BRAF and NRAS Mutations in Papillary Thyroid Carcinoma and Concordance in BRAF Mutations Between Primary and Corresponding Lymph Node Metastases

Concordance between mutations in the primary papillary thyroid carcinoma (PTC) and the paired x lymph node metastasis may elucidate the potential role of molecular targeted therapy in advanced stages. BRAF and NRAS mutations in primary PTC (n = 253) with corresponding metastatic lymph node (n = 46) were analyzed utilizing StripAssays (ViennaLab Diagnostics). Statistical analysis was performed using (SPSS, Inc.), version 24.0 with a p-value of <0.05, and concordance via kappa agreement. BRAF mutation frequency in conventional PTC (cPTC): 56.8%, papillary thyroid microcarcinoma (PTMC): 36.5%, PTMC-FV: 2.7% and PTC-FV: 4.1%. NRAS mutation frequency in PTC-FV: 28.6%, PTMC: 28.6%, PTMC-FV: 23.8%, and cPTC: 19.0%. BRAF mutation correlation with older age in cPTC (42.6 versus 33.6) years (p < 0.001) was the only significant clinicopathologic parameter. BRAF mutations were concordant in the primary and its corresponding lymph node deposits in PTC with a kappa of 0.77 (p-value < 0.0001). BRAF mutations are predominant in cPTC and PTMC while NRAS mutations in PTC-FV. BRAF mutation is conserved in metastatic lymph node deposits, thus BRAF is an early mutational pathogenetic driver. Therefore, targeted therapy is potential in recurrent and advanced stage disease.

Several studies reported an association of the BRAF mutational status with a number of PTC clinicopathologic parameters inclusive of recurrence and worse prognosis 10,11 . For instance, Nikiforova et al. correlated between BRAF status on one hand and both advanced age (5 th decade) and extrathyroidal extension on the other hand 12 , while others reported no significant correlation 13,14 . Furthermore, there is conflicting evidence with respect to BRAF mutational analysis on cytology smears as a guide for further surgical management. Alternative studies reported a prediction of lymph node status based on BRAF cytology 15,16 , while Barabaro et al. identified no significant association 17 . Yet, there is increasing evidence that coexistence of this BRAF mutation with other promoter mutations, specifically TERT promoter mutations, might form a genetic background defining PTC with the worst clinicopathologic parameters and outcomes 18 . Specifically, the C228T TERT promoter mutation has been shown to be associated with the BRAF V600E mutation, which was prevalent in the aggressive types of thyroid cancer 19 .
In the era of targeted therapy, which is based on the understanding of tumor molecular biology, it is critical to determine the molecular profiles in both the primary and metastatic sites as well. The use of anti-BRAF therapy is currently under investigation in clinical trials for cases of advanced surgically unresectable and/or radioresistant thyroid cancer cases 20 . Therefore, we aimed in this study to determine the BRAF and NRAS molecular signature concordance rates between the four main different primary PTC subtypes and the corresponding paired metastatic lymph node deposits in order to elucidate the potential clinical implication of selective molecular targeted therapy in advanced stage PTC. Additionally, we sought to determine the frequencies and types of BRAF and NRAS mutations in a cohort of Lebanese patients and correlate between the findings and the various clinicopathologic features of individual PTC subtypes: conventional PTC (cPTC), papillary thyroid microcarcinoma (PTMC) defined as tumors measuring ≤1 cm in maximum diameter, follicular variant of PTMC (PTMC-FV) and the follicular variant of PTC (PTC-FV).

Materials and Methods
Patient Selection. All patients enrolled in this retrospective clinical study gave informed consents for both participation and publication of identifying information/images (when applicable). The study with all its experimental protocols was conducted under the Institutional Review Board (IRB) approvals of the American University of Beirut Medical Center (AUBMC) and Hammoud Hospital University Medical Center (HHUMC). All experiments were performed in accordance with relevant guidelines and regulations. Archived formalin fixed paraffin embedded (FFPE) tissues of 312 PTC patients were collected from the Departments of Pathology and Laboratory Medicine, AUBMC and HHUMC, Beirut, Lebanon, between the period of January 2001 and December 2011.
Patients Tissue Sampling. Out of the 312 PTC cases, 253 PTC cases with available paraffin blocks and a minimal tumor size of 1 mm underwent mutational analysis. All the 253 PTC cases were analyzed for BRAF and KRAS, and only 202 with available extracted DNA underwent analysis for NRAS mutations (Fig. 1). As a negative control, 15 cases of multinodular goiter were used. Demographic (age and gender) and prognostic histopathologic features (tumor size, lymphovascular invasion, extrathyroidal invasion, focality, and lymph node metastasis) were evaluated and correlated with the molecular aberrations. Lymph node dissection was performed on128 cases of the 253, out of which 62 had metastatic lymph node deposits. Yet, only 46 cases with available paraffin blocks and a minimal tumor size of 1 mm underwent mutational analysis. Patients' consents were waivered by the IRB because this is a retrospective study.
DNA Extraction and Quantification. DNA was extracted from 253PTC cases utilizing the QIAamp FFPE DNA extraction kit (Qiagen, California, USA), and quantified via the Qubit fluorometer (Thermofisher Scientific, USA). The extracted DNA was stored at −20 °C until further use. StripAssays (ViennaLab Diagnostic GmbH, Vienna, Austria) were utilized to detect different point mutations and deletions in the genes coding for BRAF and NRAS. The detection sensitivity for mutant alleles is 1%, performed according to the manufacturer's instructions. Mutational analysis was performed by polymerase chain reaction (PCR) and reverse hybridization as follows: first, a multiplex PCR amplification using biotinylated oligonucleotide primers was performed for BRAF, KRAS, and NRAS gene sequence amplification; second, reverse hybridization of the amplification products was ensued via a test strip, which contains allele-specific oligonucleotide probes for mutations and controls immobilized on a parallel array; and finally, bound biotinylated sequences were visualized using streptavidin-alkaline phosphatase conjugate and enzymatic color development. Positive control samples included defined mutated cell line DNA or clones.   Statistical Analysis. Data were entered into a Microsoft Excel datasheet, and then transferred to the Statistical Package for Social Science software (SPSS, Inc.), version 24.0, which was used for data management, cleaning, and analyses. Descriptive statistics was carried out and reported as number and percent for categorical variables, whereas the mean and standard deviation (±) for continuous ones. Association between mutation and demographic, clinical and pathological data was assessed using the Chi square test or Fisher's exact test for categorical variables, and student's independent t-test or Mann Whitney test for continuous ones. Moreover, to assess for the agreement between the primary tumor and its corresponding lymph node metastasis, kappa agreement was calculated and reported along with the p-value. Statistical significance was specified at 0.05 levels.

Results
Thyroid Cancer Patients' Demographics. AUBMC  BRAF mutation in cPTC was significantly correlated to older age (BRAF mutated cPTC, mean age = 42.6 ± 14.5 years vs. wild-type cPTC, mean age = 33.6 ± 15.1 years, p = 0.005). A trend towards higher incidence of BRAF mutation was found in patients with higher tumor stage (p = 0.054). There was no significance association with respect to BRAF and NRAS mutations in the remaining cPTC clinicopathologic features ( Table 2).
In PTMC and PTMC-FV, clinicopathologic parameters were not significantly correlated with neither BRAF nor NRAS mutations. However, there was a higher trend for BRAF mutation with multifocality in PTMC (Tables 3 and 4). Similarly, PTC-FV did not correlate with any clinicopathologic feature; however, we noticed that BRAF mutations were exclusive to tumors sizes smaller than or equal to 3 cm, absence of extrathyroidal extension, and absence of lymphovascular invasion, while NRAS mutations were exclusive to females and absence of extrathyroidal extension. (Fig. 3, Table 5).

BRAF Mutational concordance between primary PTC and paired lymph nodes metastasis.
BRAF mutations were concordant in the primary and its corresponding lymph node deposits in PTC with a kappa of 0.77 (p-value < 0.0001) (Fig. 4, Table 6). Agreement coefficients for mutational concordance between primary and paired lymph node deposits were not calculated for NRAS mutations due to the small number of NRAS mutated cases and their corresponding lymph node metastasis.

Discussion
The current study evaluated the concordance rates of BRAF and NRAS mutations between primary PTC tumors and paired metastatic lymph node deposits of the four most common subtypes of PTC: cPTC, PTMC, PTMC-FV and PTC-FV. In addition, the mutational BRAF and NRAS statuses were correlated with the different clinicopathologic parameters. BRAF and RAS mutations are the most common in PTC 21,22 . In this series, we found that BRAF mutation incidence, approximated to be 60%, was closer to the higher edge of the worldwide reported range (36-69%), while NRAS was lower with approximately 11% vs. 30% reported in literature [23][24][25][26] . Comparably, we found that BRAF mutations were more prevalent than NRAS mutations in cPTC (56.8% vs. 50%) and PTMC (36.5% vs. 40%), whereas NRAS mutations showed a higher incidence than BRAF mutations in PTMC-FV (23.8%) and PTC-FV (28.6%). Interestingly, we identified a significantly elevated NRAS mutational frequency within PTMC (28.6%) similar to PTC-FV (28.6%); a finding higher than that reported by Schulten et al.  Clinicopathologic parameters' correlation with BRAF and NRAS mutations is controversial among different studies [12][13][14] . In a cohort of 129 PTMCs tested for BRAF V600E mutation and their correlation with the clinicopathologic features of patients, results showed no significant differences in age, sex, tumor size, location, and multifocality between the BRAF V600E mutated and non-mutated microcarcinomas 9 . However, there was significantly higher prevalence of infiltrative tumor borders, tumor-associated stromal desmoplasia/fibrosis and/or sclerosis, classic nuclear features of PTC, and cystic change in mutated microcarcinomas 9 . Similarly, results from another study demonstrated significant association between BRAF mutation-positive tumors and the following features: infiltrative growth, stromal fibrosis, psammoma bodies, plump eosinophilic tumor cells, and classic fully developed nuclear features of PTC, but not other clinicopathological parameters 24 . In addition, BRAF mutational status has been correlated with recurrence of PTMCs, suggesting its importance in stratifying patients for surgical management 28,29 . On the other hand, several papers concluded that BRAF positivity is not significantly associated with most clinicopathologic features redolent of aggressiveness, including tumor multicentricity, lymphovascular invasion, extranodal extension, central neck involvement, advanced stage (stage III or IV), and distant metastasis 30,31 .
In our cases, the only significant clinicopathologic correlation found was between advanced age and BRAF mutation in cPTC (p < 0.005), a potential causal link between older age and an advanced stage disease presentation. While Rodolico et al. identified BRAF mutations in 41% of PTMCs and an association with a higher age (mean = 53 years) and lymph node metastasis 32 , we reported a frequency of 36.5% BRAF mutated cases in PTMCs but with no statistically significant correlation with the various clinicopathologic parameters. Yet, a trend towards higher incidence of BRAF mutation was found in patients with higher tumor stage (p = 0.054). That being said, the clinical benefit of selective molecular targeted therapy in aggressive and advanced stage PTMC is still questionable 33 .
PTC-FV, which was initially described by Lindsay et al. 34 and categorized by Chem and Rosai due to the morphologic and biological overlap with PTC 35 , represents a unique molecular subgroup of PTC cases. At the molecular level, and in contrast to cPTC and PTMC, PTC-FV exhibits a RAS family mutation. The Cancer Genome Atlas clustered PTC into two main morphologically and molecularly distinct groups, namely BRAF driven and RAS mutated tumors 36 . Nikiforov et al. recommends that the encapsulated variant of PTC-FV is best classified as "noninvasive follicular thyroid neoplasm with papillary-like nuclear features" (NIFTP) due to the low risk malignant behavior. Only cases with the infiltrative pattern retain the PTC-FV term 37 . One case of PTC-FV harbored lymph node metastasis and was negative for the BRAF or NRAS mutation, while none of the PTMC-FV  cases exhibited lymph node metastasis. The literature on lymph node metastasis in PTC-FV varies greatly among different studies and ranges between14% and 94% 38 . Locoregional lymph node metastasis in PTC may be found in up to 46.8% 39 . In high-risk patients, characterized by older age, tumor size >3 cm, and extracapsular extension, the number and size of lymph node metastasis affects prognosis and survival. Locoregional recurrence, with a follow-up of three decades, can reach up to 30% [40][41][42] . The current study showed a highly significant concordance rate of 84% for BRAF mutation in primary PTC and corresponding paired lymph node metastasis. Similarly, Walts et al. and Vasco et al. reported concordance rates of 95.2% and 81% respectively for primary PTCs and the corresponding paired metastatic lymph node deposits 43,44 . This implies that BRAF mutation is conserved in both the primary and paired metastatic lymph nodes, thus supporting the hypothesis of a driver mutational role in the pathogenesis of PTC, particularly cPTC and PTMC, a finding reinforced by the genomic analysis of PTC via the Cancer Genome Atlas Research Network 36 . Therefore, does BRAF testing predict central lymph node metastasis and an aggressive PTC phenotype? Actually, the positive predictive value and negative predictive values of BRAF mutational testing in PTC as a marker of central lymph node metastasis were estimated to be 47% and 91%, respectively 45 . Hence, the utility of BRAF as a prognostic marker may be confined to the cPTC subtype 46 .
Argumentatively, there is a potential role of selective molecular targeted therapy in recurrent and advanced metastatic PTC cases that are surgically unresectable and radioresistant. Phase II clinical trials utilizing Selumetinib, a tyrosine kinase inhibitor targeting BRAF mutations in PTC, were conducted without any significant survival benefit 47 . Currently, a study by Dadu et al. involving treatment of advanced cPTC stage disease exhibited a 47% partial response and a 53% stable disease over a minimal 6-month period 48 . The BRAF status of the paired lymph node deposits was not determined in the study by Dadu et al. An interesting prospective study may identify responders versus non-responders with respect to metastatic lymph node BRAF status. In our study, NRAS mutations within metastatic lymph nodes were detected only in cPTC and PTMC, but the numbers are too low to conclude a significant concordance rate in either.   This study carries a number of limitations that relate to the relatively small number of cases evaluated, especially PTMC-FV and PTC-FV cases, and accordingly data may not apply to the different subtypes of PTC. Besides, the study is also limited by being retrospective in nature.

Conclusion
In conclusion, BRAF mutation is conserved in the primary and paired metastatic lymph node deposits in cPTC and PTMC. Testing for the BRAF mutation within lymph nodes is recommended in order to identify responders to the selective tyrosine kinase inhibitors in advanced stage cPTC. The high prevalence of BRAF and NRAS in PTMC and PTMC-FV with the absent significant clinicopathologic correlation undermines the role of BRAF testing in such a predominantly curable malignant thyroid disease. Finally, NRAS and BRAF testing in PTC-FV comprise a potentially diagnostically reassuring result. Further prospective studies are required to assess BRAF status within primary and paired lymph nodes for patients treated with selective targeted therapy in advanced stage cPTC. Table 6. Agreement in BRAF mutation in between primary PTC tumor and the corresponding metastatic lymph nodes. Concordance in BRAF mutation between primary PTC and the corresponding metastatic lymph nodes.