Multimodal radiomics for precision management of colorectal cancer

Colorectal cancer (CRC) treatment continues to face substantial challenges, including persistent recurrence risk, the development of drug resistance, and marked interpatient variability in therapeutic response. Multimodal radiomics (MMR), through the integrated analysis of computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET), and molecular data, enables multidimensional characterization of tumor morphology and underlying biology, thereby providing an expanded basis for precision-oriented decision-making. This review synthesizes recent progress in the clinical application of MMR across key domains, including preoperative staging, prediction of treatment response, surveillance for recurrence, and assessment of efficacy for immunotherapy and targeted therapies, with particular attention to how multimodal integration may improve characterization of metastatic heterogeneity and support individualized treatment planning. We further delineate the principal barriers to clinical translation, notably the lack of robust standardization across acquisition and segmentation workflows, limited model generalizability due to heterogeneity in datasets and validation strategies, and unresolved issues in ethical governance and data stewardship. Looking forward, the integration of radiomics with genomics, clinical records, and artificial intelligence (AI) holds promise for building patient-specific digital twin systems, supporting CRC management with more data-driven, individualized decision support, provided that harmonization, external validation, and prospective evaluation are addressed.