BRAF mutation correlates with recurrent papillary thyroid carcinoma in Chinese patients

The prevalence of thyroid cancer has been increasing worldwide1, including in China2. The disease has become the fastest-increasing cancer, and it ranks 5th among the most common cancers in women3. Notably, the current increase is almost entirely attributable to papillary thyroid carcinoma (ptc)4. The prognosis of differentiated thyroid cancer is benign, with a survival rate of 96.4% after 30 years followup5. However, a significant percentage of ptc patients experience recurrent disease or distant metastasis, reducing the survival rate to 40%6. Several staging systems have been applied to stratify the risk of a poor outcome with ptc. However, all are based on histopathologic parameters after surgery7. They therefore cannot be used before an operation to determine the extent of surgery. With the emerging understanding of molecular genetics in thyroid cancer, several specific mutations in ptc have been determined8. The BRAF V600E mutation has been associated with worse prognostic features (such as extrathyroidal extension, lymph node metastasis or advanced tumour stage), poor clinical outcome, and mortality9–13. The RET/PTC oncogenes are believed to play an important role in radiation-induced ptc14. In contrast to BRAF mutation, RET/PTC rearrangement has been associated with better prognosis15. Yet despite those findings, controversy remains16–21. In addition, questions about the clinical significance of these mutations in the management of microptc are ongoing22. ABSTRACT


INTRODUCTION
The prevalence of thyroid cancer has been increasing worldwide 1 , including in China 2 .The disease has become the fastest-increasing cancer, and it ranks 5th among the most common cancers in women 3 .Notably, the current increase is almost entirely attributable to papillary thyroid carcinoma (ptc) 4 .The prognosis of differentiated thyroid cancer is benign, with a survival rate of 96.4% after 30 years followup 5 .However, a significant percentage of ptc patients experience recurrent disease or distant metastasis, reducing the survival rate to 40% 6 .
Several staging systems have been applied to stratify the risk of a poor outcome with ptc.However, all are based on histopathologic parameters after surgery 7 .They therefore cannot be used before an operation to determine the extent of surgery.With the emerging understanding of molecular genetics in thyroid cancer, several specific mutations in ptc have been determined 8 .The BRAF V600E mutation has been associated with worse prognostic features (such as extrathyroidal extension, lymph node metastasis or advanced tumour stage), poor clinical outcome, and mortality [9][10][11][12][13] .The RET/PTC oncogenes are believed to play an important role in radiation-induced ptc 14 .In contrast to BRAF mutation, RET/PTC rearrangement has been associated with better prognosis 15 .Yet despite those findings, controversy remains [16][17][18][19][20][21] .In addition, questions about the clinical significance of these mutations in the management of microptc are ongoing 22 .

Purpose
We investigated correlations of somatic BRAF V600E mutation and RET/PTC1 rearrangement with recurrent disease in Chinese patients with papillary thyroid carcinoma (ptc).

Methods
This prospective study included 214 patients with ptc histologically confirmed between November 2009 and May 2011 at a single institute.

Results
We found somatic BRAF V600E mutation in 68.7% and RET/PTC1 rearrangement in 25.7% of the patients.Although BRAF mutation was not significantly associated with clinicopathologic features such as patient sex or age, multicentric disease, thyroid capsule invasion, tumour stage, or nodal metastasis, it was significantly associated with recurrent disease.Multivariate analysis revealed that BRAF mutation and tumour size were independent risk factors associated with recurrent disease, with odds ratios of 9.072 and 2.387 respectively.The area under the receiver operating characteristic curve increased 8.3% when BRAF mutation was added to the traditional prognostic factors, but that effect was statistically nonsignificant (0.663 vs. 0.746, p = 0.124).RET/ PTC1 rearrangement and nodal metastasis were significantly associated in all patients (p = 0.042), marginally associated in ptc patients (p = 0.051), but not associated in microptc patients (p = 0.700).RET/ PTC1 rearrangement was not significantly associated with recurrent disease.We investigated the associations of BRAF V600E mutation and RET/PTC1 rearrangement with clinicopathologic features and outcomes in Chinese patients with ptc, including microptc.

Patients
The board of medical ethics of Ruijin Hospital, Shanghai Jiaotong University, School of Medicine approved the study, and all patients gave written informed consent.The study enrolled 214 patients with histologically confirmed ptc who underwent surgery in Ruijin Hospital between November 2009 and May 2011.We performed genotyping analysis using frozen ptc tissue.Because of sample limitations, only 101 patients were analyzed for RET/PTC1 rearrangement.Patients were followed twice annually using, as necessary, tests for thyroid-stimulating hormone, thyroglobulin, thyroglobulin antibody; imaging by ultrasonography and computed tomography; and biopsy by fine-needle aspiration.Recurrent disease was defined by histologic or cytologic findings or by clinical appearance on imaging studies.Median follow-up was 36 months (range: 12-50 months), and the follow-up rate was 79.4%.

Detection of BRAF Mutation
Snap-frozen fresh tissues, including both cancerous and adjacent normal thyroid tissue, were collected by experienced pathologists and stored in liquid nitrogen until use.Evaluation of dna extracted using a dna purification kit (Omega Bio-Tek, Norcross, GA, U.S.A.) was performed using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, U.S.A.).The presence of BRAF exon 15 mutation was determined by Sanger sequencing, as previously reported 23 .Polymerase chain reactions were performed in 25 μL buffer containing 1.5 mmol/L MgCl 2 , 200 mmol/L deoxynucleoside triphosphate, 50-100 ng genomic dna, 0.5 mmol/L each primer, and 2.5 U TaKaRa Taq dna polymerase (Takara Bio, Tokyo, Japan).Thirty-five cycles with annealing temperatures optimized at 60°C.Polymerase chain reaction product was purified and sequenced using an ABI Prism 3730 dna analyzer (Applied Biosystems, Foster City, CA, U.S.A.).

Detection of RET/PTC1 Rearrangement
Total rna was isolated using trizol (Invitrogen, Carlsbad, CA, U.S.A.) as described in the manufacturer's protocol.An aliquot containing 1 μg total rna was reverse-transcribed using avian myeloblastosis virus reverse transcriptase (Promega, Madison, WI, U.S.A.), and RET/PTC1 was amplified by nested polymerase chain reaction, separated on 1.2% agarose gel, and visualized under ultraviolet illumination as previously reported 24 .As a positive control, rna from the TPC1 cell line was used.

Statistical Analysis
Statistical analyses were performed using SPSS Statistics (version 17.0: SPSS, Chicago, IL, U.S.A.) and MedCalc (version 13.0: MedCalc Software, Ostend, Belgium).The Fisher exact test was used for nominal variables, and the Mann-Whitney nonparametric test was used for continuous variables.Univariate and multivariate logistic regression were used to determine the risk factors associated with recurrent disease.Odds ratios (ors) were calculated by exponentiation of logistic regression analysis and reported with 95% confidence intervals (cis).To measure the discriminative improvement attributable to the risk score for recurrent disease, we plotted receiver operating characteristic curves for two logistic regression models: Model 1, which included traditional prognostic factors (that is, age, sex, thyroid capsule invasion, multicentricity, TNM stage), and Model 2, which included the traditional factors, plus BRAF mutation.All p values are two-tailed, and p < 0.05 was accepted as statistically significant.

Clinicopathologic Characteristics
As Table i shows, microptc accounted for 30.4% of the tumours in the study group; most patients (65.0%) presented stage i or ii disease.No distant metastases were found at first diagnosis; of the 170 patients who were subsequently followed, 15 (8.8%) experienced recurrent disease.
A somatic BRAF V600E mutation was carried by 68.7% of all study patients, and a somatic RET/PTC1 rearrangement by 25.7% of the 101 whose samples could be tested.Dual mutations (BRAF and RET/ PTC1) were harbored by 15 patients [14.9% (15 of 101)].BRAF mutation was more prevalent in patients with ptc than with microptc (73.2% vs. 58.5%,p = 0.038, Table ii).No BRAF V601 mutation was found.
No BRAF mutation or RET/PTC1 rearrangement was found in the normal tissue adjacent to the ptc tumour.

Correlations Between BRAF Mutation, RET/ PTC1 Rearrangement, and Clinicopathological Characteristics of Patients with PTC
As shown in Table iii, BRAF mutation was significantly associated with Hashimoto disease (p = 0.000), ptc (p = 0.001), and microptc (p = 0.015).BRAF mutation was less prevalent in patients with Hashimoto disease (32.3%) than in those without (74.9%).We found no association between BRAF mutation and sex, age, multicentricity, thyroid capsule invasion, tumour stage, or nodal metastasis.We observed a nonsignificant trend toward positive BRAF mutation status in patients with tumours of advanced TNM stage (p = 0.176).
In tested patients, RET/PTC1 rearrangement was significantly associated with nodal metastasis ( p = 0.042) and marginally with ptc ( p = 0.051), but not with microptc ( p = 0.700).We found no significant association of RET/PTC1 rearrangement with other clinicopathologic features (Table iv).

Prognostic Factors Associated with Recurrent Disease
We found recurrent disease in 15 patients.Tumour size and BRAF mutation status were associated with recurrence.Recurrent tumours were larger (2.0 ± 0.2 cm vs. 1.5 ± 0. We further evaluated the incremental value of BRAF (with respect to traditional prognostic factors) to predict recurrent disease.The area under the receiver operating characteristic curve was calculated for two models: Model 1 included the traditional prognostic factors (age at diagnosis, sex, thyroid capsule invasion, multicentricity, and stage); Model 2 included the Model 1 variables, plus BRAF mutation.As Figure 1 and Table viii show, the area under the curve was 0.663 (95% ci: 0.529 to 0.797) for Model 1 and 0.746 (95% ci: 0.631 to 0.860) for Model 2-achieving a nonsignificant increase of 8.3% with the addition of BRAF mutation ( p = 0.124).

DISCUSSION
In the present study, we genotyped ptc tumours and investigated the associations of mutations with clinicopathologic features and clinical outcome.We found that BRAF mutation was an independent predictor of recurrent disease, although it not associated with advanced clinicopathologic features; on the other hand, RET/PTC1 rearrangement was associated with nodal disease, but not with recurrent disease.
Although most ptc patients experience favourable outcomes, significant numbers of patients develop recurrent disease and experience poor outcomes.To improve clinical outcomes, disease staging systems have been established to stratify management strategies 25 .Large tumour size, older age, extrathyroidal invasion, male sex, multicentricity, distant metastasis, and lymph node metastasis are the main determinants of a poor outcome in ptc patients 6,[25][26][27] .Currently, the need for preoperative risk stratification and for the enlarged "grey zone" of microptc require an optimized risk scoring system.Genetic markers such as BRAF mutation and RET/PTC rearrangements seem to be able to fill the gap.Positivity for BRAF mutation has been associated with a negative prognosis and poor clinical outcome.However, conflicting data have also been reported.The clinical significance of BRAF analysis in ptc patients has been disputed because the association of positivity with poor outcome was believed to depend on the mutation's association with aggressive tumour behavior.Our data provide an important piece of evidence to help resolve the dispute.We did not find a significant association of BRAF mutation with negative prognostic indicators such as large tumour size, old age, extrathyroidal invasion, male sex, multicentricity, lymph node metastasis, or TNM stage.However, BRAF mutation and tumour size were both associated with recurrent disease (although not with tumour stage and extent of surgery).Compared with patients lacking a BRAF mutation, those with the mutation were more likely (by a factor of 9) to experience recurrent disease.BRAF analysis provided an 8.3% increment over traditional predictors for recurrent disease, but that increase was not statistically significant ( p = 0.124).Our data confirm the clinical significance of BRAF analysis in ptc patients and suggest a need for more extensive surgery and for more aggressive postsurgical management such as radioactive iodine ablation, suppression of thyroid-stimulating hormone below 0.1 mU/L, and closer follow-up in somatic BRAF carriers.However, knowing a patient's BRAF status did not substantially improve prediction overall, probably because of the limited number of recurrence events.The lack of any associations between BRAF mutation and advanced clinicopathologic features might have several explanations.First, our cohort came from a single centre serving patients from Shanghai and its surrounding area, where iodine is sufficient.Guan et al. 28 reported that high iodine intake is a significant risk factor for BRAF mutation, and a meta-analysis showed that a high prevalence of the BRAF V600E mutation tended to have a smaller meta-risk of extrathyroidal invasion and lymph node metastasis 7 .Second, a significant number of our patients underwent subtotal thyroidectomy or lobectomy, which might lead to an underestimation of aggressive pathologic characteristics 21,29 .Consistent with an earlier study 7 , BRAF mutation was confirmed as an independent predictor of recurrence in our cohort at 3 years of follow-up.
Our data indicate the importance of performing follow-up studies.
Other important findings of our study include the observation that, in addition to BRAF mutation, tumour size is another independent predictor of recurrent ptc.Also, BRAF V600E was, in general, less prevalent in microptc than in ptc, which is consistent with most earlier studies 7,13 .In addition, BRAF mutation was inversely associated with Hashimoto disease, which has also been reported in a Korean cohort 30 .This inverse correlation suggests that BRAF mutation and Hashimoto disease are exclusively involved in ptc initiation.Moreover, where dual mutation of BRAF and RET/ PTC1 occurred (15 patients), 16.7% experienced recurrent disease, compared with the 11.8% of patients harbouring solely a BRAF mutation ( p = 0.664, Table ix).That observation is consistent with an earlier finding that patients with dual mutation are more susceptible to recurrent disease 31 .
The limitations of our study include its lack of multicentre data and long-term follow-up.The limitation resulting from extent of surgery has already been discussed.Papillary thyroid carcinoma is slow-growing.Long-term follow-up, including data on both recurrence and mortality, is required to comprehensively elucidate the role of BRAF mutation analysis in the management of this most frequent thyroid cancer.

CONCLUSIONS
We analyzed both BRAF mutation and RET/PTC1 rearrangement in 214 Chinese ptc patients with a median follow-up of 36 months.We found that BRAF mutation was an independent predictor of recurrent disease despite a lack of association with advanced clinicopathologic features.Our data indicate that BRAF mutation is not just a marker of aggressiveness, but a true prognostic factor.
Although the increase was nonsignificant, BRAF mutation analysis added prognostic value over that of traditional predictors.In accord with previous studies, we suggest that BRAF mutation analysis has a positive role to play in the presurgical assessment of ptc patients.However, larger series and longerterm follow-up data are required.

figure 1
figure 1 Predictability of recurrent papillary thyroid carcinoma using combined prediction models.Model 1 included traditional prognostic factors (age, sex, thyroid capsule invasion, multicentricity, TNM stage), and Model 2 included traditional factors, plus BRAF mutation.The area under the curve was 0.663 (95% confidence interval: 0.529 to 0.797) for Model 1, and 0.746 (95% confidence interval: 0.631 to 0.860) for Model 2. The addition of BRAF mutation to traditional prognostic factors increased the predictability by 8.3%, but without statistical significance (p = 0.124).

e741 Current OnCOlOgy-VOlume 21, number 6, DeCember 2014
Copyright © 2014 Multimed Inc.Following publication in Current Oncology, the full text of each article is available immediately and archived in PubMed Central (PMC).
1 cm, p = 0.017) and more likely to harbour BRAF mutation (93.3% vs. 64.5%,p = 0.022, Table v).No association of recurrent disease with Hashimoto disease, TNM stage, type of surgery, or RET/PTC1 rearrangement was found.Multivariate analysis confirmed that tumour size and BRAF mutation were independent prognostic factors associated with recurrent ptc (Table vi).
Compared with patients whose tumours lacked the somatic BRAF V600E mutation, patients with tumours harbouring the mutation had a risk of recurrent disease that was increased by a factor of 9 (or: 9.072; 95% ci: 1.072 to 76.739; p = 0.043).Tableviishows correlations between BRAF mutation and other clinicopathologic variables.
Because of sample limitations, only 101 analyses for RET/PTC1 rearrangement were performed.c Only 170 patients were followed.
a Using the American Joint Committee on Cancer's TNM classification system for differentiated thyroid carcinoma.baTumoursgreater than 1 cm in size.bTumours 1 cm or less in size.e743CurrentOnCOlOgy-VOlume21, number 6, DeCember 2014Copyright © 2014 Multimed Inc.Following publication in Current Oncology, the full text of each article is available immediately and archived in PubMed Central (PMC).

table iii
Correlations between clinicopathologic features and BRAF mutation by papillary thyroid carcinoma (ptc) type a Tumours greater than 1 cm in size.bTumours 1 cm or less in size.cUsing the American Joint Committee on Cancer's TNM classification system for differentiated thyroid carcinoma.Current OnCOlOgy-VOlume 21, number 6, DeCember 2014Copyright © 2014 Multimed Inc.Following publication in Current Oncology, the full text of each article is available immediately and archived in PubMed Central (PMC).

table iv
Correlations between clinicopathologic features and RET/PTC1 rearrangement by papillary thyroid carcinoma (ptc) type Using the American Joint Committee on Cancer's TNM classification system for differentiated thyroid carcinoma.
a Tumours greater than 1 cm in size.b Tumours 1 cm or less in size.c

VOlume 21, number 6, DeCember 2014
Copyright © 2014 Multimed Inc.Following publication in Current Oncology, the full text of each article is available immediately and archived in PubMed Central (PMC).
Univariate and multivariate analysis of clinicopathologic features and recurrence-free survival in papillary thyroid carcinoma (ptc) a Using the American Joint Committee on Cancer's TNM classification system for differentiated thyroid carcinoma.e746 Current OnCOlOgy-VOlume 21, number 6, DeCember 2014 Copyright © 2014 Multimed Inc.Following publication in Current Oncology, the full text of each article is available immediately and archived in PubMed Central (PMC).table vi a Spearman rho.

table viii
Predictability of recurrent papillary thyroid carcinoma, measured by area under the curve (auc)