Associations of the BRAFV600E Mutation with Sonographic Features and Clinicopathologic Characteristics in a Large Population with Conventional Papillary Thyroid Carcinoma

Objective To evaluate the association of the BRAFV600E mutation with sonographic features and clinicopathologic characteristics in a large population with conventional papillary thyroid carcinoma (PTC). Methods We retrospectively reviewed the sonographic features, clinicopathologic characteristics, and presence of the BRAFV600E mutation in 688 patients who underwent thyroidectomy for conventional PTC between January and July 2010 at a single institution. The incidence of the BRAFV600E mutation was calculated. The sonographic features and clinicopathologic characteristics were compared between BRAF-positive and BRAF-negative patients. BRAF-positive patients were subdivided into those with papillary thyroid microcarcinoma (the PTMC group) and those with PTC larger than 10 mm (the PTC>10 mm group), and their sonographic features were compared. Results The BRAFV600E mutation was detected in 69.2% of patients (476 of 688). Sonographic features were not significantly different between BRAF-positive and BRAF-negative PTC, nor between PTMC and PTC>10 mm groups. The BRAFV600E mutation was associated with male sex (P = 0.028), large tumor size, extrathyroidal extension, central and lateral lymph node metastasis, and advanced tumor stage (P<0.0001). Conclusion The BRAFV600E mutation was significantly associated with several poor clinicopathologic characteristics, but was not associated with sonographic features, regardless of tumor size. We recommend that patients with a thyroid nodule with any suspicious sonographic feature undergo preoperative BRAFV600E testing for risk stratification and to guide the initial surgical approach in PTC.


Introduction
Papillary thyroid carcinoma (PTC) is the most common thyroid cancer, accounting for 85-90% of cases, with an increasing incidence globally [1,2]. The development of high-resolution ultrasonography (US) has contributed to the detection and diagnosis of PTC with high specificity, but relatively low sensitivity [3,4]. US-guided fine-needle aspiration (US-FNA) biopsy is a standard tool for diagnosing thyroid malignancies preoperatively with high specificity. Its major limitation, however is the 15-25% rate of indeterminate cytology (Bethesda category I-nondiagnostic or category III-atypia of undetermined significance or follicular lesion of undetermined significance) [5].
Various genetic analyses have improved the diagnostic performance of US-FNA. Above all, the B-type Raf kinase (BRAF) mutation has received the most attention in recent years because of its high prevalence and high specificity for PTC. The BRAF mutation induces aberrant activation of the mitogen-activated protein kinase pathway, which plays a fundamental role in cell proliferation, differentiation, and apoptosis, finally resulting in tumorigenesis [1,6,7]. The T1799A point BRAF mutation is the most common mutation found in the BRAF gene, accounting for more than 90% of mutations. It causes a V600E amino acid change in the BRAF protein, resulting in a BRAF V600E mutation that occurs exclusively in PTC with a prevalence ranging from 29 to 83% [1,7]. Several previous reports have demonstrated that BRAF V600E mutation testing may enhance the diagnostic accuracy of US-FNA for PTC [8][9][10][11][12][13].
Therefore, we evaluated the association of the BRAF V600E mutation with sonographic features and clinicopathologic characteristics in a large-scale study population with conventional PTC.

Materials and Methods
The institutional review board of Gangnam Severance hospital approved of this retrospective observational study and required neither patient approval nor informed consent for our review of patients' images and records. However, written informed consent was obtained from all patients for US-FNA and BRAF V600E mutation analysis prior to each procedure as a daily practice.

Patients
Our hospital's institutional review board approved this retrospective observational study and waived the requirement for informed consent. Between January and July 2010, 939 consecutive patients underwent thyroidectomy and were diagnosed with conventional PTC at our institution. Of these, 251 patients were excluded for the following reasons: patient refusal of BRAF V600E mutation analysis of the surgical specimen (n = 87), lack of preoperative US at our institution (n = 161) and inability to identify the lesion on US (n = 3). A total of 688 patients were included in this study (553 women and 135 men; mean age, 45 years; range, 17-83).
All patients were diagnosed preoperatively with malignancy by US-FNA at our institution (n = 222) or outside clinics (n = 466). US-FNA was performed on all suspicious thyroid nodules larger than 5 mm, and on nodules smaller than 5 mm at the patient's or clinician's request. Cytologic results were as follows: 531 papillary carcinomas, 147 cases suspicious for papillary carcinoma, 8 cases with atypia of undetermined significance, 1 benign follicular nodule, and 1 lymphocytic thyroiditis. Eight atypical lesions were surgically removed due to sonographic features compatible with papillary carcinoma (n = 7) or histologic confirmation of malignancy through core biopsy (n = 1). Two benign cases were surgically removed due to sonographic feature compatible with papillary carcinoma and lymph node metastasis on US-FNA.
All patients received curative surgery with either total thyroidectomy (n = 517) or near-total thyroidectomy (n = 171). Prophylactic or therapeutic central-compartment neck dissection was performed for all patients. Lateral compartmental lymph node dissection was performed for patients with US-FNA-proven or clinically suspicious lateral cervical lymphadenopathy (n = 61).

Ultrasound examination
US images were obtained using either HDI5000 or IU22 ultrasound scanners (Philips Healthcare, Bothell, WA) equipped with a 7.5-12 MHz linear array transducer. Four radiologists who specialize in thyroid US with between 1 and 12 years of experience (E.J.S., J.A.K., J.H.Y., and A.Y.P.) performed all US examinations before BRAF V600E mutation analysis was conducted. Two radiologists (A.Y.P and E.J.S, with 1 and 12 years of experience, respectively) blinded to BRAF status retrospectively analyzed the following sonographic features in consensus: tumor size, composition (solid or cystic), echogenicity with respect to the thyroid parenchyma and strap muscle (hyperechoic, isoechoic, hypoechoic or markedly hypoechoic), margin (circumscribed, microlobulated or irregular), calcifications (microcalcification, macrocalcification, or negative), and shape (parallel or nonparallel). US findings of microcalcification, irregular or microlobulated margin, marked hypoechogenicity, and nonparallel shape are considered indicative of malignancy. Final assessment category was classified according to the number of suspicious features, which was based on the modified thyroid imaging reporting and data system (TIRADS) suggested by Kwak et al. (4): probably benign (no suspicious US feature), low suspicion for malignancy (1 suspicious feature), intermediate suspicion for malignancy (2 suspicious feature), moderate suspicion for malignancy (3 suspicious features) or highly suggestive of malignancy (4 suspicious features).

Clinicopathologic data analysis
Medical records were reviewed to determine sex and age at diagnosis. Pathology reports were reviewed for the following data: tumor size, presence of BRAF V600E mutation, multifocality, extrathyroidal extension, and central or lateral lymph node metastasis. TNM stage was assigned based on the AJCC Cancer Staging Manual 7 th edition [32]. Patients were divided into two groups according to pathologic tumor size less than 10 mm (papillary thyroid microcarcinoma, PTMC group) or greater than 10 mm (PTC.10 mm group).

BRAF V600E mutation analysis
DNA was extracted from 10-mm-thick sections of paraffin blocks using the QIAamp DNA FFPE Tissue Kit (Qiagen, Hilden, Germany). DNA was extracted only from the marked tumor tissue. The presence of a BRAF mutation was evaluated by polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) or direct sequencing. For PCR-RFLP, a 50-mL PCR mixture was composed of extracted DNA at 100 ng/mL, 5 U Taq polymerase, 0.25 mM dNTP, 20 pmol of primers, and 106Ex Taq buffer. To examine the BRAF exon 15, a primer that created a restriction site for the BspE1 enzyme was designed. The PCR conditions were as follows: denaturation at 95uC for 10 minutes, followed by 40 cycles at 94uC for 30 seconds, 45uC for 30 seconds, 72uC for 30 seconds, and a final extension step at 72uC for 10 minutes. The PCR product was purified with the MinElute PCR purification kit (Qiagen), digested with 10 units of BspE1 (Takara, Tokyo, Japan) and electrophoresed in a 4% agarose gel containing ethidium bromide. The stained gel was photographed using an ultraviolet light transilluminator. For direct sequencing, BRAF exon 15, which contains the codon encoding the V600E mutation, was amplified. The PCR conditions were as follows: denaturation at 94uC for 5 minutes, followed by 40 cycles at 94uC for 20 seconds, 56uC for 30 seconds, 72uC for 30 seconds, and a final extension step at 72uC for 5 minutes. PCR products were purified using the Exo I/SAP Clean-Up protocol (Hilden, Germany), and direct DNA sequencing was performed using the 3730 Big Dye Terminator v 3.1 Sequencing Standard (Applied Biosystems, Foster City, CA).

Statistical analysis
The incidence of the BRAF V600E mutation was calculated. Sonographic features and clinicopathologic characteristics were compared between the patients with and without the BRAF V600E mutation. The correlation between sonographic features and the BRAF V600E mutation was also evaluated in the PTMC and PTC. 10 mm groups. The chi-squared or Fisher's exact test was used for categorical variables and the Student's t-test was used for continuous variables. Multivariate logistic regression analysis was performed for the association of the BRAF V600E mutation with sonographic features in patients with PTC. Statistical significance was accepted for P-values less than 0.05. All statistical analysis was performed using SPSS version 19.0 (SPSS, Inc., Chicago, IL).

Sonographic features
BRAF-positive PTC were significantly larger than BRAFnegative PTC on US (10.866.8 mm vs. 8.464.9 mm, P, 0.0001). There were no significant differences in sonographic features between BRAF-positive and BRAF-negative PTC, including composition, echogenicity, margin, calcification, shape or final assessment (Table 1). Multivariate logistic regression analysis also showed no association between the BRAF V600E mutation with suspicious sonographic features (

Discussion
Previous meta-analyses have published values for the overall prevalence of the BRAF V600E mutation ranging from 29 to 83% [7,[14][15][16]. This wide range may be due to variations in PTC subtype, subjects' geographical backgrounds, and research methodology. Korea appears to have a relatively high frequency of the BRAF V600E mutation, ranging from 52 to 83% [14,15,23,25,33]. Our large-scale study of conventional PTC in Korea also found a relatively high prevalence of 69.2%.
In regards to the association between the BRAF mutation and sonographic features of PTC, two recent studies reported no significant difference in sonographic features between BRAFpositive and BRAF-negative PTC [2,13]. A Korean study on PTMC alone also reported no significant difference [31]. In contrast, a recent study on 115 patients with PTC larger than 10 mm found that BRAF-positivity was associated with suspicious sonographic findings and the number of suspicious features has positive correlation with the risk of BRAF positivity [30]. The authors suggested that difficulties in accurate sonographic characterization of small PTC confounded studies including small PTC. However, our study found no significant difference in any sonographic feature between BRAF-positive and BRAF-negative PTC in the PTC.10 mm group as well as PTMC group. Final assessment category classified according to the number of suspicious features was also not significantly different between BRAF-positive and BRAF-negative PTC, although the PTC. 10 mm group had more tendency in correlation between BRAFpositivity and final assessment category than PTMC group did (Pvalue, 0.065 vs. 0.500). We assume that such discrepancy might result from different study population between ours and the abovementioned study. Most suspicious sonographic features used in both studies are oriented to diagnosis of conventional PTC, therefore only inclusion of conventional PTC in our study may predispose little distinction of sonogrpahic features between BRAF-positive and BRAF-negative PTC. The result of our study suggests that there is no specific sonographic feature to be an indication for performing additional BRAF V600E mutation analysis to FNA for the thyroid nodules suspicious of PTC.
The tumor size of BRAF-positive PTC, both sonographic and pathologic, were larger than that of BRAF-negative PTC. Previous studies have also reported an association between the BRAF mutation and large tumor size, which suggests the possibility that the mutation induces tumor progression and aggressiveness [23,26,34,35]. However, other studies have reported conflicting results [2,14,18,22,33], indicating that the relationship between the BRAF mutation and tumor size remains controversial. Our results also showed that the BRAF mutation was more frequent in men, consistent with previous studies [26,36].
In respect to the relationship between the BRAF mutation and clinicopathologic PTC characteristics, many studies have reported that one or more high-risk clinicopathologic parameters were associated with the BRAF mutation [14][15][16][17][18][19][20][21][22][23][24][25][26][27][28]. Two recent metaanalyses that included 5655 and 2470 PTC patients, respectively, found a significant association between the BRAF mutation and lymph node metastasis, extrathyroidal extension, advanced tumor stage and recurrence [15,16]. In addition, a recent study with a median 15 years of follow-up demonstrated that the BRAF mutation was related to advanced tumor stage, vascular invasion, and mortality [19]. Our results also revealed a relationship between BRAF-positive PTC and high-risk clinicopathologic characteristics, including extrathyroidal extension, lymph node metastasis and advanced tumor stage. On the contrary, several studies employing multivariate analysis with adjustment of Table 1. Association of the BRAF V600E mutation with sonographic features in patients with PTC. confounders found no such relationship with these characteristics [14,33,[37][38][39]. Possible reasons for this discrepancy in results include variations in PTC subtypes, geographic or ethnic factors, scale of enrolled data, disease extent at the time of diagnosis, methodology of BRAF analysis, and the use of prophylactic central neck compartment dissection. Our study has several unique strengths. First, this is the largest single-center study on conventional PTC, which limits participant heterogeneity and variations in tumor subtypes. Second, our finding that the BRAF mutation was associated with poor clinicopathologic parameters suggests the utility of preoperative BRAF analysis in risk stratification and surgical management, especially in cases of equivocal extrathyroidal extension or cervical node metastasis on preoperative US, or nondiagnostic result of cervical lymph node metastasis on US-FNA. Lastly, we analyzed the relationship between BRAF mutation and sonographic features in both PTMC and PTC.10 mm groups to establish the previous controversial results according to the tumor size and observed the result that there was no difference in sonographic features between the BRAF-positive PTC and BRAF-negative PTC in both PTMC and PTC.10 mm.
Our study has several limitations. First, this was a retrospective observational study, a design that prevents long-term follow-up or analysis of actual clinical outcomes. Second, we did not evaluate interobserver variability of PTMC sonographic features, despite known difficulty in accurate characterization of small tumors. Finally, we did not evaluate the Doppler or elastographic tumor findings, which can provide additional diagnostic information. Future areas of research include prospective long-term follow-up and Doppler or elastographic evaluation of PTC.

Conclusion
The BRAF mutation was not associated with particular sonographic features in conventional PTC, regardless of tumor size. However, the mutation was significantly associated with poor clinicopathologic parameters including male gender, large tumor size, extrathyroidal extension, lymph node metastasis and advanced tumor stage. Our results suggest the utility of preoperative BRAF V600E mutation analysis of thyroid nodules with any suspicious sonographic feature for risk stratification and determination of the initial surgical approach in PTC.

Author Contributions
Conceived and designed the experiments: EJS. Performed the experiments: EJS AYP. Analyzed the data: AYP EJS. Contributed reagents/materials/ analysis tools: JAK JHY YJP CSP HSC. Wrote the paper: AYP EJS. Drafted the article or revised it critically for important intellectual content: EJS AYP JAK JHY.