Harnessing the power of artificial intelligence for clinical trials in cancer

Cancer research has become increasingly data-intensive, with digital pathology, imaging, and genomic sequencing generating vast, heterogeneous datasets. Artificial intelligence (AI) now plays a growing role across the oncology continuum, offering tools to interpret complex data, streamline workflows, and enhance clinical decision-making. In the context of clinical trials, AI is emerging as a catalyst for more efficient, inclusive, and data-driven research.

This review summarizes how AI, encompassing foundational machine learning (ML) models alongside advanced deep learning (DL) and large language model (LLM) systems, is being applied across the oncology trial lifecycle - from design and recruitment to data management and outcome assessment. Tools such as Trial Pathfinder, TrialGPT, and PRISM demonstrate the ability to emulate trial criteria, accelerate patient matching, and improve eligibility accuracy. The review also highlights key challenges related to algorithmic bias, explainability, accountability, and evolving regulatory oversight by the FDA and EMA.

AI is transitioning from conceptual promise to operational utility in oncology clinical research. As regulatory frameworks mature, harmonizing innovation with patient protection will be essential. When responsibly implemented, AI can bridge the gap between research and real-world care, transforming oncology trials into faster, fairer, and more reliable engines of discovery.