Evaluation of prognostic and predictive values of bespoke circulating tumor DNA assay in patients with advanced epithelial ovarian cancer.

5570 Background: Despite the standard-of-care (SOC) treatment series, about 90% of EOC eventually relapse. Monitoring molecular residual disease (MRD) with high sensitivity and clinical accuracy to alert for early recurrence is critical to achieve better outcome. Methods: We conducted a prospective observational study (NCT05446545) enrolling newly diagnosed patients who underwent primary debulking surgery or platinum sensitive relapsed patients with secondary cytoreductive surgery. The patients received SOC therapies followed by PARPi or bevacizumab for maintenance. Samples were assessed using a tumor-informed ctDNA assay (Huajianwei bespoke MRD test). For each patient, tumor tissue was sequenced by a whole exome panel at 500× coverage, a personalized set of primers targeting tumor-specific variants was designed. ctDNA from plasma samples was then amplified with the primer set before being sequenced at an ultra-high depth of >100,000×. Results: A total of 98 patients (861 plasma samples) were analyzed between Sep. 2021 and Dec. 2023 (data cut-off), including 80 (81.6%) newly diagnosed patients and 18 (18.4%) relapsed patients, respectively. The median follow-up time was 16.6 months (4.4 - 27.6 months) and 22.4% of patients relapsed. ctDNA was detectable in 98.9% (94/95 qualified samples) of patients at pre-OP, 73.5% (72/98) at post-OP, and 21.2% (18/85) after adjuvant chemotherapy (post-ACT). In paired samples analysis, residual disease was cleared in 70% (42/60) of patients through ACT. In comparison, CA125 had similar positive rates of 93.9% (92/98) at pre-OP and 78.4% (69/88) at post-OP, but only 1.3% (1/75) at post-ACT, resulting in an overly high clearance rate of 98.6%. Although the follow-up time is still insufficient, the data have clearly shown that CA125 detection at post-ACT is not credible (sensitivity: 39% (ctDNA) vs. 0% (CA125)). Moreover, ctDNA has a superior negative predictive value (NPV) than CA125. Specifically, all 21 patients remained ctDNA-negative after surgery and did not experience relapse while 6/19 CA125-negative patients relapsed. Excitingly, serial ctDNA monitoring during maintenance demonstrated an outstanding sensitivity of 95.0% and specificity of 86.6%, with a positive predictive value (PPV) of 67.9% and an NPV of 98.3% (58/59), compared to serial CA125 which only showed a sensitivity of 55.6% and a specificity of 90.2%, with a PPV of 62.5% and an NPV of 87.3%. Patients with positive biomarkers at any time experienced a significantly shorter recurrence-free survival (RFS) than those with negative status (ctDNA: HR 57.5, P < 0.001; CA125: HR 6.3, P < 0.001). Conclusions: Bespoke ctDNA assay achieved much higher clinical accuracy than CA125. Longitudinal monitoring in maintenance provides the best predictive and prognostic value, which may provide vital evidence for treatment decisions and patient management. Clinical trial information: NCT05446545 .