AI Tool Improves Speed and Accuracy of Cervical Cancer Treatment Planning

Cervical cancer affects around 600,000 women globally each year and causes approximately 340,000 deaths. Brachytherapy is among the most effective treatments for this cancer, but its use is often limited by the time and expertise required to create individualized treatment plans. Planning can take over an hour, varies widely between hospitals, and may require patients to remain under sedation while clinicians fine-tune radiation delivery. Researchers have now demonstrated an automated approach that can generate high-quality, personalized brachytherapy plans in minutes, addressing time constraints, variability in care, and the risk of human error.

The solution was developed by researchers at the University of California San Diego (La Jolla, CA, USA) who leveraged advanced computing infrastructure to tackle long-standing inefficiencies in radiation treatment planning. The solution integrates deep learning models with streamlined data processing and was trained and tested using large-scale computational resources. It has been embedded into a widely used clinical software platform, allowing clinicians to access the tool without changing their existing workflows.

The artificial intelligence (AI) system analyzes patient medical images and automatically generates a customized brachytherapy plan with a single click. Instead of relying on prolonged manual adjustments, the tool produces optimized radiation plans in under four minutes while maintaining clinical quality. Development and benchmarking were carried out using the Voyager supercomputer, an NSF-funded system optimized for AI workloads. This high-performance environment enabled rapid model training, testing, and validation across hundreds of patient cases.

The AI-generated plans were evaluated against those created by experienced clinicians using hundreds of real patient cases. Researchers found that the automated plans matched expert-level quality while requiring only a fraction of the time. The results, published in Brachytherapy, indicate that the system can significantly reduce planning time, minimize patient discomfort related to prolonged procedures, and lower the likelihood of errors associated with manual planning.

In addition to improving efficiency, the tool could help standardize brachytherapy care across hospitals, including clinics with limited resources or specialized expertise. By reducing time pressure, clinicians may be able to focus more on refining treatment quality rather than meeting procedural constraints. The research team plans to adapt the AI platform for other cancers treated with radiation, including breast and prostate cancer, with the goal of broad deployment across healthcare systems worldwide.

“The new tool uses AI to automate and speed up treatment planning and with just one click, the system analyzes a patient’s medical images and creates a high-quality, customized plan in less than four minutes — potentially reducing both the patient’s discomfort and the risk of human error,” said AI researcher Lance Moore. “This approach could help standardize care, especially in clinics with fewer resources or less specialized staff, and may allow more focus on improving plan quality rather than rushing to finish.”

Related Links:
University of California San Diego