Abstract

Deep learning (DL) models can harness electronic health records (EHRs) to predict diseases and extract radiologic findings for diagnosis. With ambulatory chest radiographs (CXRs) frequently ordered, we investigated detecting type 2 diabetes (T2D) by combining radiographic and EHR data using a DL model. Our model, developed from 271,065 CXRs and 160,244 patients, was tested on a prospective dataset of 9,943 CXRs. Here we show the model effectively detected T2D with a ROC AUC of 0.84 and a 16% prevalence. The algorithm flagged 1,381 cases (14%) as suspicious for T2D. External validation at a distinct institution yielded a ROC AUC of 0.77, with 5% of patients subsequently diagnosed with T2D. Explainable AI techniques revealed correlations between specific adiposity measures and high predictivity, suggesting CXRs’ potential for enhanced T2D screening.

Traditional type 2 diabetes (T2D) screening methods often depend on age, BMI guidelines and glucose measurements. Here, authors use a deep learning model that leverages chest radiographs and electronic health record data to screen for T2D, highlighting potential for early detection and intervention.

Details

Title
Opportunistic detection of type 2 diabetes using deep learning from frontal chest radiographs
Author
Pyrros, Ayis 1 ; Borstelmann, Stephen M. 2 ; Mantravadi, Ramana 3 ; Zaiman, Zachary 4 ; Thomas, Kaesha 4 ; Price, Brandon 5   VIAFID ORCID Logo  ; Greenstein, Eugene 6 ; Siddiqui, Nasir 7 ; Willis, Melinda 7 ; Shulhan, Ihar 8   VIAFID ORCID Logo  ; Hines-Shah, John 7 ; Horowitz, Jeanne M. 9 ; Nikolaidis, Paul 9 ; Lungren, Matthew P. 10 ; Rodríguez-Fernández, Jorge Mario 11   VIAFID ORCID Logo  ; Gichoya, Judy Wawira 4   VIAFID ORCID Logo  ; Koyejo, Sanmi 12   VIAFID ORCID Logo  ; Flanders, Adam E 13 ; Khandwala, Nishith 14 ; Gupta, Amit 15 ; Garrett, John W. 16   VIAFID ORCID Logo  ; Cohen, Joseph Paul 17 ; Layden, Brian T. 18 ; Pickhardt, Perry J. 16 ; Galanter, William 18 

1  Duly Health and Care, Department of Radiology, Downers Grove, USA; University of Illinois Chicago, Department of Biomedical and Health Information Sciences, Chicago, USA (GRID:grid.185648.6) (ISNI:0000 0001 2175 0319) 
2  University of Central Florida, Department of Radiology, Orlando, USA (GRID:grid.170430.1) (ISNI:0000 0001 2159 2859) 
3  Brainnet, Inc., West Harrison, USA (GRID:grid.170430.1) 
4  Emory University, Department of Radiology, Atlanta, USA (GRID:grid.189967.8) (ISNI:0000 0004 1936 7398) 
5  Florida State University, Department of Radiology, Tallahassee, USA (GRID:grid.255986.5) (ISNI:0000 0004 0472 0419) 
6  Duly Health and Care, Department of Cardiology, Downers Grove, USA (GRID:grid.255986.5) 
7  Duly Health and Care, Department of Radiology, Downers Grove, USA (GRID:grid.255986.5) 
8  EPAM, Inc, Newtown, USA (GRID:grid.255986.5) 
9  Northwestern University, Department of Radiology, Chicago, USA (GRID:grid.16753.36) (ISNI:0000 0001 2299 3507) 
10  UCSF, Department of Biomedical and Health Information Sciences, San Francisco, USA (GRID:grid.266102.1) (ISNI:0000 0001 2297 6811); Stanford University, Center for Artificial Intelligence in Medicine, Stanford, USA (GRID:grid.168010.e) (ISNI:0000 0004 1936 8956); Microsoft Corporation, Microsoft, Redmond, USA (GRID:grid.419815.0) (ISNI:0000 0001 2181 3404) 
11  The University of Texas Medical Branch, Department of Neurology, Galveston, USA (GRID:grid.176731.5) (ISNI:0000 0001 1547 9964) 
12  Stanford University, Department of Computer Science, Stanford, USA (GRID:grid.168010.e) (ISNI:0000 0004 1936 8956) 
13  Thomas Jefferson University, Department of Radiology, Philadelphia, USA (GRID:grid.265008.9) (ISNI:0000 0001 2166 5843) 
14  Bunkerhill, Palo Alto, USA (GRID:grid.265008.9) 
15  University Hospitals Cleveland Medical Center, Department of Radiology, Cleveland, USA (GRID:grid.443867.a) (ISNI:0000 0000 9149 4843) 
16  University of Wisconsin, Department of Radiology, Madison, USA (GRID:grid.14003.36) (ISNI:0000 0001 2167 3675) 
17  Stanford University, Center for Artificial Intelligence in Medicine, Stanford, USA (GRID:grid.168010.e) (ISNI:0000 0004 1936 8956) 
18  University of Illinois Chicago, Department of Medicine, Chicago, USA (GRID:grid.185648.6) (ISNI:0000 0001 2175 0319) 
Pages
4039
Publication year
2023
Publication date
2023
Publisher
Nature Publishing Group
e-ISSN
20411723
Source type
Scholarly Journal
Language of publication
English
DOI
https://doi.org/10.1038/s41467-023-39631-x
ProQuest document ID
2834368622
Copyright
© The Author(s) 2023. corrected publication 2024. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.