diagnostic validity. Aside from simply diagnosing malignancy, RAS mutations may serve as prognostic indicators. They have been shown to be highly predictive of follicular neoplasms, which tend to be indolent with slow growth and are associated with a relatively good prognosis, even when diagnosed as carcinoma. 13 This contrasts with the finding that BRAF mutations confer a higher risk of disease persistence or recurrence. 55 A difference in outcomes has also been shown in the comparison of RAS mutation subtypes. In 1 study, the probability of a malignant outcome for the mutation subtypes followed the order of KRAS (100%), NRAS (74%), and HRAS (56%). However, this study was restricted to follicular adenoma and carcinoma 44 Radkay et al found that NRAS61 and HRAS61 mutations had a higher positive predictive value and were more closely associated with follicular-variant papillary thyroid carcinoma than KRAS12/ 13 mutations. 56 Thus, in the future, patterns of mutational status could guide the extent of surgical management. There are, however, limitations to this study. First, although our search strategy was designed to maximize inclusivity, it is possible that some existing data meeting inclusion criteria were missed. Second, the studies included were somewhat heterogeneous, potentially affecting the

6 Clinkscales et al 477 results. Sample sizes ranged from 19 to 513, while RAS mutation incidence ranged from 24% to 79%. Furthermore, our methods do not account for the variability in pathologist expertise across institutions. While the Bethesda classification system offers a great deal of standardization, realistically an inconsistent accuracy of FNA readings presents a potential source of confounding error. Last, the included studies were not limited to any particular mutational detection method. All but 1 of the papers used some combination of polymerase chain reaction for genomic amplification and either pyrosequencing or fluorescence melting curve analysis for point mutation detection. One paper used a rapid multiplex molecular analysis. These differing methods could have also had a confounding effect. Conclusion The present study suggests with some certainty that detecting RAS mutations alone should not alter the decision to undergo surgery. Although the literature seems to question the contribution of RAS mutations, as part of a panel, it may guide the extent of surgical management. Further investigation could identify some confounding variable such as age, mutation subtype, or ethnicity that, once stratified, could render RAS mutations more useful. Until then, our results contribute to the current understanding that, in general, molecular markers are not sensitive enough to definitively alter management, and the clinical role of mutation analysis continues to need defining. Author Contributions William Clinkscales, designed study, gathered data, authored manuscript; Adrian Ong, designed study, gathered data, authored manuscript; Shaun Nguyen, analyzed data, revised paper; Elizabeth Emily Harruff, analyzed data, revised paper; Marion Boyd Gillespie, designed study, revised manuscript, senior author. Disclosures Competing interests: Shaun Nguyen, CSL Behring, Roche consultant. Sponsorships: None. Funding source: None. References 1. Russ G, Leboulleux S, Leenhardt L, Hgedus L. Thyroid incidentalomas: epidemiology, risk stratification with ultrasound, and workup. Eur Thryoid J. 2014;3: Ohori NP, Schoedel KE. Variability in the atypia of undetermined significance/follicular lesion of undetermined significance diagnosis in the Bethesda System for Reporting Thyroid Cytopathology: sources and recommendations. Acta Cytol. 2011;55: Cibas ES, Ali SZ. The Bethesda System for Reporting Thyroid Cytopathology. Am J Clin Pathol. 2009;132: Goldstein RE, Netterville JI, Burkey B, Johnson JE. Implications of follicular neoplasms, atypia, and lesions suspicious for malignancy diagnosed by fine-needle aspiration of thyroid nodules. Ann Surg. 2002;235: Kato MA, Buitrago D, Moo TA, et al. Predictive value of cytologic atypia in indeterminate thyroid fine-needle aspirate biopsies. Ann Surg Oncol. 2011;18: Ghanem M, Levy Y, Mazeh H. Preoperative diagnoses of benign thyroid nodules with inderterminate cytology. Gland Surger. 2012;1: Alexander EK, Kennedy GC, Baloch ZW, et al. Preoperative diagnosis of benign thyroid nodule with indeterminate cytology. N Eng J Med. 2012;367: Kwak JY, Kim EK. Cancer: Indeterminate thyroid nodules added testing, added value? Nat Rev Endocrinol. 2013;6: Haugen BR, Alexander EK, Bible KC, et al American Thyroid Association management guidelines for adult patients with thyroid nodules and differentiated thyroid cancer: the American Thyroid Association Guidelines Task Force on Thyroid Nodules and Differentiated Thyroid Cancer. Thyroid. 2015;26: Bryson PC, Shores CG, Hart C, et al. Immunohistochemical distinction of follicular thyroid adenomas and follicular carcinomas. Arch Otolaryngol Head Neck Surg. 2008;134: Moses W, Weng J, Sansano I, et al. Molecular testing for somatic mutations improves the accuracy of thyroid fineneedle aspiration biopsy. World J Surg. 2010;34: Howell GM, Hodak SP, Yip L. RAS mutations in thyroid cancer. Oncologist. 2013;18: Gupta N, Dasyam AK, Carty SE, et al. RAS mutations in thyroid FNA specimens are highly predictive of predominantly low-risk follicular-pattern cancers. J Clin Endocrinol Metab. 2013;98:E914-E Jung CK, Little MP, Lubin JH, et al. The increase in thyroid cancer incidence during the last four decades is accompanied by a high frequency of BRAF mutations and a sharp increase in RAS mutations. J Clin Endocrinol Metab. 2014;99:E276- E Jia Y, Yu Y, Li X, et al. Diagnostic value of B-RAF(V600E) in difficult-to-diagnose thyroid nodules using fine-needle aspiration: systematic review and meta-analysis. Diagn Cytopathol. 2014;42: Sterne JAC, Egger E. Funnel plots for detecting bias in metaanalysis: guidelines on choice of axis. J Clin Epidemiol. 2001; 54: DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7: Freeman MF, Tukey JW. Transformations related to the angular and the square root. Annals of Mathematical Statistics. 1950;21: 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. J Clin Endocrinol Metab. 2011;96: Nikiforov YE, Steward DL, Robinson-Smith TM, et al. Molecular testing for mutations in improving the fine-needle aspiration diagnosis of thyroid nodules. J Clin Endocrinol Metab. 2009;94: Beaudenon-Huibregtse S, Alexander EK, Guttler RB, et al. Centralized molecular testing for oncogenic gene mutations

7 478 Otolaryngology Head and Neck Surgery 156(3) complements the local cytopathologic diagnosis of thyroid nodules. Thyroid. 2014;24: Cantara S, Capezzone M, Marchisotta S, et al. Impact of proto-oncogene mutation detection in cytological specimens from thyroid nodules improves the diagnostic accuracy of cytology. J Clin Endocrin Met. 2010;95: Eszlinger M, Piana S, Moll A, et al. Molecular testing of thyroid fine-needle aspirations improves presurgical diagnosis and supports the histologic identification of minimally invasive follicular thyroid carcinomas. Thyroid. 2015;25: Stence AA, Gailey MP, Robinson RA, Jensen CS, Ma D. Simultaneously detection of 50 mutations at 20 sites in the BRAF and RAS genes by multiplexed single-nucleotide primer extension assay using fine-needle aspirates of thyroid nodules. Yale J Biol Med. 2015;88: An JH, Song KH, Kim SK, et al. RAS mutations in indeterminate thyroid nodules are predictive of the follicular variant of papillary thyroid carcinoma. Clin Endocrin. 2015;82: Yoon JH, Kwon HJ, Lee HS, Kim EK, Moon HJ, Kwak JY. RAS mutations in AUS/FLUS cytology: does it have an additional role in BRAF V600E mutation-negative nodules? Medicine (Baltimore). 2015;94:e Ohori NP, Nikifovora MN, Schoedel KE, et al. Contribution of molecular testing to thyroid fine-needle aspiration cytology of follicular lesion of undetermined significance/atypia of undetermined significance. Cancer Cytopathol. 2012:118: 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-generation sequencing assay. Cancer. 2014;120: Ohori NP, Wolfe J, Hodak SP, et al. Colloid-rich follicular neoplasm/suspicious for follicular neoplasm thyroid fineneedle aspiration specimens: cytologic, histologic, and molecular basis for considering an alternate view. Cancer Cytopathol. 2013;121: Duman BB, Kara OI, Uguz A, Ates BT. Evaluation of PTEN, PI3K, MTOR, and KRAS expression and their clinical and prognostic relevance to differentiated thyroid carcinoma. Contemp Oncol. 2014;18: Eszlinger M, Krogdahl A, Münz S, et al. Impact of molecular screening for point mutations and rearrangements in routine air-dried fine-needle aspiration samples of thyroid nodules. Thyroid. 2014;24: Guerra A, Carrano M, Angrisani E, et al detection of RAS mutation by pyrosequencing in thyroid cytology samples. Int J Surg. 2014;12: Kozio1ek M, Bińczak-Kuleta A, Stepaniuk M, et al. Frequency assessment of BRAF mutation, KRas mutation, and RASSF1A methylation in nodular goitre based on fine-needle aspiration cytology specimens. Endokryn Polska. 2015;66: Liu S, Gao A, Zhang B, et al. Assessment of molecular testing in fine-needle aspiration biopsy samples: an experience in a Chinese population. Exp Mol Path. 2014;97: Mathur A, Weng J, Moses W, et al. A prospective study evaluating the accuracy of using combined clinical factors and candidate diagnostic markers to refine the accuracy of thyroid fine needle aspiration biopsy. Surgery. 2010;148: Park SJ, Sun JYH, Hong K, et al. Application of BRAF, NRAS, KRAS mutations as markers for the detection of papillary thyroid cancer from FNAB specimens by pyrosequencing analysis. Clin Chem Lab Med. 2013;51: Park JY, Kim WY, Hwang TS, et al. BRAF and RAS mutations in follicular variants of papillary thyroid carcinoma. End Path. 2013;24: Guo HQ, Zhao H, Zhang ZH, Zhu YL, Xiao T, Pan QJ. Impact of molecular testing in the diagnosis of thyroid fine needle aspiration cytology: data from mainland China. Dis Markers. 2014;2014: Rivera M, Ricarte-Filho J, Knauf J, et al. Molecular genotyping of papillary thyroid carcinoma follicular variant according to its histological subtypes (encapsulated vs infiltrative) reveals distinct BRAF and RAS mutation patterns. Mod Pathol. 2010;23: Proietti A, Sartori C, Borrelli N, et al. Follicular-derived neoplasms: morphometric and genetic differences. J Endocrinol Invest. 2013;36: Proietti A, Borrelli N, Giannini R, et al. Molecular characterization of 54 cases of false-negative fine-needle aspiration among 1347 papillary thyroid carcinomas. Cancer Cytopath. 2014;122: Lee SR, Jung CK, Kim TE, et al. Molecular genotyping of follicular variant of papillary thyroid carcinoma correlates with diagnostic category of fine-needle aspiration cytology: values of RAS mutation testing. Thyroid. 2013;23: Hwang TS, Kim WY, Han HS, et al. Preoperative RAS mutational analysis is of great value in predicting follicular variant of papillary thyroid carcinoma. Biomed Res Int. 2015;2015: Fukahori M, Yoshida A, Hayashi H, et al. The associations between RAS mutations and clinical characteristics in follicular thyroid tumors: new insights from a single center and a large patient cohort. Thyroid. 2012;22: Nikiforov YE, Nikiforova MN. Molecular genetics and diagnosis of thyroid cancer. Nat Rev Endocrinol. 2011;7: Rodrigues HGC, de Pontes AAN, Adan LFF. Use of molecular markers in samples obtained from preoperative aspiration of thyroid. Endocrine J. 2012;59: Bongiovanni M, Spitale A, Faquin WC, Mazzucchelli L, Baloch ZW. The Bethesda System for Reporting Thyroid Cytopathology: a meta-analysis. Acta Cytol. 2012;56: Yip L, Wharry LI, Armstrong MJ, et al. A clinical algorithm for fine-needle aspiration molecular testing effectively guides the appropriate extent of initial thyroidectomy. Ann Surg. 2014;260: Aragon Han P, Olson MT, Fazeli R, et al. The impact of molecular testing on the surgical management of patients with thyroid nodules. Ann Surg Oncol. 2014;21: Hodak SP, Rosenthal DS. Information for clinicians: commercially available molecular diagnosis testing in the evaluation of thyroid nodule fine-needle aspiration specimens. Thyroid. 2013;23:

8 Clinkscales et al Yip L, Farris C, Kabaker AS, et al. Cost impact of molecular testing for indeterminate thyroid nodule fine-needle aspiration biopsies. J Clin Endocrinol Metab. 2012;97: Li H, Robinson KA, Anton B, Saldanha IJ, Ladenson PW. Cost-effectiveness of a novel molecular test for cytologically indeterminate thyroid nodules. J Clin Endocrinol Metab. 2011; 96:E1719-E Rossi M, Buratto M, Tagliati F, et al. Relevance of BRAFV600E mutation testing versus RAS point mutations and RET/PTC rearrangements evaluation in the diagnosis of thyroid cancer. Thyroid. 2015;25: Filicori F, Keutgen XM, Buitrago D, et al. Risk stratification of indeterminate thyroid fine-needle aspiration biopsy specimens based on mutation analysis. Surgery. 2011;150: Xing M, Alzahrani A, Carson K, et al. Association between BRAF V600E mutation and mortality in patients with papillary thyroid cancer. JAMA. 2013;309: Radkay LA, Chiosea SI, Seethala RR, et al. Thyroid nodules with KRAS mutations are different from nodules with NRAS and HRAS mutations with regard to cytopathologic and histopathologic outcome characteristics. Cancer Cyto. 2014;122: