The researchers used data from the medical records of patients in both the United States and Denmark from 1977 to 2020. They looked at a cohort of 6.2 million Danish patients, 23,985 of whom had pancreatic cancer, and 3 million military personnel undergoing treatment. Through Veterans Affairs, 3,864 of them were eventually identified.
The researchers used a machine learning model to analyze the data, teaching it to predict cancer risk based on symptoms and various diagnostic codes in patients’ medical records.
Some of the symptoms associated with a high-risk prognosis are not traditionally linked to pancreatic cancer. Gallstones, type 2 diabetes, anemia and gastrointestinal symptoms such as vomiting and abdominal pain were all linked to a higher risk score three years before diagnosis.
In a real-world scenario, the researchers write, the AI model would develop pancreatic cancer in 320 of every 1,000 people identified as being at high risk. By targeting surveillance to high-risk patients, the tool could make screening more affordable, they write.
Currently, the US Preventive Services Task Force does not recommendation Screening asymptomatic people for pancreatic cancer. Screening of high-risk patients related to However, there is a higher chance of long-term survival.
“An AI tool that can zero in on people who are at high risk for pancreatic cancer, and who can benefit from more tests, could go a long way toward improving clinical decision-making,” said Chris Sander, co-author of the study. Harvard Medical School Laboratory In one message, biology is dedicated to using machine learning and other technologies to solve problems liberation.
Used at scale, such a tool could extend lifespan and improve treatment outcomes, Sander said.
