The development of oncology therapeutics is currently impeded by exorbitant costs, protracted timelines, and high clinical attrition rates stemming from the inherent complexity of tumor biology. Artificial Intelligence (AI) is transforming oncology drug discovery, shifting the paradigm from trial-and-error experimentation to one of data-driven rational design. In this paper, we review recent AI advances in four key areas. First, regarding Target Identification, we examine how multi-omics integration and deep learning uncover novel vulnerabilities, such as synthetic lethal pairs and immune checkpoints. Second, we analyze the evolution of Virtual Screening, moving from classical docking to graph neural networks that efficiently explore vast chemical spaces. Third, we highlight the shift toward Generative Molecular Design, where AI models create de novo small molecules, protein binders, and nucleic acid therapeutics with tailored functional properties. Fourth, we discuss AI applications in Preclinical Evaluation for predicting toxicity and efficacy. Finally, we critically assess current challenges—including data standardization deficits and the “black box” nature of deep learning—and propose emerging strategies, such as automated design-make-test workflows, to bridge the gap between computational prediction and clinical reality.
Tang et al. (Mon,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: