Beyond radiogenomics: advancing imaging integration and multi-omics horizons in cancer precision medicine

Cancer remains the leading cause of death worldwide, presenting substantial challenges to precision medicine due to its complex heterogeneity. Radiogenomics, as a method combining quantitative radiologic data with genomic information, provides a robust analysis framework to assess tumor heterogeneity and cancer progression. Here, we summarize the application of radiogenomics in two key fusion methods: feature-level and decision-level fusion. Feature-level fusion combines multimodal data into a rich feature set to improve the predictive power of models, while decision-level fusion integrates decision results from multiple independent models to improve robustness and reliability. Furthermore, we explore the integration of radiomics with various omics technologies, including transcriptomics, metabolomics, and proteomics. This integration enables a deeper understanding of the dynamic tumor microenvironment, metabolic dysregulation, and cancer progression mechanisms. Finally, we provide a detailed overview of publicly available datasets relevant to radiogenomics research, such as The Cancer Imaging Archive, cBioPortal, UK Biobank, and Human Connectome Project; and further describe multiple types of omics data and sample characteristics for each resource for the benefit to readers. In summary, this review charts a path beyond radiogenomics by advancing radiomics and multi-omics horizons to transform precision medicine in cancer.