Key points are not available for this paper at this time.
Abstract Artificial intelligence (AI) is reshaping oncology by extracting clinically actionable signals from complex cancer data and accelerating drug development. In this narrative review, we summarize how machine learning and deep learning models support cancer phenotyping, including tumor detection, molecular subtyping, prognosis, and treatment response prediction across histopathology, radiology, and multi-omics data. We then discuss AI-enabled virtual screening, drug repurposing, generative molecular design, and hybrid computational–experimental pipelines that streamline oncology drug discovery and optimization. Cross-cutting limitations are examined, including data quality and representativeness, class imbalance, bias and fairness, model interpretability, and ethical, privacy, and regulatory challenges in clinical deployment. Finally, we highlight emerging directions such as multimodal foundation models, federated learning, AI stewardship, and patient-specific digital twins, and outline a roadmap for integrating trustworthy AI into precision oncology. Realizing the full potential of AI will require rigorous validation, transparent reporting, and close collaboration between clinicians, data scientists, regulators, and patients to ensure equitable, patient-centred benefit. Graphical Abstract
Wang et al. (Thu,) studied this question.