Ultra-sensitive molecular residual disease detection via tumor-informed whole-genome sequencing-based ctDNA assay in resectable urothelial cancer in the MONSTAR-SCREEN-3 project.

810 Background: Circulating tumor DNA (ctDNA) has emerged as a promising biomarker for detecting molecular residual disease (MRD) and is increasingly integrated into both clinical practice and translational research in various cancers. Conventional methods often lack sensitivity for MRD, particularly in low-shedding tumors. To overcome this limitation, the MONSTAR-SCREEN-3 study is evaluating the utility of a tumor-informed whole-genome sequencing (WGS)-based MRD assay across more than a dozen cancer types. Here we report preliminary results from the urothelial cancer cohort, including upper tract urothelial cancer (UTUC) and bladder cancer (BC). Methods: MONSTAR-SCREEN-3 is a prospective, multicenter study enrolling 1,100 patients receiving curative-intent therapy in the definitive cohort. Personalized ctDNA panels were generated using a WGS-based tumor-informed platform incorporating up to 1,000 tumor-specific alterations. Serial plasma samples were collected at baseline, after neoadjuvant chemotherapy (if applicable), 1 month post-surgery, quarterly during the first year, and biannually thereafter for up to two years. Results: As of September 2025, 62 patients with resectable UTUC (n = 22) and BC (n = 40) had been enrolled. MRD results were available for 63 samples from 20 patients. Median age was 75 (range 50-90), and 70% were male. Among the 20 patients with available WGS data, a median of 20,081 panel-eligible alterations per patient were identified (range: 1,733-184,521). Nine patients received neoadjuvant chemotherapy and 11 had upfront surgery. Two patients also underwent transurethral resection of bladder tumor (TURBT). Seventeen patients underwent radical surgery, with pathological stages: pStage 0 (n = 4), I (n = 2), II (n = 5), III (n = 5), and IV (n = 1). ctDNA results were available for 18 patients at baseline, and ctDNA was detectable in 100% of evaluable cases (10/10 BC and 6/6 UTUC), with 18.8% (3/16) detected at ultra-sensitive levels (tumor fraction < 100 parts per million ppm; minimum: 2.9 ppm). The remaining two patients showed undetectable ctDNA at enrollment, consistent with prior TURBT. Post-surgery MRD positivity rates were 63% (10/16) at 1 month and 36% (5/14) at 3 months. One patient with 1-month post-surgery MRD positivity ( < 1 ppm) developed radiographic recurrence, with MRD detection preceding imaging findings by 1.7 months. One patient with intravesical recurrence was MRD-negative at 1-month post-surgery. Extended follow-up and longitudinal ctDNA dynamics will be presented. Conclusions: The WGS-based personalized ctDNA assay demonstrated technical feasibility in urothelial cancer highlighting the critical importance of ultra-sensitive platforms for low-shedding tumors.