Molecular Testing for BRAFV600E and RAS Mutations from Cytoscrapes of Thyroid Fine Needle Aspirates: A Single-Center Pilot Study.

CONTEXT AND AIM: Molecular testing of thyroid FNA has been advocated in the indeterminate categories of The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) 2018. The utility of cytoscrapes of thyroid FNA samples for BRAF V600E and RAS mutations was evaluated in this pilot study. METHODS AND MATERIALS: Thyroid FNA samples between 2015 and 2018 from TBSRTC categories 3-6 were included. DNA was extracted from one to two representative smears (cytoscrape). Real-time PCR for BRAF V600E and RAS (KRAS, NRAS, and HRAS) gene mutations was performed. Histopathology correlation was available in 44 cases. STATISTICAL
METHODS: Chi-square test and calculation of sensitivity, specificity, and positive/negative predictive values were performed.
RESULTS: A total of 73 thyroid FNA cases and 11 nodal metastases of papillary thyroid carcinoma (PTC) were evaluated. The DNA yield ranged from 1.9 to 666 ng/μl (mean 128 ng/μl) in 80 cases and was insufficient in four cases. Overall, mutations were seen in 45 (56.25%) cases with BRAF V600E, NRAS, HRAS, and KRAS in 21 (46.7%), 19 (42.2%), 4, and 1 cases, respectively. BRAF V600E mutation was seen in PTC (11/18, 61%), nodal PTC metastases (5/10, 50%), and occasionally in TBSRTC category 3 (1/18, 5.5%). NRAS mutations were seen across all categories and were maximum in the AUS/FLUS group (6/18, 33%). BRAF V600E /RAS testing had an overall sensitivity, specificity, and positive and negative predictive values of 61.7%, 80%, 91.3%, and 38%, respectively, for the detection of malignancy. In indeterminate thyroid nodules, the sensitivity, specificity, PPV, and NPV were 56.2%, 80%, 81.8%, and 53.3%, respectively.
CONCLUSION: BRAF V600E/RAS mutation testing from cytoscrapes are useful as a rule-in test for indeterminate thyroid nodules and provide molecular confirmation in nodal metastases of PTC. Copyright:

INTRODUCTION

Fine needle aspiration cytology (FNAC) is a reliable technique for diagnosing malignant thyroid nodules with papillary thyroid carcinoma (PTC), the most common thyroid cancer in our setup.[1] The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) is the commonly used classification system for reporting thyroid FNAC and categories 3, 4, and 5 of TBSRTC are regarded as grey-zone areas with intermediate malignancy risk.[123] The recent TBSRTC system has advocated the application of molecular studies to identify the genetic mutations/gene rearrangements in category 3 or AUS (atypia of undetermined significance) in line with the guidelines of the American Thyroid Association.[4] This would potentially identify the patients for surgery which includes near-total thyroidectomy or lobectomy. The most common alteration involves the B-type RAF kinase (BRAF) and the RAS genes followed by rearrangements of RET/PTC and PAX8/PPARγgenes. BRAF V600E mutation is present in 45–80% of classical PTC, 5–25% of follicular variant of PTC or FVPTC, 1.4% of follicular thyroid carcinoma, 5–15% of poorly differentiated thyroid carcinoma or PDC, and 10–50% of anaplastic thyroid carcinoma.[56] The rat sarcoma (RAS) oncogene family includes three genes HRAS, NRAS, and KRAS. The prevalence of RAS mutations in thyroid cancer is 20–40% and is found in follicular lesions including adenomas, carcinomas, and FVPTCs.[67] The PAX8/PPARγ gene fusionis present in thyroid follicular lesions, both in adenoma and in carcinoma, while the RET/PTC gene rearrangement is seen in 5–35% of PTC.[56]

We aimed to evaluate the utility of detection of BRAF V600E and RAS (HRAS, KRAS, and NRAS) gene mutations by real-time polymerase chain reaction across the TBSRTC spectrum and with special focus on its value for the indeterminate thyroid nodules including categories 3, 4 and 5 of TBSRTC. We also included cases of nodal metastases of PTC in this analysis. We exclusively used DNA obtained by cytoscrapes of representative smears of the thyroid lesions to confirm its feasibility.

MATERIALS AND METHODS

This was a retrospective analysis of cases referred to the Department of Cytology & Gynaec Pathology, PGIMER, Chandigarh, for FNAC of thyroid swelling between July 2015 and June18. The approval for the study was obtained from the Institute Ethics Committee vide letter no. INT/IEC/2019/001424 dated 18.07.19 with a waiver for obtaining individual consent. A total of 84 cases in category 3-6 of TBSRTC were included in the study. The number of cases was restricted to 84 based on the resources available. The inclusion criteria was a TBSRTC category 3–6 with adequate material available on the smears. All unsatisfactory and benign thyroid aspirates (category 2) were excluded for molecular testing. In all these cases, FNAC was routinely done using a 23G needle attached to a 20-ml syringe in the specially designed holder (Cameco, AB Taby) and May Grünwald–Giemsa (MGG) stained air-dried smears and Haematoxylin & Eosin stained alcohol fixed smears were routinely evaluated. The histopathological (HPE) follow-up of these cases wherever available was also recorded and compared with cytology findings.

Extraction of DNA: Genomic DNA was extracted generally from air-dried, May-Grünwald–Giemsa stained smears of the cases included. The slides chosen were moderately cellular containing at least 10 clusters of cells with a minimum of 20 cells. If the cellularity was low, then one more smear was used. The entire slide was dipped in water for 2–3 seconds and scraped manually using a sterile scalpel blade and the material was transferred to an Eppendorf tube. This was referred to as “cytoscrape.” Genomic DNA was extracted from the material in the Eppendorf tube using a commercially available DNA extraction kit (Qiagen DNeasy Blood and Tissue kit, GmBH, Hilden, Germany) following the manufacturer's protocol. The DNA yield was quantified with a spectrophotometer by measuring the absorbance at 260 nm. The purity of DNA was determined by the ratio of A260/280 and a ratio of 1.8 to 2.0 was taken as an acceptable value for DNA. Two microliters of DNA was also run on a 2% agarose gel to check for its quality.

Real-Time PCR: Real-time PCR (Agilent technologies—AriaMx Real-time PCR system, Agilent Technologies, Santa Clara, USA) was performed using an EntroGen thyroid mutation analysis kit (THDNA-RT64, Entrogen Inc, CA, USA). Each reaction well contains primer sets and probes for the detection of somatic mutations as well as an internal control gene. The assay works by amplifying the mutant-specific sequences in the samples that contain a mixture of mutant and wild-type DNA. The probes used were labeled with fluorochromes, Fluorescein amidite (FAM) and VIC (2′-chloro-7′phenyl-1,4-dichloro-6-carboxyfluorescein). The method was standardized using 20 ng of DNA for each gene mutation. Each reaction contained 15 μl of the reaction mix, 1.5 μl of 20 ng of DNA, 6 μl of primer, and 7.5 μl of water in a final volume of 30 μl. The RT-PCR was carried out using the following run conditions: initial denaturation at 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 secs, and 60°C for 60 secs. The results were analyzed using the AGILENT AriaMx 1.0 software and called out as positive or negative for mutation based on the cycle threshold (Ct) values for the two probes used as per the manufacturer's instructions. This assay uses two probes, VIC and FAM which detect the internal control and test mutation, respectively. A cycle threshold (Ct) values ≤38 for FAM and ≥25 for VIC indicate positivity for mutation (internal control and test positive); values of <38 for FAM and <25 for VIC indicated excess DNA; values >38 for FAM and <31 for VIC indicate negativity for the mutation, and values >38 for FAM and >31 for VIC indicated insufficient DNA.

Statistical analysis: The statistical analysis was carried out using the IBM Statistical Package for Social Sciences software version 22. Qualitative data or categorical variables were described as frequencies and proportions. Proportions were compared using a Chi-square test. A P value

RESULTS

A total of 84 cases were included in the study. There were 64 females and 20 males with a mean age of 42.5 (2–70) years and a median age of 43 years. These included 73 primary thyroid lesions and 11 cases of lymph nodal metastases from PTC. The break-up of cases as per TBSRTC categories along with the histopathology follow-up and mutations observed are detailed in Table 1. The cohort included 28 thyroid malignant neoplasms (including 19 conventional PTC and one follicular variant of PTC), 11 nodal metastases of PTC, and the remaining 45 cases were indeterminate thyroid nodules with 20, 9, and 16 cases of TBSRTC categories 3, 4, and 5, respectively. The histopathology outcomes in 44 of these cases is also shown [Table 1].

Table 1

TBSRTC categories, mutational analysis, and histopathology outcomes in thyroid aspirates (n = 84)

TBSRTC Category DNA range/mean	Mutations	No. of cases with HPE	HPE outcome	No. of cases with no HPE follow-up
n = 84		n = 44		n = 40
Cat 3 (n = 20) 5.0-183.2 ng/µl	None	6	MNG-2, FA-3, HCA-1	3
Mean: 62.1 ng/µl	NC			2
	BRAF			1
	NRAS	3	FVPTC-1, PTC-1, FA-1	3
	HRAS	1	FA-1
	KRAS			1
Cat 4 (n = 9)18-666 ng/µl	None	3	FVPTC-1, FC-1, PDC-1
Mean: 132.7 ng/µl	NRAS	2	FC-2	1
	HRAS	2	FC-2	1
Cat 5 (n = 16)1.9-111.9 ng/µl	None	6	PTC-3, FVPTC-1, FA-2	3
Mean: 36.2 ng/µl	BRAF	1	PTC-1	2
	NRAS	2	FC-2	2
Cat 6, PTC = 19	None	2	PTC-2	4
15.5-433 ng/µl	BRAF	6	PTC-6	5
Mean: 113.9 ng/µl	NC			1
	NRAS			1
FVPTC = 1	BRAF	1	FVPTC-1
HCC = 2	None	1	HCC-1	1
PDC = 2	NRAS	1	FC-1	1
AC = 4	None			2
	NRAS			2
Nodal (n = 11)	None	3	PTC-2, FVPTC-1	1
Metastases	NC			1
12.1-46 ng/µl	BRAF	4	PTC-4	1
48.2 ng/µl	NRAS			1

TBSRTC-The Bethesda System of Reporting Thyroid Cytology; Cat-category; NC-Noncontributory; HPE-Histopathology examination; MNG-Multinodular goiter; FA- Follicular adenoma; FC- Follicular carcinoma; PTC- Papillary thyroid carcinoma; FVPTC- Follicular variant of PTC; PDC- Poorly differentiated carcinoma; - HCA- Hürthle cell adenoma; HCC- Hürthle cell carcinoma; AC-anaplastic carcinoma; BRAF-BRAFV600E mutation

DNA extraction from cytoscrapes and its quantitation in various categories

Overall, sufficient DNA was available for molecular analysis in 80 cases, and in the remaining four cases, it was insufficient and hence were excluded from the study. The quantity of DNA extracted from the cytoscrapes ranged from 1.9 to 666 ng/μl, with a mean value of 128 ng/μl. The range and mean DNA extracted for each of the TBSRTC categories is shown in Table 1. Category 5 cases showed the lowest mean DNA extracted followed by the nodal metastases group and categories 3, 6, and 4.

Mutational analysis for BRAF and RAS genes

Molecular testing could be performed in 80 cases, and 45 cases (54.9%) showed positivity for BRAF/RAS gene mutations whereas the remaining 35 cases were negative [Table 2]. The most common mutation was BRAF V600E mutation which is seen in 21 cases (46.7%) followed by NRAS mutation in 19 cases (42%), HRAS mutation in four cases (8.9%), and KRAS mutation in only one case (2%). The cases which showed BRAF V600E mutation were mostly in cases confirmed as PTC and in nodal metastases of PTC and a single case in category 5, suspicious of malignancy. It was also detected in one case in the AUS/FLUS group which did not have a histopathology follow-up. A representative case of PTC positive for BRAF mutation is illustrated in Figure 1a–c. On the other hand, RAS gene mutations were spread across categories 3–6. NRAS gene mutations were seen in 19 cases and across all categories with six, three, four, and five cases in categories 3, 4, 5, and 6, respectively, and in one case of nodal metastases [Tables 1 and 2]. In eight cases that were positivefor NRAS mutation, follow-up histopathology revealed malignancy in seven cases with five follicular carcinomas, one FVPTC, and one conventional PTC. Hence, detection of NRAS gene mutation had a high (87.5%) predictive value for malignancy.

Table 2

BRAF and RAS mutations in different TBSRTC categories [N = 80]

TBSRTC	No. of	No mutation	Mutation	BRAF	NRAS	HRAS	KRAS
Category*	Cases	detected	detected
Cat 3	18	9	9 (50%)	1	6	1	1
Cat 4	9	3	6 (66.7%)	0	3	3	0
Cat 5	16	9	7 (43.75%)	3	4	0	0
Cat 6	27	10	17 (66.7%)	12	5	0	0
Nodal metastases	10	4	6 (55.6%)	5	1	0	0
Total	80	35	45 (56.25%)	21	19	4	1
				(46.7%)	(42.2%)	(8.9%)	(2.2%)

*cases with technically noncontributory results excluded

Figure 1

Each row represents one illustrative case. a–c, Papillary thyroid carcinoma on cytology (category 6) with BRAF V600E mutation and PTC confirmed on histopathology; d–f, Follicular lesion of undetermined significance (category 3) with NRAS mutation and follicular adenoma on histopathology; g–i, follicular neoplasm on cytology (category 4) with mutation and follicular carcinoma on histopathology. [a, g x100; g x200 - May-Grünwald Giemsa stain; c x400, f and i x200 – Hematoxyline-Eosin stained sections; b, e, and h—screenshots of plots obtained]

In 35 cases, no mutations were detected [Table 2]. Four cases were nodal PTC metastases. Histopathological follow-up was benign in eight cases (two nodular goiters, one Hürthle cell adenoma, and five follicular adenomas) and malignant in 13 cases (five conventional PTC, two nodal PTC metastases, three FVPTC, one follicular carcinoma, one Hürthle cell carcinoma, and one poorly differentiated carcinoma), thereby indicating the low predictive value of a negative test.

TBSRTC category 3, 4, and 5 or indeterminate thyroid nodules

Molecular testing could be performed in 18 of 20 cases of AUS/FLUS from cytoscrapes of representative smears and 9/18 (50%) tested positive. Fifteen cases showed architectural atypia (AA), two cases showed cytological atypia (CyA), and one showed both (CyA/AA). The most common mutation was in the NRAS gene (six cases, 33.3%), followed by one case each positive for BRAF V600E(5.6%), HRAS (5.6%), and KRAS (5.6%)mutations. Of the 15 cases subcategorized as AA, no mutations were detected in seven cases, NRAS mutations were seen in four cases, and HRAS and BRAF V600E mutations were seen in one case each, respectively; in two cases, molecular testing was technically compromised. Among two cases with CyA, one showed NRAS mutation while others showed no mutations. The CyA/AA case showed NRAS mutation. Follow-up histopathology was available in ten cases with one PTC (BRAF V600E mutation positive), one FVPTC (NRAS mutation positive), five follicular adenomas (one HRAS, one NRAS and three no mutations), one case each of Hurthle cell adenoma, colloid goiter, and follicular hyperplasia that were negative for any mutation. A representative case of FLUS (follicular lesion of undetermined significance) with NRAS mutation whose histopathology revealed follicular adenoma is shown in Figure 1(d–f).

There were nine cases in category 4, of which three showed HRAS and NRAS mutation each and the rest three were negative. Of the three cases with no detectable mutations, follow-up histopathology revealed FVPTC, follicular carcinoma, and poorly differentiated carcinoma in one case each, respectively. Follicular carcinoma was seen on histopathology in two out of three cases that showed NRAS and HRAS mutations, respectively, and HPE was not available in the remaining cases. A representative case of follicular neoplasm on cytology with HRAS mutation whose histopathology revealed follicular carcinoma is shown in Figure 1g–i.

Among 16 cases in category 5, suspicious for malignancy, seven mutations (43.8%) were seen with four NRAS (25%) and three (18.8%) BRAF V600E mutations. The histological outcome was conventional PTC in four cases (one BRAF V600E and three no mutations), FVPTC in one case (no mutation), follicular carcinoma in two cases with NRAS mutation, and follicular adenoma in two cases with no mutations detected. Two representative cases in category 5, suspicious for PTC, one with no mutations detected and FVPTC on histopathology and the second with BRAFV600E and PTC on histopathology are depicted in Figure 2a–c and d–f.

Figure 2

Each row represents one illustrative case. a–c, Suspicious for PTC on cytology (category 5) with no mutations detected and FVPTC on histopathology; d–f, Suspicious for PTC on cytology (category 5), positive for BRAF V600E mutation, and PTC on histopathology; g–i, Metastatic PTC in lymph node on cytology, positive for BRAF V600E mutation, and confirmed on histopathology [a, g x400, d x200- May-Grünwald–Giemsa stain;

Mutational analysis of nodal metastases (N = 11)

Mutational analysis was performed from 10 out of 11 lymph nodal metastases of PTC as in one case the results were technically noncontributory. Overall, 5/10 (50%) showed BRAF V600E mutation; NRAS mutation was present in one case and none detected in four cases [Table 1]. In all these cases which initially presented as nodal metastases and diagnosed by FNA, the thyroid primary lesion was detected subsequently by imaging and ultrasound-guided FNA confirmed PTC of the thyroid. HPE correlation was available for seven cases, which showed PTC with lymph nodal metastases, six conventional and one FVPTC. A representative case positive for BRAF V600E mutation is illustrated in Figure 2g–i.

Histopathological outcome and correlation to gene mutational analysis

The gene mutations in histologically proven cases (N = 44) are shown in Table 3. It was observed that BRAF V600E mutation was restricted to PTC including FVPTC; however, NRAS mutation was seen in follicular neoplasms as well as few cases of PTC. All cases reported as follicular adenoma and Hürthle cell adenoma were assumed to have a benign outcome for calculation of the sensitivity and specificity of molecular testing for BRAF and RAS gene mutations. The results of the correlation of gene mutation detection with the histological outcomes are tabulated in Table 4A and 4B. There was a positive correlation of gene mutation detection with the overall malignant histological outcome (Chi-square, 5.402; P = 0.02). However, in the indeterminate category, this association was not seen (p = 0.06). The sensitivity of molecular testing for BRAF and RAS genes was 61.7%, specificity was 80%, positive predictive value (PPV) was 91.3%, and negative predictive value (NPV) was 38%. In the indeterminate category, the sensitivity of molecular testing for BRAF and RAS genes was 56.2% with a specificity of 80%, PPV of 81.8%, and NPV of 53.3%.

Table 3

Correlation of genetic mutation with histopathological outcomes [N = 44]

Histopathology	BRAF	NRAS	HRAS	None
	V600E			detected
PTC (13)	7 (53.8%)	1 (7.6%)	0 (0%)	5 (38.4%)
Nodal PTC (6) metastases	4 (66.7%)	0	0	2 (33.3%)
FVPTC (5)	1 (20%)	1 (20%)	0	3 (60%)
Follicular (8) Carcinoma	0	5 (62.5%)	2 (25%)	1 (12.5%)
PD carcinoma (1)	0	0	0	1 (100%)
Hurthle cell (1) Carcinoma	0	0	0	1 (100%)
Follicular (7) Adenoma	0	1 (14.3%)	1 (14.3%)	5 (71.4%)
Hurthle cell (1) Adenoma	0	0	0	1 (100%)
Multinodular (2) Goiter	0	0	0	2 (100%)
Total (44)	12	8	3	21

PTC- Papillary thyroid carcinoma; FVPTC- Follicular variant of PTC; PD-poorly differentiated

Table 4

Correlation of gene mutations with histological outcomes

A: Overall histological outcome [N = 44]
Histological	Mutation	Mutation	Total
Outcome	Positive	Negative
Benign#	2	8	10
Malignant	21	13	34
Total	23	21	44
chi-square = 5.402; p-value =0.0201
B: Histological outcome in indeterminate (category 3, 4, and 5) lesions [N = 26]
Histological	Mutation	Mutation	Total
Outcome	Positive	Negative
Benign#	2	8	10
Malignant	9	7	16
Total	11	15	26

Chi-square=3.313; p-value = 0.068. #includes Follicular and Hürthle cell adenoma

DISCUSSION

The current version of TBSRTC recommends molecular testing for category 3 (atypia of undetermined significance) in line with the guidelines of the American Thyroid Association.[2] In the present study, we evaluated a commercially available simple 5-gene mutation kit in TBSRTC categories 3–6 and nodal metastases of PTC. The material for molecular testing may be obtained as a direct sample at the time of performing the FNA. However, this is operator dependent and may not be taken at the time of FNA due to the paucity of the sample. The cytoscrape technique was therefore employed in this retrospective analysis, wherein DNA was extracted from one representative moderately or highly cellular air-dried May-Grünwald–Giemsa stained smear, thereby ensuring lesional representation. The quantity of DNA for molecular testing was insufficient in four cases with adequate DNA in the remaining 80 cases. Category 5 smears showed the least amount of DNA obtained followed by category 3. The reason for this was the lower smear cellularity and in these two categories, we used two smears generally to obtain adequate DNA. DNA in direct cytology smears is easily extractable and is stable for six months to a few years.[8] The use of stained rather than unstained smears offers the advantage of ensuring lesional representation. Diff-Quik stained smears can provide high-quality DNA for sophisticated molecular studies.[89] Further, DNA appears to be preserved better in such smears as compared to Papanicolaou-stained smears, where DNA degradation occurs as a function of aging.[9] DNA extracted from microdissected direct cytologic smears from various sites like lung, lymph node, liver, soft tissue, and thyroid was also found suitable for next-generation sequencing.[10] In another study, Ferraz et al. have also extracted RNA from air-dried smears of thyroid aspirates for molecular testing of gene rearrangement.[11] We reaffirm that cytoscrapes from air-dried MGG stained smears provide sufficient DNA for molecular testing in thyroid aspirates in all categories. The RT-PCR based kit employed required only 100 ng per sample for testing five gene mutations (BRAF V600E, NRAS, and HRAS, and two mutations in KRAS genes) which is an advantage in cytology samples. This is a sensitive assay with a quick turnaround time of 3–4 hours and cost per test being approximately INR 3000 makes it affordable to our patients with limited resources.

PTC can present as metastasis in the cervical lymph node with or without a palpable thyroid lesion. In all the cases of nodal metastases included in this cohort, the thyroid primary was not clinically palpable. Of 11 cases, molecular testing was possible in ten cases and five showed BRAF-V600E mutation and one showed NRAS mutation. The thyroid primary lesion was confirmed subsequently upon imaging and ultrasound-guided fine-needle aspiration cytology. In such a clinical scenario, confirmation of the cytomorphological diagnosis by molecular testing can be a useful ancillary technique.

The overall mutation (BRAF or RAS) positive cases were 54.9%. Taking histopathology as the gold standard, the sensitivity, specificity, PPV, and NPV of detecting a malignant lesion with any of these mutations were 61.7%, 80%, 91%, and 38%, respectively. Eszlinger et al.[12] showed a sensitivity of 75.3% and 90.4% specificity for detecting the malignant lesions by combining the FNA and molecular studies. Similar results were seen in literature where the presence of any mutations has been found to be a good predictor of malignancy.[51314] Studies in the literature have also shown that by performing molecular tests, the sensitivity of detecting malignant thyroid nodules increases to 80–90%[1415]; however, these studies have also included detection of RET-PTC and PAX/PPARγ. The detection of gene rearrangements requires the extraction of total RNA from the sample which was not performed in this study. Few authors have demonstrated the feasibility of extracting RNA from routine air-dried FNA smears for the detection of PAX8/PPARG and RET/PTC rearrangements with RT-qPCR.[1112] This will add to the overall cost of the test as cDNA would need to be synthesized from the extracted mRNA followed by appropriate controls and capillary sequencing of any gene fusion detected by qRT-PCR.

BRAF V600E mutation was seen in 53.8% of histologically proven PTC which is concordant with several studies and is reviewed by Nikiforov.[15] In category 3, BRAF V600Emutation was seen in just one case (5%) but was seen in 19% in category 5; none was seen in category 4. Studies have shown 15–39% BRAF positivity in indeterminate/nondiagnostic cases.[151617] In a meta-analysis, Su et al reported BRAF positivity rates of 43.2% in suspicious for malignancy, 13.77% in AUS/FLUS, and 4.43% in follicular neoplasm categories.[18] However, it is highly specific for a diagnosis of PTC and in another meta-analysis, the prediction of malignancy (POM) of BRAF for PTC was found to be 99.8%.[17] In suspected nodal metastases of PTC, BRAF mutation was positive in just 50% cases, which was lower than expected as BRAF V600E-positive PTCs are well-known to have aggressive behavior and nodal metastases.[19]

NRAS mutation was seen predominantly in the category 3 (AUS/FLUS) group with a 66.7% positivity rate. Histology follow-up of eight NRAS mutation–positive cases revealed malignancy in seven cases with follicular carcinoma most commonly observed with just a single case of follicular adenoma. HRAS mutation positivity was restricted to Bethesda categories 3 and 4 with either follicular carcinoma or adenoma outcome. Some studies have shown that NRAS and HRAS– positive follicular adenomas are precursors for follicular carcinoma.[671120] However, in a recent meta-analysis, Nabhan et al. reported the wide variability in the frequency of RAS mutations in the indeterminate cytology category. They concluded that future studies are required to clarify the management of RAS-mutated nodules and to understand the potential pathways from a RAS-mutated benign neoplasm to invasive carcinoma.[21]

In the present study, FVPTC cases (two cytologically suspected and five histologically proven) showed NRAS and BRAF V600E mutations in one case each which was concordant with the other similar studies.[6722] Interestingly, we did not encounter any histologically proven case of noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) in this cohort of patients which differs from the previously reported histological outcome for the indeterminate category from other parts of the world.[712232425] A total of 35 cases did not reveal any mutations in the BRAF/RAS genes with an equal probability of benign or malignant outcome clearly implying the low negative predictive value of this molecular assay. Further, no mutations were detected in 21/43 (48.8%) cases in the indeterminate cases. Hwang et al.[24] reported 21.3% of indeterminate cases were negative for NRAS or BRAF mutations. Nikiforov et al.[515] studied seven gene panels including RET/PTC1, RET/PTC3, and PAX8/PPARG rearrangements and showed only 6% of indeterminate nodules were negative for these mutations. This clearly implies the need for an extended panel to include gene fusions to resolve a greater number of cases. However, this requires RNA extraction from the sample which was not performed due to limited resources. In a recent report by Bellevicine et al.,[26] a seven-gene panel was tested in 162 AUS/FLUS cases of which, only four gene fusions (PAX8-PPARg and RET/PTC) were detected, implying their rarity in this category; RAS gene mutations were the most common supporting our observations. A correlation of BRAF V600E with a CyA qualifier was observed. Hence, testing for 5-gene mutations only may be justified in the indeterminate nodules.

The limitations of this study are the small number of subjects who were tested only for gene mutations and not evaluated for RET-PTC and PAX8-PPARγ gene rearrangements. To conclude, this is the first study from India to demonstrate the application of molecular testing in FNA of thyroid using cytoscrapes. The commercially available multiplex RT-PCR assay kit testing for 5-genes requires only a small amount of DNA and is easy to perform in routine molecular cytology laboratories with modest resources, as it is a low cost procedure with a quick turnaround time. Among the indeterminate thyroid categories, this 5-gene molecular assay showed a PPV of 81% making it a good rule-in test.

Financial support and sponsorship

The study was supported by a Departmental grant, PGIMER, Chandigarh.

Conflicts of interest

There are no conflicts of interest.

Presented at: International Cytology Congress, Sydney, 2019.