AI Boosts Cancer Radiotherapy Planning, Cuts Wait Times

Researchers at an unnamed institution present a new artificial intelligence system, BAT-RM, designed to automatically outline cervical cancer targets for radiotherapy planning. The system combines transformer and Mamba architectures to efficiently analyze medical images and define tumor boundaries and surrounding organs.

In tests against existing methods, BAT-RM improved the accuracy of target volume delineation, particularly for gross tumor volume and clinical target volume, as well as organs at risk like the rectum and bladder. A multi-center study involving 13 radiation oncologists found the AI increased the accuracy of junior doctors’ outlining to near-senior levels.

The system has already been deployed at a partner hospital, reducing patient wait times for treatment from days to hours. This suggests AI-assisted auto-contouring can address radiotherapy access issues where demand exceeds specialist availability. Further research could explore the system’s adaptability to other cancer types and imaging modalities.