A pilot study evaluating somatic mutational burden and signatures among endometrial cancer patients: Is there prognostic benefit to whole genome sequencing?

Abstract

e17629 Background: Whole genome sequencing (WGS) has emerged as a tool to characterize mutational burden and potentially identify new therapeutic targets and predictive biomarkers. In this study, we describe the mutational landscape of diverse patients with endometrial cancer using WGS and explore associations between mutations and clinicopathologic characteristics. Methods: Patients with endometrial cancer were prospectively consented and WGS was performed on tumors and blood. Using the the New York Genome Center Cancer pipeline, we identified high confidence somatic mutations in our patient cohort (identified by 2 or more variant callers). Somatic variants were categorized into mutational signatures based on the Catalog of Somatic Mutations in Cancer (COSMIC) v2. Associations between COSMIC signatures and clinicopathologic variables were evaluated using ANOVA tests. Results: Sixteen patients with endometrial cancer were enrolled; 10 (62.5%) self-identified as White, 4 (25.0%) as Black, and 2 (12.5%) as Asian. Tumor histology was endometrioid (13, 81.3%) or serous/clear cell (2, 12.5%), with both low grade (9, 56.3%) and high grade (7, 43.8%) tumors. The most frequently mutated genes were PTEN (15, 93.8%), ARID1A (9, 56.3%), PIK3CA (9, 56.3%), and TP53 (6, 37.5%). PIK3CA mutations were found in all mismatch repair (MMR) deficient tumors (5), compared to only 30.0% (3) of patients with intact MMR mechanisms (FDR p = 0.0256). Signature 3, associated with BRCA mutations and loss of DNA double-strand break-repair by homologous recombination, was enriched in p53 mutated tumors (p = 0.0036). Signature 1, associated with age, was enriched in patients with MMR intact tumors (p = 0.0475), while signatures 6 and 20, associated with defective DNA MMR and microsatellite unstable tumors, were enriched in patients with in MMR deficient tumors (p < 0.0001 and p = 0.0022, respectively). Signature 8, of unknown etiology, was enriched in patients with MMR intact tumors, advanced disease (p = 0.0045), disease spread to lymph nodes (p = 0.0231), and disease recurrence (p = 0.0021). Signature 12, also of unknown etiology, was enriched in MMR deficient tumors (p = 0.0005) and stage 1 disease (p = 0.0235). Conclusions: WGS revealed distinct patterns of mutational burden associated with clinicopathologic variables. MMR status was associated with PIK3CA mutations and multiple mutational signatures. Mutational signature 8 was associated with multiple clinicopathologic characteristics associated with poor prognosis, suggesting utility as a biomarker. Further investigation regarding the mechanism of the mutational signatures and correlation with prognostic factors may shed light on disparate biologic endometrial cancer outcomes and serve as a source of hypothesis-generation for the development of targeted therapies.