Pancreatic cancer risk prediction using deep sequential modeling of longitudinal diagnostic and medication records.

Summary

A new artificial intelligence model identifies individuals at high risk for pancreatic cancer by analyzing sequences of medical diagnoses and medication histories. This approach allows for the detection of high-risk patients far more accurately than age and sex alone, offering a potential pathway for earlier diagnosis and improved survival rates.

Key result

A transformer-based model identifying the top 1,000 to 5,000 highest-risk patients among 1 million individuals demonstrated a 3-year pancreatic ductal adenocarcinoma incidence 70 to 115 times higher than age- and sex-based estimates.

Abstract

Pancreatic ductal adenocarcinoma (PDAC) is a rare, aggressive cancer often diagnosed late with low survival rates, due to the lack of population-wide screening programs and the high cost of early detection methods. To enable early detection of high-risk individuals, we develop a transformer-based model trained on longitudinal Veterans Affairs electronic health record (EHR) with 19,426 PDAC cases and ∼15.9 million controls. Our model combines diagnostic and medication trajectories to predict PDAC risk within a 6-, 12-, and 36-month assessment window. Incorporating medication significantly improved performance; among the top 1,000-5,000 highest-risk patients in a cohort of 1 million patients, 3-year PDAC incidence is 115-70 times higher than a reference estimate based on age and sex alone. Furthermore, analysis of most predictive features highlights the role of events such as chronic inflammatory conditions and specific medications on overall PDAC risk. Our work provides an AI-driven identification of high-risk individuals, with a potential to improve early detection, enhance patient care, and reduce healthcare costs.

Key figure