Using AI to Screen for Cancers Now May Boost Survival Down the Road

In an interview with The American Journal of Managed Care^® (AJMC^®), Douglas Flora, MD, FACCC, executive medical director of oncology services at St. Elizabeth Healthcare, and editor in chief of AI in Precision Oncology, discussed how artificial intelligence (AI) can expedite processes for physicians. In this part of the interview, Flora highlights the ways AI can help providers screen for cancers at earlier stages and tailor treatments for individual patients, and how he measures whether an AI tool is being helpful.

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

Transcript

What factors do you think are most important to consider when assessing whether an AI tool is beneficial in cancer care?

We have 2 goals. We want to do things that are the right thing for our patients. Assessing things like outcomes, for us, may mean how many patients we were able to detect with an earlier stage cancer, which should cause us a stage migration and a shift in survivals, ultimately, 5 years down the road. Were we better at choosing tools that let us choose the right agents, at the right time, for the right patient, at the right dose?

On the hospital side, the business side, did it have a real financial ROI [return on investment]? Did we reduce the number of patients who had to go to an ER [emergency room] or got readmitted? Or did we reduce our lengths of stay or hospital-acquired infections or the other things that measure quality? I think this is one of those rare circumstances where the AI has emerged as a tool that can help everybody—help better patient outcomes, help the doctors provide safer, more efficient care—and also hopefully keep the lights on at our hospital, because we're more efficient at delivering the care that we promised we would.

How can health systems ensure AI tools are being trained and used in a way to ensure equitable care for each patient?

I think we're trying to keep pace, as everyone is. AI is flying right now. Where we've started is in a very conservative but hopeful way. We've used a lot of it for screening, trying to detect incidental pulmonary nodules before the human eye would discern them as one. Blood tests are emerging that may be better at detecting cancers at an earlier stage, even [earlier] than a mammogram or a colonoscopy. We're using tools like colonoscopy augmenters that might put a red box around something that needs to be biopsied that may not have been elucidated in a normal test. Having our mammograms overread. Those things, I think, are very safe.

As we lean forward into this, maybe a year or 2 from now, the arrival of agentic AI and the agents that will assist us in our daily lives—both personal and professional—are exciting to me. I think that's when the doctors will have sort of a J.A.R.V.I.S. [Editor's note: This is a reference to Tony Stark's AI in Marvel Comics] on their shoulder that can help guide them and say, "This is the current evidence, this is current standard of care. Maybe this drug doesn't mix with that drug well. Would you like to be made aware of that?" I would like to.