Integrating experimental biology, computational methods, and artificial intelligence in anticancer drug discovery:...

Cancer remains a leading cause of mortality globally, with the incidence projected to reach 28.4 million new cases annually by 2040. Traditional drug discovery is notoriously inefficient: 10-17 years and up to $2.8 billion per approved drug, with fewer than 10% of clinical candidates succeeding. This review critically examines how the integration of experimental methods with computational biology and artificial intelligence (AI) can accelerate anticancer drug development. We argue that an iterative feedback loop in which sophisticated in vitro and in vivo models provide biological context for validation while AI accelerates target identification, de novo molecular design, and ADMET prediction offers a path forward. FDA-approved drugs, including imatinib, crizotinib, and larotrectinib, have emerged from such integrated pipelines. A persistent translational gap still limits progress owing to three major barriers: poor data fidelity and lack of standardization, limited interpretability of AI models (the black box problem), and regulatory complexity. Addressing these challenges requires physiologically relevant organ-on-a-chip systems, explainable and physics-informed AI, federated learning ecosystems, and adaptive regulatory frameworks. Overall, the continuous integration of computational prediction with experimental validation in a rigorous, data-driven pipeline is critical for accelerating the development of next-generation anticancer therapies.