Leveraging Human Nuance in the AI Era of Radiation Oncology

As artificial intelligence (AI) and automated workflows rapidly redefine the technical landscape of radiation oncology, the role of the clinician is shifting from primary data processor to expert interpreter. Although AI tools offer unprecedented efficiency, they often lack the real-time clinical nuance derived from direct patient interaction.

During the 2026 American College of Radiation Oncology (ACRO) Summit, CancerNetwork® sat down with Sunil W. Dutta, MD, assistant professor in the Department of Radiation Oncology at Emory University School of Medicine, to discuss how the next generation of radiation oncologists can maintain their edge. Dutta noted that the most critical soft skill for modern residents isn’t found in a coding manual or a software suite, but in the follow-up clinic. By witnessing the subtle toxicities and lived experiences of patients after treatment, physicians develop the specialized expertise required to provide truly informed consent and sophisticated case planning that AI cannot yet replicate.

Transcript:

As a Residency Site Director, what soft skill do you find most essential for the next generation of radiation oncologists to master as the field becomes more automated and AI-driven?

There are automated and advanced providers running a lot of the follow-up clinics. One of the things that I found most helpful is seeing your patients during follow up and seeing the nuance between different [adverse] effects from the radiation. Seeing those [adverse] effects and talking to those patients and their experiences will shape how you plan cases in the future and the informed consent process with future patients. It’s important for new graduates or residents to see patients and follow up as [often] as they can while still having clinic time to see new patients. That’s going to get you to that experience level, or that expert radiation oncologist, where you can give a thorough, informed consent with your new patients and learn how to better treat your patients [who are experiencing adverse] effects.

I personally use AI. It’s a great tool. One of the things I’ve noted is that its knowledge goes up to, generally, at least 1 to 2 years ago, so I would just encourage coming to conferences like this. Seeing presentations of the latest data and latest experiences is one way that we can offer patients the most up-to-date care in the age of AI, which may assist us but not replace the current standard of care.