Surgery and Radiology Team Up to Evaluate AI Chatbot for Patients with Adrenal Nodules

Alexander Chiu, MD

Each year, over five million patients in the United States have an adrenal nodule that is identified purely by chance during a CT scan or an MRI, and many of these nodules produce excess adrenal hormones that can cause serious health issues. While clinical guidelines recommend that these nodules be evaluated, in practice the vast majority are not tested due to difficulty accessing specialty care. With a new Translational, Basic, & Clinical Research Pilot Award from the UW Institute for Clinical and Translational Research (ICTR), Division of Endocrine Surgery Assistant Professor Alexander Chiu, MD and UW Department of Radiology Professor Alan MacMillan, PhD will be testing a patient-facing artificial intelligence (AI) chatbot to streamline adrenal nodule counseling visits and increase the completion of recommended testing.

“Currently, patients with adrenal nodules require multiple face-to-face visits with specialists. This resource-heavy approach creates logistical barriers for patients, introduces care delays, and strains the capacity of endocrine specialists,” explained Chiu. “By designing an AI chatbot that can provide patient education, answer commonly asked questions, and provide instructions for laboratory testing, we hope to improve healthcare access and convenience for patients, reduce unnecessary specialty clinic visits, and triage patients to the correct provider based on their lab results.”

Chiu and MacMillan previously partnered with UW’s Computer Science Department and the N+1 Institute to create a beta version of the chatbot. With the support of their new ICTR grant, they will refine the beta version to ensure it is safe, accurate, and understandable for patients and then pilot test it in clinical settings.

“Approximately 20% of adrenal nodules produce excess hormones that can increase the risk of life-threatening complications, such as stroke, metabolic disease, and sudden, severe spikes in blood pressure, so it’s essential that we find a solution that helps us identify these patients,” said Chiu. “If the intervention ultimately proves successful after both pilot and future large-scale, multi-site testing, it will help us improve the care of patients with adrenal nodules. The lessons learned in this project could also be directly applied to similar interventions that can help improve access to specialty care of all types, including other high-volume incidental findings.”