Much of the GI community has anchored its view of AI to adenoma detection rates, but Neil Parikh, MD, a gastroenterologist with Farmington-based Connecticut GI, says the bigger disruption will come from predictive analytics and chronic disease management.
“We’re really falsely caught up in AI being solely in polyp detection,” Dr. Parikh told Becker’s. “I feel like we were confused by this — the GI community just assumes AI means polyp detection — but where I see it going, and where I think other specialties are also taking it, is really in both predictive analysis and chronic care management.”
Polyp detection has drawn the most attention because its benefits are easier to see and measure, he said.
“A lot of what drives change is obviously clinical benefit, and secondary return on investment,” he said. “So you need to have the patient first — and I wholeheartedly believe most are driven by patient first — but you also have to be able to show a financial return.”
Additionally, an Aug. 12 study published in The Lancet found that routine use of artificial intelligence in colonoscopy procedures may reduce endoscopists’ ability to detect precancerous polyps when AI is not available. Adenoma detection rates in non-AI procedures dropped from 28.4% before AI adoption to 22.4% after, representing a 20% relative decline.
Inflammatory bowel disease is a prime example of where AI could meaningfully improve chronic care management, Dr. Parikh said. AI could synthesize patient data to predict when a flare might occur, anticipate treatment needs and forecast the course of illness before symptoms worsen.
Many gastroenterologists already collect the underlying data, he said, but typically through traditional workflows where patients are seen in the office, symptoms are documented, lab work is ordered and follow-up actions are based on results.
“I think AI is going to flip that,” he said. “I think AI is going to let us synthesize. It’s going to help us collect the data ahead of time, synthesize it even before the visit or in between visits, and predict the flare.”
He also expects AI to expand in GI oncology risk analysis and stratification. While some tools are already in use, Dr. Parikh predicts broader adoption.
“There’s a variety of tools where you look at Barrett’s esophagus, which probably has the most, but I think you’re going to see it in all GI cancers — hopefully eventually in pancreatic cancer as well — where we can do risk assessment,” he said. “In colon cancer, there was some work being done where you just took an entire population’s blood counts, put them into an AI algorithm, and started predicting — based on blood count and other factors — who needs maybe even earlier colon cancer screening intervention.”
Dr. Parikh also said CMS and payer support for chronic care management could accelerate the shift toward AI-enabled monitoring and management.
“I applaud CMS and the payers who are starting to see value in chronic care management and see value in remote patient monitoring,” he said. “And as they show that value, it will probably motivate clinicians and health systems to put the capital in for the infrastructure. Because you’ve got to have those RPM tools available. You have to have the ability to offer that to your patients. And that all requires money. So I think that’s where the shift is going.”
Still, he worries about GI bandwidth if AI expands identification of higher-risk patients without a parallel increase in clinical capacity.
“If we open this new AI reflex technology in any disease state, and we identify X more patients’ volume that needs to be clinically assessed,” he said. “We need to have the ability to triage that clinical population — unless eventually AI does that triaging as well, but that’s a second-stage process. But upfront, we need to have the bandwidth to handle the clinical volume that AI may bring us.”
While physician burnout has dipped slightly in recent years, concerns remain. Administrative burdens and rising patient volumes continue to drive stress and dissatisfaction. Combined with an aging population and physician shortages, many GI leaders predict burnout to influence workforce trends.
