Ultra-sensitive whole-genome sequencing–based molecular residual disease detection in resectable sarcoma in MONSTAR-SCREEN-3.

Abstract

11544 Background: Circulating tumor DNA (ctDNA)–based molecular residual disease (MRD) detection has shown strong prognostic value in multiple epithelial cancers, but its utility in sarcoma remains poorly defined. Sarcomas are highly heterogeneous and often represent low-shedding tumors, limiting the sensitivity of conventional assays. The MONSTAR-SCREEN-3 evaluates the clinical performance of a whole-genome sequencing (WGS)–based MRD assay in a pan-cancer cohort, including patients with sarcoma. Methods: Personalized ctDNA panels were generated using a WGS-based tumor-informed platform (Myriad Genetics), incorporating up to 1,000 tumor-specific variants identified through WGS of matched tumor tissue. Serial plasma samples were collected at baseline, 1 month post-surgery, quarterly during the first year, and biannually thereafter for up to two years. Results: As of November 2025, 35 patients with resectable soft tissue sarcoma (n=33) and osteosarcoma (n=2) were enrolled; MRD results were available for 72 samples from 24 patients. Of whom, histology was myxofibrosarcoma (n=7), undifferentiated pleomorphic sarcoma (n=7), myxoid liposarcoma (n=5), dedifferentiated liposarcoma (n=3), epithelioid sarcoma (n=3) and others (n=10). Clinical staging distribution included Stage I/II/III/IV: 15%/19%/50%/15%. Most of patients underwent upfront radical surgery (2 cases received neoadjuvant chemotherapy). Personalized panel creation succeeded in 100% of patients (24/24), identifying a median of 3,510 highly confident tumor-specific alterations per patient (range: 730-7,931) and yielding bespoke panels containing 386-1,000 alterations. Customized panels were created with 96.4% SNVs and 3.6% indels. The assay demonstrated 87.5% baseline ctDNA detection (21/24), with 16.7% detected at ultra-sensitive levels (tumor fraction <100 parts per million [ppm]). In the ctDNA-negative cases, ctDNA tumor fractions were 0.0, 0.8 and 6.2 ppm, classified as ctDNA-negative based on a statistical threshold set at 99.615% specificity. Post-surgical MRD positivity rates were 20.0% (4/20) at 1 month, 7.1% (1/14) at 3 months, and 28.6% (2/7) at 6 months. Among these patients, four patients developed radiographic recurrence, with MRD detection preceding imaging by median 1.1 months (0.1-3.0). Extended follow-up and comprehensive longitudinal ctDNA dynamics will be presented. Conclusions: The WGS-based ctDNA assay demonstrated high technical feasibility in sarcoma, enabling detection of patient-specific tumor variants across a highly heterogeneous and low-shedding tumor population. In our cohort, MRD positivity was observed exclusively in patients who subsequently developed recurrence, highlighting the potential utility of WGS-based tumor-informed MRD analysis for surveillance and risk stratification in sarcoma. Clinical trial information: UMIN000053975.