The SCRUM-MONSTAR Cancer-Omics Ecosystem: Striving for a Quantum Leap in Precision Medicine

Abstract

The SCRUM-Japan MONSTAR-SCREEN consortium is a nationwide molecular profiling project employing artificial intelligence-driven multiomics analyses for patients with advanced malignancies, aiming to develop novel therapeutics and diagnostics and deliver effective drugs to patients. Concurrently, studies assessing molecular residual disease-based precision medicine for resectable solid tumors, including CIRCULATE-Japan, are ongoing. The substantial data generated by these platforms are stored within a state-of-the-art supercomputing infrastructure, VAPOR CONE. Since 2015, our project has registered over 24,000 patients as of December 2023. Among 16,144 patients with advanced solid tumors enrolled in MONSTAR-SCREEN projects, 5.0% have participated in matched clinical trials, demonstrating a 29.2% objective response rate and 14.8-month median survival (95% CI, 13.4-16.3) for patients treated in the matched clinical trials. Notably, patients who received matched therapy demonstrated significantly prolonged overall survival compared with those who did not (hazard ratio 0.77; 95% confidence interval, 0.71-0.83). Significance: Our nationwide molecular profiling initiative played pivotal roles in facilitating the enrollment of patients with advanced solid tumors into matched clinical trials and highlighted the substantial survival benefits of patients treated with matched therapy. We aim to facilitate an industry-academia data-sharing infrastructure ecosystem, fostering new drug discovery paradigms and precision medicine.

Figure 1.

Overview of the SCRUM-MONSTAR project. The SCRUM-Japan MONSTAR-SCREEN project, a nationwide molecular profiling project, uses AI-driven multiomics analyses for patients with advanced malignancies. The project encompasses various sub-studies for advanced solid tumors, including GI-SCREEN, GOZILA, MONSTAR-SCREEN, and MONSTAR-SCREEN-2. Concurrently, the CIRCULATE-Japan and COSMOS platforms manage MRD for resectable solid tumors. The BRANCH project explores the relevance of germline mutations in patients with solid tumor and their families. The project comprises several specialized committees, each diligently administered and operated by its respective group. These committees cover 12 areas: AI, ancillary research, database management, ethical, legal, and social implications (ELSI), genetic information, germline board, meta-information management, pathology, patient and public involvement (PPI), quality assurance/quality control (QA/QC), regulatory affairs, and trial management.

Figure 2.

The MRD project for resectable solid tumors. A, Annual patient enrollment in the CIRCULATE-Japan and COSMOS project, exceeding 7,000 participants (light green: CIRCULATE-Japan; blue: COSMOS). B, The demographics of the enrolled patients according to cancer types in the CIRCULATE-Japan and COSMOS project. C, The schema of CIRCULATE-Japan related trials and COSMOS projects. CIRCULATE-Japan is composed of GALAXY, VEGA and ALTAIR studies. The GALAXY study is a prospective large-scale registry designed to monitor ctDNA in patients with CRC. The COSMOS projects are composed of six cancer specific studies: CRC, GC, PC, HCC, MM, and BTC to evaluate the utility of early postoperative MRD analysis and early detection of various cancer types. NEXUS, VOLTAGE II, and Nir-Bev trials are linked to the CIRCULATE-Japan platform. D, Comparison of the predictive performance of various biomarkers for cancer recurrence. The HR with 95% CIs for recurrence, recurrence-free survival, or disease-free survival were previously reported for several biomarkers, including tumor budding status, lymphovascular invasion status, tumor differentiation, tumor sidedness, serum carcinoembryonic antigen level, tumor-infiltrating lymphocytes, Immunoscore, CpG island methylator phenotype, microsatellite Instability status, gene expression signature (Recurrence Score and ColoPrint), gene mutational status (BRAF, KRAS, TP53, and PIK3CA), and molecular expression status (MUC2, CDX2, and HER2), and ctDNA status in the GALAXY study. Among these biomarkers, ctDNA positivity after surgery in the GALAXY study demonstrated a substantially higher HR for recurrence compared with the other biomarkers. ACT, adjuvant chemotherapy; CAPOX, capecitabine plus oxaliplatin; CDX2, caudal-type homeobox transcription factor 2; CEA, carcinoembryonic antigen; CIMP, CpG island methylator phenotype; cy, cycles; FTD/TPI, trifluridine/tipiracil; GIST, gastrointestinal stromal tumor; HCC, hepatocellular carcinoma; HER2, human epidermal growth factor receptor 2; LVI, lymphovascular invasion; MM, malignant melanoma; MSI, microsatellite instability; MUC2, mucin 2; R, randomization.

Figure 3.

Consortium of the clinical trial platform and SCRUM-Japan registry. A, Annual participant enrollment across different projects, exceeding 16,000 participants (light green: GI-SCREEN, blue: GOZILA, light red: MONSTAR-SCREEN, light purple: MONSTAR-SCREEN-2, light orange: BRANCH, red line: total enrollment). B, The demographics of the enrolled participants according to cancer types in the SCRUM-Japan MONSTAR-SCREEN project. C, The schema of SCRUM-MONSTAR–related trials and SCRUM-Japan registry projects. Patients with specific genomic alterations are incorporated into pertinent clinical trials, with several tumor-specific and tumor-agnostic trials on the basis of MONSTAR-SCREEN platform. D, Nine trials completed enrollment. Of these trials, four (TRIUMPH, HERB, TiFFANY, and HERALD) met the primary endpoint, whereas two did not. Three trials are in the process of analyzing data. E, The structure of the SCRUM-Japan registry, which collects prospective data on patients with rare genetic alterations, informing future drug development and serving as an external control group. Currently, more than 500 patients with rare molecular alterations have been pooled, and their data will be used as the external control data for approval. BRCA, breast cancer; CervC, cervical cancer; EC, esophageal cancer; EDC, electronic data capture; EnC, endometrial cancer; GIST, gastrointestinal stromal tumor; HCC, hepatocellular carcinoma; HNC, head and neck cancer; IIT, investigator-initiated trial; MM, malignant melanoma; NEN, neuroendocrine neoplasms; OVC, ovarian cancer; PRAD, prostate adenocarcinoma; RCC, renal cell carcinoma; SIC, small intestinal cancer; UroC, urothelial cancer.

Figure 4.

The overview of target biomarkers in clinical trials related to SCRUM-MONSTAR-SCREEN. A, Background characteristics of patients participating in clinical trials. B, Left, Treatment lines examined in clinical trials; Right, class of drugs examined in clinical trials. C, Targeted biomarkers in clinical trials. D, Overall drug efficacy in matched clinical trials. CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease. E, Objective response rate by each targeted drug. FGFR, fibroblast growth factor receptor; HER2, human epidermal growth factor receptor 2; MEK, MAPK/ERK; MET, mesenchymal-epithelial transition; mTKI, multitargeted tyrosine kinase inhibitors; PARP1, poly(ADP-ribose) polymerase 1.

Figure 5.

Long-term outcomes of patients matched to clinical trials and comparison of matched vs. nonmatched therapy. A, PFS in patients treated with biomarker-matched therapy in clinical trials. B, OS in patients treated with biomarker-matched therapy in clinical trials. C, OS of patients treated with matched therapy vs. nonmatched therapy. m, months; MST, median survival time.

Figure 6.

Structure of SCRUM-Japan MONSTAR-SCREEN-3 and SCRUM-MONSTAR cyber space (VAPOR CONE). MONSTAR-SCREEN-3 consists of three cohorts: the Advanced (N = 1,700), Definitive (N = 1,100), and Hematology (N = 400) cohorts. Comprehensive multiomics analyses are designated per cohort, including various sequencing methods and evaluations for tissue and blood samples; WES, WTS, and spatial ST sequencing for tissue samples; WES, WTS, germline sequencing, and proteome analysis for blood samples; and microbiome analysis for fecal samples. Quality of life (QoL) evaluation using patent-reported outcomes from electric records, digital image collection, and digital pathology collection are also conducted. Sequential tissue–informed WGS-based MRD analysis is performed in the Definitive and Hematology cohorts, supplemented by IgH/TCR-based MRD in the Hematology cohort. VAPOR CONE represents a virtual space dedicated to the generation of quality-assured real-world evidence (RWE) within a secure HPC environment. This platform facilitates the development of RWE extraction, data analysis technologies, and various AI models, propelling forward the fields of cancer genomics and translational research across basic, applied, and clinical domains. Using generative AI technology, existing models within VAPOR CONE undergo continuous refinement, aiming to establish an advanced social infrastructure for cancer genome medicine, achieving decentralized clinical trial (DCT). Data within VAPOR CONE are merged with the collective expertise of a nationwide network of academic professionals. This synergy enables the reorganization, enrichment, and standardization of the VAPOR CONE database, aligning it with international standards in real time. The database serves as a robust, accurate data source, laying the groundwork for sophisticated applications in healthcare and pharmaceutical research. Continuous learning, facilitated by generative AI technology, enhances the machine intelligence of VAPOR CONE’s existing AI models, improving auxiliary diagnostic support, interpretation, report generation, and predictive capabilities for research outcomes. By bolstering the speed, efficiency, and cost-effectiveness of these processes, we aim to pioneer an unparalleled social infrastructure for advanced cancer genome medicine in Japan. ctNA, circulating tumor nucleic acid; CTx, chemotherapy; DCT, decentralized clinical trial; ePRO, electronic patient reported outcome; IgH/TCR, immunoglobulin receptor/T-cell receptor.