10521 Background: Early detection and diagnosis is key to reducing cancer-related mortality. We developed a blood-based multi-cancer early detection test, called THEMIS, using multimodal analysis of plasma cell-free DNA (cfDNA) methylation, fragmentation, and copy number alteration features by whole-methylome sequencing (WMS). The real-world performance of THEMIS was investigated with a prospective cohort of high-risk populations and the preliminary results were reported here. Methods: A total of 1,374 adults over 65 y with additional cancer risk factors for five common cancers (colorectal, esophageal, liver, lung, or stomach cancer) were recruited in the city of Wuxi, Jiangsu province, China. Additional risk factors included fecal occult blood, smoking history, H. pylori infection, HBV infection, and serum protein biomarkers. Participants had no known cancer diagnosis at the time of administration. Blood samples were collected and subject to low-pass WMS assay (~2X) and cancer probability was predicted by the THEMIS model, which was developed and locked in previous studies. Participants predicted as cancer positive by THEMIS were followed up until positive cancer diagnosis in a clinic or one year after receiving the test report, whichever was earlier. Results: At a fixed cutoff rendering 99% training specificity, THEMIS predicted 58 (4.15%) positive cases, among whom 19 participants consented to and completed follow-up. Ten participants received confirmed cancer diagnosis within the follow-up period, including five lung cancers, three liver cancers, one esophageal cancer, and one colorectal cancer. These results yielded a positive predictive value (PPV) of 52.6%. Those who declined follow-up upon a positive THEMIS report were predicted with digestive tract cancers, which need gastrointestinal endoscope for diagnostic confirmation. Conclusions: The nonivasive blood-based THEMIS test demonstrated promising performance and patient compliance for clinical utility to detect cancer early and improve patient outcome. Given that clinical follow-ups were only conducted with a subset of participants, future investigations of the precise real-world performance of THEMIS are needed with more complete patient follow-up.