UC San Diego Develops AI Tool to Reduce Cervical Cancer Radiation Planning Time to Minutes
Researchers at UC San Diego have created an AI tool that reduces the planning time for brachytherapy treatments for cervical cancer from hours to under four minutes. This system, detailed in the journal Brachytherapy, integrates a deep learning model into Varian's BrachyVision treatment planning software.
Developed using the Voyager supercomputer at the San Diego Supercomputer Center, the tool automates the generation of customized treatment plans with a single click. The AI model was trained on over 1,800 historical brachytherapy plans and predicts patient-specific radiation dose distributions based on medical images.
In tests involving 28 cases, the AI-generated plans differed from clinical standards by less than 4%, significantly shortening planning times compared to traditional methods. The researchers suggest this advancement could help standardize cervical cancer care, especially in resource-limited settings, while noting that automated plans should still be reviewed by clinicians.
