Auditable CT Phenotyping Through Report-derived Radiological Observations
Abstract
Medical image foundation models can predict clinical phenotypes from computed tomography (CT), but strong performance leaves open whether they read disease-specific findings or shortcuts that correlate with the diagnosis. We tested this in 221 electronic-health-record (EHR) phenotypes using Auditable CT phenotyping (ACT), built on report-derived radiological observations. We trained ACT on 38,317 patients, mined 376,194 observations and evaluated it in 25,183 held-out patients. ACT exceeded five vision-language baselines on zero-shot annotation, and CT-CLIP across 221 phenotypes from unseen CT pulmonary angiography, both under zero-shot scoring (0.651 versus 0.572) and under linear probing (0.709 versus 0.662). Reading each probe exposes what accuracy conceals: only 97 observations occupy the 221 rank-1 positions, and one phrase describing aortic and coronary calcification ranks first for 20 phenotypes, including osteoporosis, urinary tract infection and major depressive disorder. Restricting the bank to clinician-specified evidence redirects those probes onto phenotype-related observations in 86 phenotypes at no accuracy cost (0.751 versus 0.741). Accurate CT-based EHR phenotyping can therefore rest on observations that are not valid evidence for the coded phenotype and that ACT can identify and intervene on.
Authors 12
-
Affiliation as printed
Artificial Intelligence in Biomedical Imaging Laboratory (AIBIL) , Perelman School of Medicine , University of Pennsylvania , Philadelphia , PA , USA
-
Affiliation as printed
Department of Radiology , Perelman School of Medicine , University of Pennsylvania , Philadelphia , PA , USA
-
Affiliation as printed
Department of Radiology , Perelman School of Medicine , University of Pennsylvania , Philadelphia , PA , USA
-
Universitätsklinikum Aachen · RWTH Aachen University
Affiliation as printed
Department of Diagnostic and Interventional Radiology , University Hospital RWTH Aachen , Aachen , Germany
-
TUM Klinikum · Technical University of Munich
Affiliation as printed
Department of Diagnostic and Interventional Radiology , School of Medicine and Health , Klinikum rechts der Isar , TUM University Hospital , Technical University of Munich , Munich , Germany
-
Affiliation as printed
Department of Biostatistics, Epidemiology and Informatics , Perelman School of Medicine , University of Pennsylvania , Philadelphia , PA , USA
Division of Pulmonary, Allergy and Critical Care , Department of Medicine , Perelman School of Medicine , University of Pennsylvania , Philadelphia , PA , USA
-
Affiliation as printed
Department of Biostatistics, Epidemiology and Informatics , Perelman School of Medicine , University of Pennsylvania , Philadelphia , PA , USA
-
Affiliation as printed
Artificial Intelligence in Biomedical Imaging Laboratory (AIBIL) , Perelman School of Medicine , University of Pennsylvania , Philadelphia , PA , USA
Center for Artificial Intelligence and Data Science for Integrated Diagnostics (AI2D) , Perelman School of Medicine , University of Pennsylvania , Philadelphia , PA , USA
-
Affiliation as printed
Department of Radiology , Perelman School of Medicine , University of Pennsylvania , Philadelphia , PA , USA
-
Universitätsklinikum Aachen · RWTH Aachen University
Affiliation as printed
Department of Diagnostic and Interventional Radiology , University Hospital RWTH Aachen , Aachen , Germany
-
Affiliation as printed
Center for Artificial Intelligence and Data Science for Integrated Diagnostics (AI2D) , Perelman School of Medicine , University of Pennsylvania , Philadelphia , PA , USA
Department of Radiology , Perelman School of Medicine , University of Pennsylvania , Philadelphia , PA , USA
Cited by 0 stored of 0
References 8
8 results