Optimizing the delivery of radiotherapy with artificial intelligence

A Nature Reviews Clinical Oncology review argues that radiation oncology, with its complex and data-intensive workflows, is among the medical specialties best suited to artificial intelligence, and catalogues applications already succeeding in practice, such as automated contouring of tumours and organs at risk, treatment planning and adaptive therapy. The authors set out the main obstacles to routine clinical adoption, including validation across diverse patient populations, regulatory approval, workflow integration and the risk of automation bias among clinicians. They stress that safe, supervised deployment with ongoing quality assurance is the precondition for these tools to translate into better patient outcomes. As a narrative review rather than a trial report, it consolidates existing evidence instead of adding new clinical data.