Molecular residual disease detection for ovarian cancer using a WGS-based ctDNA assay: Initial findings from the MONSTAR-SCREEN-3.

202 Background: Ovarian cancer, including fallopian tube and primary peritoneal carcinoma, is often diagnosed at advanced stages. Although cytoreductive surgery and perioperative chemotherapy is a fundamental strategy, treatment outcomes remain unfavorable. Molecular residual disease (MRD) assessed by circulating tumor DNA (ctDNA) has emerged as a novel prognostic factor in multiple cancer types. However, ctDNA data in ovarian cancer remain limited, particularly using whole-genome sequencing (WGS)–based approaches. Methods: MONSTAR-SCREEN-3 is a prospective, multicenter study involving 1,100 patients (pts) with solid tumors, including ovarian cancer, undergoing curative treatment. Personalized panels (Precise MRD, Myriad Genetics) incorporating up to 1,000 targeted variants identified by WGS of matched tumor tissue were constructed. Plasma samples were collected sequentially at baseline, after preoperative chemotherapy, and at 1 month (M) post-surgery, every 3M during year 1, and every 6M during year 2. Results: As of December 2025, 65 pts with ovarian cancer were enrolled. Of whom, 40 ovarian cancer pts with both ctDNA results and clinical data available were included. 72.5% (29/40) were in FIGO stage III/IV, and 30 pts underwent upfront surgery and 10 pts received preoperative chemotherapy. Serous carcinoma was the most common histological type (50%). Thirty-five pts received curative-intent surgery, the majority of whom received adjuvant chemotherapy. Personalized panel creation was successful for all pts, with 82.5% achieving the maximum 1,000-variant panel. Among a total of 112 plasma samples passed quality control, 87 (77.7%) showed ctDNA positivity, with 26.4% (23/87) detected at an ultra-sensitive level (tumor fraction [TF] < 100 parts per million [ppm]). ctDNA positivity at baseline (100% detected at TF over 100 ppm), 1M, and 3M post-surgery, were 100%, 76.7%, and 62.5%, respectively. Among post-operative samples, ultrasensitive-level detection was observed in 34.8% at 1M and 60% (9/15) at 3M, with a minimum TF of 2.4 ppm. Even in cases with macroscopic complete resection, only 27.8% (5/18) were MRD-negative. With a median follow-up of 3.7 months (0-13.3 months), three pts relapsed with a median lead time of 2.9 months (range, 0-6 months). Extended follow-up and longitudinal ctDNA data will be presented. Conclusions: In advanced ovarian cancer undergoing curative treatment, we reported the dynamics of ctDNA using a WGS-based ctDNA detection assay. Ovarian cancer demonstrated a high ctDNA-shedding phenotype; nevertheless, a clinically relevant subset of post-operative samples was detectable at ultrasensitive levels, underscoring the necessity of high-sensitivity assays. Further follow-up is warranted to establish the clinical significance of ultra-sensitive MRD detection in ovarian cancer. Clinical trial information: UMIN000053975 .