Nat Rev Cancer. 2026 Apr 21. doi: 10.1038/s41568-026-00922-2. Online ahead of print.
ABSTRACT
The extensive heterogeneity of cancer across biological scales necessitates a holistic approach beyond single-analyte methods. Integrating multi-omics data - from genomics to proteomics - with multimodal information, such as clinical records and medical imaging, offers a comprehensive, systems-level view of tumorigenesis. Artificial intelligence (AI) has emerged as the essential technology to decipher these complex, high-dimensional datasets, powering substantial advances in early diagnosis, precise patient stratification, prediction of therapeutic response and the elucidation of mechanisms of drug resistance. To translate these powerful predictive models into practice, explainable AI is critical for building clinical trust and generating novel, testable biological hypotheses. While challenges in data accessibility and model generalizability persist, the field is advancing toward patient-specific digital twins, promising to simulate individual disease trajectories and optimize treatments, thereby heralding a new era of precision oncology.
PMID:42014628 | DOI:10.1038/s41568-026-00922-2