Exploring associations between tumor mutational burden and survival in uterine cancers (220)

Abstract

<b>Objectives:</b> Molecular profiling of tumors from patients with uterine cancer has led to new therapeutic options with the use of immunotherapy. The FDA approved pembrolizumab for patients with tumor mutational burden-high (TMB-H) (≥10 mutations/megabase [mut/Mb]) solid tumors. We characterized patients with uterine cancer who had undergone next-generation sequencing to describe tumor mutational burden, microsatellite instability, and clinical outcomes within different epithelial and molecular subtypes. <b>Methods:</b> A retrospective cohort study was performed on patients with uterine cancer who underwent Foundation One testing between 2015 and 2021 at a single academic institution. Demographic and clinical characteristics were identified, including age, race/ethnicity, stage, histology, molecular subtype, using The Cancer Genome Atlas molecular classification, microsatellite instability (MSI) status, TMB, and follow-up time in months. Kaplan-Meier curves and Cox regression analyses were used to estimate overall survival (OS) and progression-free survival (PFS) based on TMB, MSI, and molecular subtypes. <b>Results:</b> We identified 129 patients with uterine cancer with Foundation One testing results: 54 (41.9%) had endometrioid histology, 36 (27.9%) serous, six (4.7%) clear cell, 17 (13.2%) carcinosarcoma, five (3.9%) endometrial stromal sarcoma, five (3.9%) leiomyosarcoma, and six others (4.7%). The median age at diagnosis was 62 years. TMB was available for 106/129 patients: 71 (67.0%) had low TMB, 33 (31.1%) had intermediate, and two (1.9%) had high. TMB scores were higher among those with MSI (mean, 13.5±5.2) compared to those with microsatellite stable tumors (mean, 4.5±2.6; p<0.001). Mean TMB was highest among those with the MSI hypermutated molecular subtype (median 10, mean 31.7±52.4), followed by <i>POLE</i> ultra-mutated (median 10, mean 10.1±6.3), copy-number low (median 4, mean 5.1±3.6), and copy-number high (median 4, mean 4.0±1.7). Patients with MSI-high tumors had statistically significantly higher TMB compared to patients with copy-number high tumors (p=0.001). There was also a statistically significant difference between TMB from patients with <i>POLE</i> hypermutated tumors those with copy-number high (p<0.001) and those with copy-number low tumors (p=0.004). TMB was not an independent predictor of OS (p=0.61) or PFS (p=0.07) for the entire cohort, or for any particular molecular subtype. Among patients who developed recurrent uterine cancer, there was a significant difference in PFS (p=0.002) and OS (p=0.002) by molecular classification (Figure 1). <b>Conclusions:</b> TMB was not an independent predictor of PFS or OS for patients with primary or recurrent uterine cancer. There are significant differences in TMB in the TCGA molecular subgroups of uterine cancer. Further work is needed to determine how TMB within the various molecular subgroups of uterine cancers predicts response to therapy.