AI Model IDs Pancreatic Ductal Adenocarcinoma at Prediagnostic Stage

FRIDAY, May 1, 2026 (HealthDay News) -- An artificial intelligence (AI) model surpasses radiologists for detecting pancreatic ductal adenocarcinoma (PDA) at its visually occult prediagnostic stage, according to a study published online April 28 in Gut.Sovanlal Mukherjee, Ph.D., from the Mayo Clinic in Rochester, Minnesota, and colleagues developed and validated an AI framework to identify subvisual radiomic signatures of prediagnostic PDA on standard-of-care computed tomography (Radiomics-based Early Detection MODel [REDMOD]). The model was trained on a multi-institutional cohort of 969 patients (156 prediagnostic, 813 control) and was tested on an independent set of 493 patients (63 prediagnostic, 430 control), simulating a low-prevalence, early detection paradigm. AI-driven segmentation was coupled with a heterogeneous ensemble architecture trained on a 40-feature radiomic signature. Validation included direct comparison with radiologists, test-retest analysis, and external specificity validation in two independent cohorts with 539 and 80 participants.The researchers found that REDMOD identified occult PDA on an independent test set, with an area under the receiver operating characteristic curve (AUC) of 0.82 and 73.0 percent sensitivity, at a median 475-day lead time. This represented a nearly twofold higher sensitivity than seen for radiologists (38.9 percent), which increased at >24 months lead time (68.0 versus 23.0 percent). Strong longitudinal stability was seen with REDMOD (90 to 92 percent concordance), and specificity was generalizable across multi-institutional and public datasets (81.3 and 87.5 percent, respectively)."This AI can now identify the signature of cancer from a normal-appearing pancreas, and it can do so reliably over time and across diverse clinical settings," senior author Ajit Harishkumar Goenka, M.D., also from the Mayo Clinic, said in a statement.Abstract/Full Text.Sign up for our weekly HealthDay newsletter

AI Model IDs Pancreatic Ductal Adenocarcinoma at Prediagnostic Stage

Related Stories

Model Provides Well-Calibrated Risk Estimates for First Metachronous cSCC

Metabolic Syndrome Tied To Cancer Risk

Machine Learning Predicts Asthma Risk in Children With Early-Life Atopic Dermatitis

Elraglusib/GnP Improves Survival in Metastatic Pancreatic Adenocarcinoma