Clinical development of molecular residual disease (MRD) and multi-cancer early detection (MCED) using liquid biopsy multiomics with artificial intelligence (AI)

Early detection of cancer and precise recurrence monitoring remain major unmet needs in oncology. Conventional screening is limited to a few cancer types, leaving nearly half of cancers without established programs. Multi-cancer early detection (MCED) tests based on circulating tumor biomarkers have shown promise, but sensitivity for early-stage remains a challenge. In parallel, detection of molecular residual disease (MRD) using circulating tumor DNA (ctDNA) has emerged as a powerful prognostic and predictive tool, though current assays remain limited in sensitivity and specificity. This study aims to integrate multi-omics data to develop more refined and highly sensitive MCED and MRD assays.

This study leverages clinical information and biospecimens from patients with cancer and cancer-naïve individuals. Samples from patients with cancers will be derived from the MONSTAR-SCREEN-3 study, while those from cancer-naïve individuals will be obtained from the Tohoku Medical Megabank Project. Comprehensive analyses will include whole-genome sequencing (WGS), whole-exome sequencing (WES), whole-transcriptome sequencing (WTS), proteomics, metabolomics, and microbiome profiling using stool and saliva. Artificial intelligence (AI)-based multi-omics integration will be performed to develop novel MCED and MRD assays and to evaluate their clinical performance. The primary endpoints are the sensitivity and specificity of MCED and MRD assays.

This is the first large-scale study to integrate comprehensive multi-omics profiling with AI for MCED and MRD assay development. The findings are expected to advance precision oncology by improving early diagnosis and recurrence monitoring.

UMIN000053815, approved by the Institutional Review Board of the National Cancer Center Hospital East.