|
Getting your Trinity Audio player ready...
|
After radiation therapy, certain cancer patients face a lingering question: Is a troubling lesion a return of the cancer, or is it damage caused by treatment? At Stanford, researchers are testing an artificial intelligence model designed to help doctors tell the difference.
Dr. Michael Chang is an assistant professor in the department of otolaryngology at Stanford University School of Medicine, where he practices medicine as a specialized ear, nose, and throat doctor called a rhinologist. In December, Dr. Chang and a team of researchers published a study which found that an AI model could identify a serious radiation-related complication called skull base osteoradionecrosis with 85% accuracy — roughly on par with the abilities of experienced clinicians, Chang said.
“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.”
The model focuses on nasopharyngeal carcinoma, a rare but deadly cancer that disproportionately affects Asian American communities. The tumor develops deep behind the nose and symptoms can be subtle, and many cases are diagnosed late.
In the study, a computer vision model was trained on about 1,500 endoscopic images of the nasal pharynx from 192 patients. While the model performed strongly at detecting osteoradionecrosis versus healthy tissue, the system showed moderate accuracy in identifying recurrences of nasopharyngeal carcinoma. At times, the model confused recurrent NPC with radiation-related damage or normal-appearing tissue. Still, Dr. Chang said with a larger dataset, the models could perform even better. The research was funded by the Center for Asian Health Research and Education, a medical center at Stanford dedicated to advancing the health of Asian people worldwide.
There are 100,000 new cases of nasopharyngeal carcinoma, Dr. Chang said, and the NIH estimates that roughly 80,000 people die from the cancer each year. While NPC is generally treatable with radiation treatment if diagnosed early, the cancer is often not detected until more advanced – and deadly – stages. While previous AI medical research has focused on making an initial diagnosis, Dr. Chang’s research breaks new ground by using AI to identify images after treatment.
Dr. Chang’s research can help improve health equity in several areas. For one, it can bolster health outcomes for people of Asian descent in a cancer with a high fatality rate. More broadly, Dr. Chang expects that AI can bridge discrepancies in health care by democratizing expertise. AI imaging models may also be a more objective system than human clinicians whose judgment may be subjective.
“A patient getting treated at Stanford is going to have access to all the expert surgeons and pathologists and radiologists,” Dr. Chang said. “But a patient somewhere else may not be able to have the same access to that level of expertise.”
Dr. Chang expects the overall impact of AI in the medical field to be “enormous in all aspects of health care.” AI can help automate tasks like billing or charting that are time-intensive and, as his research shows, assist with medical decision-making. Though physical AI – technology that works in physical space like robots — is less advanced, Dr. Chang said he can envision AI being co-pilots for surgeons in the future. But he doesn’t intend for AI to replace clinicians or replicate the doctor-patient relationship.
“The goal is to have AI augment the clinician’s decision-making,” Dr. Chang said. “I think there is certainly a big role for AI in diagnosis and in treatment and surveillance.”
‘The end of an era’: After 36 years in Palo Alto, Il Fornaio prepares to closeJanuary 13, 2026 2:05 pm