Integrating Genomics, Radiomics, and Pathomics in Oncology: A Scoping Review and a Framework for AI-Enabled...

Area of Science:

• Oncology
• Artificial Intelligence
• Bioinformatics

Background:

• Multimodal AI integration in oncology is rapidly advancing.
• Evidence is heterogeneous and unevenly distributed across genomics, radiomics, and pathomics modalities.
• Current research faces challenges in data integration and validation.

Purpose of the Study:

• To map empirical studies integrating two or more omic modalities.
• To summarize AI integration and validation approaches.
• To identify research gaps for future surgomics development.

Main Methods:

• Scoping review following PRISMA-ScR guidelines.
• Searched major databases (PubMed, Ovid, Wiley, Google Scholar) for studies from January 2020 to March 2025.
• Charted study characteristics, modalities, fusion strategies, AI models, validation, and performance metrics.

Main Results:

• 11 studies met inclusion criteria, totaling 1078 participants.
• Most studies focused on radiomics-pathomics integration; genomics incorporation and tri-modal fusion were uncommon.
• Significant variation in clinical tasks, endpoints, preprocessing, and validation limited comparability.

Conclusions:

• Methodological activity is growing in radiopathomics and cross-scale association modeling.
• Tri-modal pipelines and clinically deployable multimodal workflows are underdeveloped.
• Surgomics is proposed as a conceptual roadmap, not a current clinical capability, informed by identified research gaps.