The “Secret Sauce” in Oncology: The Partnership Between Clinicians and AI

Debra Patt, MD, PhD, MBA, MPH, serves as executive vice president of Texas Oncology, medical director for public policy for The US Oncology Network, and most recently, president of the Community Oncology Alliance. In this interview, Patt highlights how artificial intelligence (AI) is poised to revolutionize oncology care, from diagnostics to clinical decision support.

While AI can improve efficiency and accuracy, Patt warns of risks such as bias and data fabrication, stressing that the technology must be integrated thoughtfully alongside clinicians, and calls for better education and training to ensure AI enhances community oncology care without leaving smaller practices behind.

This transcript has been lightly edited; captions were auto generated.

Transcript

You’ve spent most of your career working with technology in practice and now with AI in particular. Can you discuss the promise of AI in oncology but also the equity issues? How do we balance the potential while also ensuring that some practices are not left behind?

I think AI will touch every aspect of care delivery, from how we answer the phones, how we communicate with patients, how patients get screened, how their mammograms get read, how their colonoscopies get interpreted, how their pathology is read, how their diagnostic imaging is read, the clinical decision support that manages the therapies that we prescribe, how each of our administrative partners works. So, I think it's going to be in every aspect, and I think that by and large, that will create efficiencies when we know how to use it.

However, there are risks that need to be managed. Bias is one of them with the large language models; hallucination is another, [as is] data fabrication. Those are some hazards and liabilities that need to be managed. But that said, the error rates with large language models are incredibly low, and they continue to get better.

To error is human, and also to error is in machine learning. But it happens actually much less frequently in machine learning. I believe the secret sauce in medicine is the partnership between machine learning and AI and the clinicians that deliver care, and the staff that support practices because I don't think any of them are going to act in a silo. I think it's always going to be a partnership. I think that AI will act as clinical decision support, that we'll be empowered to give patients better education. But there's a huge gap in educating the workforce in order to get us where we need to be.

So, I like to say that we have our North Star in sight, and we're sort of on a journey, but that journey will change as we go along that journey, and we need to do the important work of educating stakeholders on how to use these tools to make sure that they can improve community care as much as possible.