4 Discussion

In recent years, researches designed to better understand the significance of BRCA mutations are changing clinical practice, especially in the domain of ovarian cancer treatment. There were success in several key clinical trials, including SOLO-1¹⁹, PAOLA,²⁰ PRIMA,²¹ etc., which facilitated the National Comprehensive Cancer Network (NCCN) recommendation of PARP inhibitors as first-line maintenance therapy in epithelial ovarian cancer, and also urged the delivery of BRCA testing as a common practice.²² In EC, however, we have limited understanding of BRCA mutational status since few research evidences supporting related clinical applications. In this study, we have proved the role of BRCA mutations in predicting EC patients’ survival. Our results demonstrate that high-grade endometrioid EC patients with BRCA1/2 mutations have longer survival than those with wild-type BRCA, though the time to recurrence is not significantly affected.

Previous work suggested that patients with germline BRCA1 mutations were linked to an increased risk of developing serous or serous-like EC.²³ A recent study carried out by de Jonge et al.¹⁴ further proved that ECs developed in the context of germline BRCA mutation and subsequent loss-of-heterozygosity were more commonly of non-endometrioid type. Our study, however, showed that in EC cohort with BRCA1/2 somatic mutations, there were significantly more endometrioid cases (88.9% vs. 72.4%, p = 0.001). In addition, the distribution of histological types was similar among patients with BRCA1, BRCA2 or both mutations (p = 0.181). These results suggested the possible differences in tumorigenesis-related mechanisms after germline or somatic BRCA mutations. Besides, based on the study results of de Jonge et al.,¹⁴ ECs with germline BRCA mutation and loss-of-heterozygosity showed more adverse prognostic factors compared with those without loss-of-heterozygosity, indicating unfavorable survival. But in another recent study by Kadan et al.,²⁴ serous ECs with BRCA mutations showed similar survival with BRCA wild-type ones, which was in accordance with our data (p = 0.115 and 0.226 for OS and RFS in non-endometrioid EC).

According to previous studies, BRCA1/2 mutations were associated with homologous recombination defects and increased mutation burden in the genome.²⁵ And higher tumor mutation burden was proved to be associated with better survival outcome in multiple cancer types from direct and indirect evidences.^6,26–28 These results might partly explain why high-grade endometrioid EC patients with BRCA mutations exhibited better survival than BRCA wild-type ones. While for non-endometrioid type EC, more invasive biological behavior endowed it generally poor survival compared with endometrioid ones, in which context, the positive survival impact of BRCA mutation might be not prominent enough to show a statistically significant effect. But still, more studies were needed to validate the above proposed mechanism.

In our study, detailed clinicopathological features of patients with distinct BRCA mutational status were compared, and stratification of patients according to clinicopathological features was performed. But there were also several limitations. Firstly, as a retrospective study, the capacity of controlling inter-group confounders was limited, though statistical method, such as PSM, was used. Besides, though the rates of receiving radio-, chemo- and molecular targeted therapies were basically comparable between the groups, detailed information about adjuvant therapy use was lacking in TCGA database. Secondly, in the study, since the number of patients with BRCA1 mutations was limited, we did not manage to compare the survival influence of BRCA1 and BRCA2 mutations and discuss their roles separately. Larger cohorts may be helpful for more detailed comparison and better robustness of the results. Thirdly, since germline and somatic BRCA mutations may be distinct in their prognostic roles, it is better to set another cohort of germline BRCA mutant patients for further comparisons, which was not achieved in our study.

Currently, the trend toward individualized tumor management based on both clinicopathological features and molecular profiles is coming to the mainstream of clinical practice in gynecologic oncology treatment, and gene testing is getting increasing impact in revealing certain prognostic features and suggesting specific treatment modalities. One example is the currently ongoing PORTEC-4a trial, which combined EC patients’ molecular and clinicopathological features in the clinical-decision system for choosing radiation therapy mode.²⁹ Our results about BRCA mutation provide another promising biomarker for prognostic prediction in the subgroup of high-grade endometrioid EC patients, and might be helpful for future molecular profile based immuno- and targeted therapy strategy planning. Nonetheless, further studies are needed in the future to fully validate the clinical significance of different BRCA mutation types, and more genetic researches might be helpful for better understanding the underlying mechanisms so as to better guide clinical practice.