Ultra-sensitive circulating tumor DNA (ctDNA) detection as predictor of survival outcomes in endometrial cancer patients undergoing frontline treatment.

5611 Background: For endometrial cancer (EC), clinicopathologic and molecular characterization add to the risk stratification complexity, but remain limited to predict risk of recurrence. Precise biomarkers are needed to guide treatment. This study seeks to investigate the value of circulating tumor DNA (ctDNA) in EC patients. Methods: In an ongoing, single-institution prospective study of EC patients undergoing primary surgical management, serial plasma samples were collected at multiple timepoints: pre-(pre-op)/post-operatively (post-op), post-definitive treatment (dTx), and surveillance. If patients did not have adjuvant therapy, the post-op time point was considered the dTx timepoint. ctDNA analysis was performed on banked samples at the non-surveillance timepoints using NeXT Personal, a personalized whole genome sequencing-based tumor-informed assay tracking up to 1800 mutations, (Personalis, Fremont, CA). Primary endpoint was ctDNA detection rate. Secondary endpoints included correlation with survival outcomes (recurrence free survival [RFS] and overall survival [OS]). Multivariable analysis was performed for RFS adjusting for dTX ctDNA detection, stage, and TP53 mutation status. Given the limited number of deaths, multivariable analysis was not performed for OS. Results: 99 patients and 254 samples were included for ctDNA analysis. Median age at surgery was 65.1 (range 42-85) years. After a median follow-up of 24.1 (range 3.4 – 67) months, there were 25 recurrences and 10 deaths. Majority had stage I (68.7%), grade 3 (73.7%) tumors, non-endometrioid histologic tumor components (63.7%), and received adjuvant therapy (79.8%; therapy declined in 4 [4%] patients). Adjuvant radiation and chemotherapy was given in 67 (67.7%) and 53 (53.5%) patients, respectively. ctDNA was detected in 83.3% (75/90) of pre-op, 19.5% (17/87) of post-op, and 14.1% (9/64) of dTx samples. Although pre-op ctDNA was not associated with survival, post-op ctDNA detection was associated with worse RFS (HR 6.3, 95% CI 2.6-15.5; p<0.001) and OS (HR 24.6, 95% CI 3-205; p=0.003). Post-op detection in the ultrasensitive range (<100 PPM) were also significantly prognostic of worse RFS (HR: 5.1, 95% CI 1.3-19.5, P=0.02). Similarly, dTx detection was associated with worse RFS (HR 27.7, 95% CI 8.9-86.6; p<0.001) and OS (HR 7.5, 95% CI 1.7-33.7; p=0.01). On multivariable analysis, dTx detection was independently associated with worse RFS (HR 27.7, 95% CI: 8.9-86.6; p<0.001). Conclusions: In a predominantly early-stage, high-risk cohort of EC patients, post-op and dTx ctDNA detection is associated with inferior survival. ctDNA detection in the ultrasensitive range may hold promise to identify EC patients at risk for recurrence which may not be identified in conventional assays. These results support ctDNA as a promising biomarker for clinical decision-making in EC.