Stanford Researchers Use AI to Monitor Rare Cancers after Radiation Therapy

The research team led by Dr Michael Chang published findings showing that their AI model could identify this complication with about 85% accuracy.

Stanford University researchers have developed and tested an AI-based imaging system that helps doctors distinguish between cancer recurrence and radiation-related damage in patients treated for a rare and deadly cancer, highlighting how AI can strengthen post-treatment monitoring and clinical decision-making.

The AI model analyzes endoscopic images and has shown accuracy comparable to experienced clinicians, pointing to a growing role for AI in post-treatment cancer surveillance.

After radiation therapy, patients with head and neck cancers often develop lesions that are difficult to interpret. These changes may signal recurrent disease or may reflect complications from radiation itself. One such complication is skull base osteoradionecrosis, a serious condition caused by radiation damage to bone.

In December, a Stanford research team led by Dr Michael Chang published findings showing that their AI model could identify this complication with about 85% accuracy.

“The major finding was that this is a very feasible way that we can apply AI in this area of health care,” Chang said. “Because a lot of what we do is very reliant on images and interpretation of images, there’s a lot of opportunity for computer vision to help augment clinicians’ ability to diagnose, to treat and to surveil different disease processes.”

To train the AI system, researchers used roughly 1,500 endoscopic images collected from 192 patients. The computer vision model performed well when distinguishing healthy tissue from osteoradionecrosis.

The study was supported by Stanford’s Center for Asian Health Research and Education, reflecting an effort to address health disparities through targeted AI innovation.

Unlike many AI studies that focus on initial detection, this research applies AI to the post-treatment phase, where clinical uncertainty is often highest. Chang believes this approach could also advance health equity by extending expert-level image interpretation beyond major academic centers.

“The goal is to have AI augment the clinician’s decision-making,” Chang said. “I think there is certainly a big role for AI in diagnosis and in treatment and surveillance.”

Looking ahead, Chang expects AI to have a broad impact across healthcare, from automating administrative tasks to assisting clinicians during diagnosis, treatment, and long-term surveillance. While he does not see AI replacing doctors, he envisions it becoming a trusted clinical partner.

Stay tuned for more such updates on Digital Health News