Artificial Intelligence and Multi-Omics Integration in Precision Oncology: A Review of Clinical Applications and Challenges

The convergence of artificial intelligence and multi-omics technologies is transforming precision oncology by enabling the integration of heterogeneous biomedical data to decode cancer complexity. This review synthesizes recent advances in AI-driven multi-omics integration, examining how machine learning and deep learning approaches are applied across genomics, transcriptomics, proteomics, metabolomics, radiomics, and spatial omics to construct comprehensive disease atlases. The clinical utility of these integrative models spans early detection, prognostic risk stratification, and prediction of therapeutic responses. AI models integrating multi-omics data have improved diagnostic accuracy by up to ۱۵% in ovarian cancer and ۲۰% in lung cancer. However, critical challenges remain, including data heterogeneity, limited model generalizability, the "black box" nature of deep learning, algorithmic bias, and a scarcity of prospective validation studies. Emerging approaches such as foundation models, federated learning, and digital twins offer promise in addressing these barriers and accelerating clinical implementation.

کلیدواژه ها:

Artificial Intelligence ، Multi-Omics ، Precision Oncology ، Machine Learning ، Cancer Genomics ، Radiomics ، Spatial Transcriptomics ، Clinical Translation