TERT promoter Mutation and Its Association with Clinicopathological Features and Prognosis of Papillary Thyroid Cancer: A Meta-analysis

We performed a meta-analysis to elucidate the associations of the clinicopathological characteristics and prognostic factors of papillary thyroid cancer (PTC) with TERT promoter mutations. A literature search was performed of the PubMed and EMBASE databases using Medical Subject Headings and keywords. Individual study-specific odds ratios (ORs) and confidence intervals (CIs) were calculated. The average prevalence rate of TERT promoter mutations was 10.1%. TERT promoter mutations occurred more frequently in patients with larger tumors (p = 0.003). TERT promoter mutations were associated with advanced stage (OR = 3.11, 95% CI = 2.22–4.36), lymph node metastasis (OR = 1.82, 95% CI = 1.12–2.96), distant metastasis (OR = 4.18, 95% CI = 1.61–10.81), BRAF mutation positivity (OR = 2.71, 95% CI = 1.45–3.24), recurrence (OR = 3.91, 95% CI = 1.83–8.34), and mortality (OR = 8.13, 95% CI = 3.77–17.53). The associations of TERT promoter mutations with extrathyroidal invasion (OR = 1.98, 95% CI = 0.96–4.07), unifocality (OR = 1.36, 95% CI = 0.90–2.07), and vascular invasion (OR = 1.45, 95% CI = 0.92–2.30) were not significant. TERT promoter mutations are closely associated with aggressive clinicopathological characteristics and poorer prognosis in PTC.


Meta-analysis of the Effects of TERT promoter mutations on Clinicopathological Features.
In each category of prognostic factors, the pooled ORs were higher in patients with TERT promoter mutations than in those with the wild-type gene. Six studies presented clinical data including extrathyroidal invasion. Extrathyroidal invasion was present in 70 (47.3%) of 148 patients with TERT promoter mutations and in 456 (34.9%) of 1306 patients without TERT promoter mutations ( Fig. 2A). The average odds ratio (OR) from the six studies was 1.98 (95% confidence interval [CI] = 0.96-4.07). The heterogeneity of the data was significant (P = 0.006), and the I 2 estimate of the variance between the studies was 69%. According to our analysis, the difference between the occurrence of extrathyroidal invasion according to TERT promoter mutation positivity was not significant (P = 0.06).
We analyzed three studies that reported tumor size (Fig. 2B). A fixed-effects model was adopted, as heterogeneity was not significant between tumor size and TERT promoter mutations (P = 0.90, I 2 = 0%). According to our analysis, tumor size was larger in patients with TERT promoter mutations than in those without TERT promoter mutations. The mean difference was 0.74 cm (95% CI = 0.26-1.23, P = 0.003).
TNM stage was reported for patients in seven studies (Fig. 3A). Advanced TNM stage (III/IV) was identified in 82 (48.8%) of 168 patients with TERT promoter mutations and in 364 (23.0%) of 1582 patients without TERT promoter mutations. A fixed-effects model was adopted because heterogeneity was not significant between TNM stage and TERT promoter mutations (P = 0.47, I 2 = 0%). According to our analysis, advanced TNM stage (III/IV) occurred more frequently in patients with TERT promoter mutations than in those without TERT promoter mutations. The overall OR was 3.11 (95% CI = 2.22-4.36, P < 0.00001). Regarding patients with unifocality (Fig. 3B), four studies were included in the meta-analysis. Unifocality was present in 71 (65.1%) of 109 patients with TERT promoter mutations and in 627 (58.2%) of 1078 patients without TERT promoter mutations. The overall OR was 1.36 (95% CI = 0.90-2.07). A fixed-effects model was adopted, as heterogeneity was not significant between multifocality and TERT promoter mutations (P = 0.50, I 2 = 0%). According to our results, there was no significant association between TERT promoter mutations and tumor foci (P = 0.15).
Vascular invasion was reported in four studies (Fig. 4A). Vascular invasion was present in 33 (33%) of 100 patients with TERT promoter mutations and in 232 (26.2%) of 885 patients without TERT promoter mutations.  A fixed-effects model was adopted, as heterogeneity was not significant between vascular invasion and TERT promoter mutations (P = 0.37, I 2 = 4%). According to our results, the difference in the occurrence of vascular invasion according to TERT promoter mutation positivity was not significant. The overall OR was 1.45 (95% CI = 0.92-2.30, P = 0.11).
Regarding cases of lymph node metastasis ( Fig. 4B), nine studies were included in the meta-analysis. Lymph node metastasis was present in 115 (57.8%) of 199 patients with TERT promoter mutations and in 869 (47.0%) of 1849 patients without TERT promoter mutations. The overall OR was 1.82 (95% CI, 1.12-2.96). A random-effects model was adopted because the heterogeneity of the data was significant (P = 0.05), and the I 2 estimate of the variance between the studies was 48%. According to our analysis, lymph node metastasis occurred more frequently in patients with TERT promoter mutations than in those without TERT promoter mutations (P = 0.02).  Concerning the presence of distant metastasis (Fig. 5A), four studies were included in the meta-analysis. Distant metastasis was present in 31 (28.4%) of 109 patients with TERT promoter mutations and in 81 (9.1%) of 886 patients without TERT promoter mutations. The overall OR was 4.18 (95% CI, 1.61-10.81). A random-effects model was adopted because the heterogeneity of the data was significant (P = 0.06), and the I 2 estimate of the variance between the studies was 59%. According to our analysis, distant metastasis occurred more frequently in patients with TERT promoter mutations than in those without TERT promoter mutations (P = 0.003).
Regarding patients with BRAF mutations (Fig. 5B), five studies were included in our meta-analysis. BRAF mutation positivity was present in 72 (62.1%) of 116 patients with TERT promoter mutations and in 547 (47.7%) of 1147 patients without TERT promoter mutations. The overall OR was 2.17 (95% CI = 1.45-3.24). A fixed-effects model was adopted because the heterogeneity of the data was not significant (P = 0.61), and the I 2 estimate of the variance between the studies was 0%. According to our analysis, BRAF mutations occurred more frequently in patients with TERT promoter mutations than in those without TERT promoter mutations (P = 0.0002).

Meta-analysis of the Effect of TERT Mutation on Persistent/Recurrence and Mortality.
Concerning persistence/recurrence, five studies were included in this meta-analysis. Persistence/recurrence occurred in 58 (46.4%) of 125 patients with TERT promoter mutations and in 164 (14.8%) of 1109 patients without TERT promoter mutations. A random-effects model was adopted because the heterogeneity of the data was significant (P = 0.05), and the I 2 estimate of the variance between the studies was 57%. According to our analysis, persistent/recurrence was more frequent in patients with TERT promoter mutations than in those without mutations. The overall OR was 3.91 (95% CI = 1.83-8.34), and the P-value was 0.0004 (Fig. 6A). Furthermore, individual studies were assessed further by sequentially excluding each study from our meta-analysis. In this manner, we found that I 2 was 0% when we excluded the study by Qasem, but this did not occur when other articles were excluded; thus, we concluded that the heterogeneity was mainly caused by this particular study.
Mortality data were reported in three studies (Fig. 6B). Mortality occurred in 25 (45.5%) of 55 patients with TERT promoter mutations and in 31 (7.6%) of 406 patients without TERT promoter mutations. A fixed-effects model was adopted, as heterogeneity was not significant between mortality and TERT promoter mutations (P = 0.39, I 2 = 0%). According to our analysis, mortality was more frequent among patients with TERT promoter mutations than among those without mutations. The overall OR was 8.13 (95% CI = 3.77-17.53, P < 0.00001).

Subgroup Analyses of the Effects of TERT promoter mutations on Aggressive Clinicopathological
Features and Prognostic Factors. Subgroup analysis was conducted according to the patients' country of origin to investigate potential sources of heterogeneity and to assess whether the effects of TERT promoter mutations on aggressive clinicopathological features and poor prognosis of PTC were associated with geographic regions ( Table 3). The effect estimates were broadly consistent among the subgroups that were analyzed. Heterogeneity was markedly decreased in the subgroup analyses of extrathyroidal invasion, distant metastasis, and recurrence. However, this may have been caused by two Asian countries (Saudi Arabia and Korea) being distantly located geographically and being different in many aspects. The two studies involved were among the smallest in our meta-analysis, and the total number of cases and relatively low prevalence of TERT promoter mutations led to a lack of statistical power. The results of the subgroup analyses also indicated that TERT promoter mutations were not significantly associated with some high-risk tumor features (e.g., extrathyroidal extension and lymph node metastasis) of PTC in patients from Asia.

Discussion
To some extent, molecular alterations such as genetic and epigenetic alterations in signaling pathways have changed the treatment of thyroid cancer through the development of targeted therapies. Many somatic genetic alterations including those in BRAF, HRAS, KRAS, NRAS, PTEN, and HER1 have been revealed to play fundamental roles in the tumorigenesis of thyroid carcinoma. The close association of TERT promoter somatic mutations with tumorigenesis is widely recognized. TERT expression confers infinite proliferation potential and many other biological activities on cancer cells by maintaining telomere length [16][17][18] . In addition, recent data also demonstrated that the functional TERT rs2736100 SNP as a novel genetic component of PTC etiology 19 .
TERT promoter mutations, which might contribute to telomerase activation, have been reported in various cancers including melanoma, anaplastic oligodendroglioma, and bladder cancer [20][21][22][23][24] . Recently, the presence of TERT promoter somatic mutations has been reported in PTC; however, their diagnostic and prognostic significance remains unclear 7,8 . The present study aimed to determine the influence of TERT promoter somatic mutations on the clinicopathological outcomes and prognosis of PTC via a meta-analysis.
Larger tumor size, extrathyroidal invasion, stage III or IV disease, lymph node metastasis, and distant metastasis were correlated with poor prognostic features such as recurrence and short overall survival (OS) 15,25 . Several authors suggested that TERT promoter somatic mutations were associated with these aggressive clinicopathological characteristics 7,14 . This view is mainly in line with the results of our meta-analysis. Our findings illustrated that TERT promoter mutations were more likely to be present in patients with larger tumors, stage III or IV disease, lymph node metastasis, and distant metastasis. However, extrathyroidal invasion was at a critical level in terms of an association with TERT promoter mutations (P = 0.06).
One interesting finding in the current study was that the associations of TERT promoter mutations with multifocality and vascular invasion were not significant. This finding may be explained by the fact that the studies including data on focus number and vascular invasion were relatively small.
Most authors previously demonstrated that recurrence and poor OS were more likely to occur in patients with TERT promoter mutations than in those without mutations. By contrast, Qasem et al. found that the relationships of recurrence and short OS with TERT promoter mutations were not significant 14 . In our meta-analysis, however, patients with PTCs with TERT promoter mutations were more likely to experience recurrence and worse OS. Therefore, we conclude that TERT promoter mutations may represent an independent prognostic factor for prognosis in PTC.
A classic alteration found in PTC is the V600E mutation in the serine/threonine kinase BRAF 9,26 . Multicenter studies confirmed the association of BRAF V600E with poor clinicopathological outcomes such as aggressive pathological features, increased recurrence, and resistance to radioiodine therapy in PTC 27 . Recently, a positive correlation between TERT promoter mutations and the BRAF V600E mutation was reported in advanced PTCs, suggesting functional cooperation, and Xing et al. suggested that patients with PTC and coexistent TERT promoter and BRAF V600E mutations tend to have poorer prognosis compared with patients harboring single mutations or wild-type TERT promoter and BRAF 7 . However, evidence reported by Biase et al. illustrated that TERT promoter mutations do not predict unfavorable features in papillary thyroid microcarcinoma, either alone or in combination with the BRAF V600E mutation 8 , and Melo et al. found no differences in persistence between tumors with TERT promoter mutations alone and those harboring both TERT and BRAF mutations 10 . Furthermore, Gandolfi et al. observed a positive association between BRAF V600E and TERT C228T mutations in a cohort of patients with PTC and distant metastases 11 . However, the associations of BRAF V600E with TERT C250T and all TERT promoter mutations were not significant. Our meta-analysis illustrated that TERT promoter mutations occurred more frequently in patients with BRAF mutations. Whether the coexistence of these two mutations predicts a poor prognosis is unclear and needs further investigation.
Our systematic review and meta-analysis assessed the associations of TERT promoter mutations with clinicopathological outcomes and prognosis in PTC. Although our study cannot identify a causal relationship between TERT promoter mutations and outcomes and TERT promoter mutations may be correlated with some other mutation or some confounder that may be even more useful, our findings may still be of great value in evidence-based clinical decision-making. In other words, our analysis should be interpreted with caution because of the heterogeneity of the data. Possible explanations for this heterogeneity include patient demographics as well as thyroidectomy, approaches to lymph node dissection, radioactive iodine (RAI) treatment, and the duration of follow-up. The population included patients from the USA, Portugal, Italy, and China, among other countries. Details of the approaches used for thyroidectomy, lymph node dissection, and RAI treatment were not all provided; furthermore, the outcomes might be interrelated. For instance, patients with more advanced disease tend  to have extrathyroidal invasion, and thus, disease stage may confound the association between TERT promoter mutations and extrathyroidal invasion. Apart from potential heterogeneity caused by varied management practices, differences in pathology reporting, follow-up duration, and the definition of remission may affect the conclusions of this meta-analysis. Another limitation is that the number of included articles is relatively small, and relevant unpublished data could not be obtained for further analysis. Therefore, our conclusions should be used cautiously. Larger studies would be helpful to definitely address the role of TERT promoter mutations in the prognosis of PTC. Furthermore, we did not distinguish between the C228T and C250T mutations from collective mutations in some studies 9,11 , and in the study by de Biase et al. 8 , only microcarcinomas were analyzed, which may have interfered with our results.

Conclusion
Our meta-analysis demonstrated that TERT promoter mutations are closely associated with aggressive clinicopathological factors and poorer prognosis in PTC. TERT promoter mutations may be considered poor diagnostic and prognostic markers in PTC, and patients with such mutations may require aggressive management. More studies are needed to better evaluate the role of TERT promoter mutations in PTC.

Materials and Methods
Search strategy and literature selection. A systematical literature search was performed using online electronic databases (PubMed, EMBASE, and ISI Web of Science) for published papers through November 2015, and the search was supplemented by manual searching and reference backtracking using the following Medical Subject Headings and keywords: "TERT", " telomerase reverse transcriptase", "mutation", "thyroid", "neoplasm(s)", "tumor", "cancer", and "carcinoma". The searches were limited to studies conducted in humans and written in English. Furthermore, the reference lists of retrieved articles were also reviewed to identify additional studies. Relevant unpublished data that were presented at international meetings such as the American Thyroid Association meeting were also included. We contacted the authors for additional tabular data when necessary 13 .
The following criteria were applied into the literature selection for studies that examined the association of TERT promoter mutations with high-risk clinic pathological factors and prognostic outcomes: (1) the studies had a randomized controlled trial or retrospective comparative study (cohort or case-control study) design; (2) the studies compared the TERT mutation-positive and TERT mutation-negative groups for patients with primary PTC; (3) the study investigated at least one outcome of interest; and (4) weighted mean differences (WMDs) and ORs with 95% CIs were reported or were available to be calculated. Studies lacking a control population, duplicates of previous publication, animal studies, abstracts, single-case reports, reviews, and unpublished reports were excluded. Additionally, studies that included TERT mutation analysis for patients with preoperative fine-needle aspiration biopsies were excluded to avoid false-negative results for the cytologic specimens. For studies with the same or overlapping data published by the same investigators, we selected studies with complete designs and larger sample sizes in our meta-analysis. Data extraction. All data were extracted independently by two authors (CP Liu and ZM Liu) and cross-checked to resolve any discrepancies. The following information was extracted from each included study regarding the association of TERT promoter mutations with clinical and pathological risk factors of PTC: first author, publication year, study location, ethnicity, number of cases, the percentage of females, age in both the TERT mutation-positive and TERT mutation-negative groups, TERT mutation rate, and the incidence rate of clinicopathological features in both the TERT mutation-positive and TERT mutation-negative groups. Regarding the associations of TERT promoter mutations with recurrent/persistent disease and mortality, the following data were extracted: first author, publication year, study location, number of follow-up cases, TERT mutation rate, stage of disease, initial treatment, poorer outcome, and the incidence rates of persistent/recurrence and mortality in both the TERT mutation-positive and TERT mutation-negative groups.
The following outcomes were extracted to compare the TERT mutation-positive and TERT mutation-negative groups for patients with primary PTC: the presence of extrathyroidal invasion, larger tumor size, advanced TNM stage (stage III or IV), unifocality, vascular invasion, lymph node metastasis, distant metastasis, BRAF mutation positivity, persistent/recurrent disease, and cancer-specific mortality. All of the procedures conformed to the guidelines for the meta-analysis of observational studies in epidemiology 28 .
Statistical Analysis. The summary ORs with 95% CIs and WMDs with 95% CIs were calculated to compare dichotomous and continuous variables, respectively. The χ 2 -based Cochran's Q statistic test and I 2 statistics were used to evaluate heterogeneity between the studies, and heterogeneity was considered significant when P < 0.1 for the Q statistic or for an I 2 statistic > 50% 29 . A fixed-effects model (Mantel-Haenszel method) was used when no significant heterogeneity was detected; otherwise, a random-effects model (DerSimonian-Laird method) was applied. Subgroup analyses were also performed according to geographic region. In addition, we performed sensitivity analysis to assess the influence of each study on the overall estimate. Moreover, the potential publication bias was assessed using Egger's test and funnel plot analysis. All meta-analyses were performed using Review Manager (version 5, Cochrane Collaboration, Oxford, UK). The P-values were two-tailed with a level of significant at 0.05.