Imaging Tool Using AI to Predict Pancreatic Cancer Being Adapted to Black Patients

Cedars-Sinai investigators who previously developed an imaging tool that used artificial intelligence (AI) to predict pancreatic cancer are now working to adapt that tool specifically for Black patients, who have disproportionately high rates of the disease.

“The incidence of pancreatic cancer among the Black population is at least 50% higher than the incidence of other racial groups. Furthermore, research has shown that Black patients have the lowest survival rate,” said Debiao Li, PhD, director of the Biomedical Imaging Research Institute and professor of Biomedical Sciences and Imaging at Cedars-Sinai, and co-principal investigator of the study. “We know that there are genetic, socioeconomic and lifestyle differences between ethnic and racial populations, and we suspect some of these differences might affect pancreatic tissue and pancreatic cancer risk.”

Li and co-investigators received a grant from the National Cancer Institute in 2022 and developed a tool that uses CT scans and AI to detect minute changes in the pancreas, which is part of the digestive system. These changes can indicate a patient is likely to develop pancreatic cancer years before they develop the disease, thus enabling an earlier diagnosis when treatments are most effective. This investigative team has received a new grant to conduct a pilot study that will allow them to determine whether those characteristics differ in Black patients.

The team will collaborate with Cemal Yazici, MD, associate professor of Medicine at the University of Illinois Chicago, which, like Cedars-Sinai, serves a large population of Black patients.

Predictive models for pancreatic cancer based on symptoms, genetics and weight haven’t been shown to be accurate on their own, said Stephen Pandol, MD, director of Basic and Translational Pancreas Research at Cedars-Sinai and co-principal investigator of the study. Meanwhile, blood and urine tests for pancreas cancer are still in development.

“Packaging these tests with our imaging tool could eventually prove effective in predicting who will develop pancreatic cancer,” Pandol said. “Our goal is to perfect our tool, then apply it to a population to see if we can diagnose individuals earlier.”

Pancreatic cancer is one of the deadliest cancers, with only around 10% of patients surviving five years after diagnosis. Early detection and surgical intervention can increase five-year survival rates to more than 50%, but this is rare because early pancreatic cancer causes few symptoms.

“Drilling down to understand cancer risk at precision levels is key to the Cedars-Sinai Cancer mission to bring personalized cancer care to every patient,” said Dan Theodorescu, MD, PhD, director of Cedars-Sinai Cancer and the PHASE ONE Foundation Distinguished Chair. "This work in early detection nicely complements recent work showing that a new AI-based precision medicine approach called the Molecular Twin Precision Oncology Platform can predict outcomes of pancreatic cancer."

With the help of improved predictive tools, patients at the highest risk can be monitored via annual imaging so any cancer that develops is detected at the earliest stage.

“If we can diagnose a patient’s pancreatic cancer two or three years earlier, that could make a life-and-death difference,” Li said.