Paper Title: Automated diagnosis of chronic obstructive pulmonary disease using deep learning applied to electrocardiograms
Journal: eBioMedicine, Volume 123, January 2026
Authors: Monica Kraft, MD, Health System Chair of the Department of Medicine; Girish N. Nadkarni, MD, MPH, CPH, Chief AI Officer and Chair of the Department of Artificial Intelligence and Human Health; Akhil Vaid, MD, Assistant Professor of Medicine (Data Driven Digital Medicine) at the Icahn School of Medicine at Mount Sinai and the Mount Sinai Health System; and other coauthors.
Bottom Line: Chronic Obstructive Pulmonary Disease (COPD) is a leading cause of morbidity and mortality globally. Effective management hinges on early diagnosis, which is often impeded by non-specific symptoms and resource-intensive diagnostic methods. This study assesses the effectiveness of electrocardiograms (ECGs) analyzed via deep learning as a tool for early COPD detection.
How: Mount Sinai researchers utilized a Convolutional Neural Network model to analyze ECGs, or medical tests that record the heart's electrical activity, and can detect COPD. The primary outcome was the accuracy of a new clinical COPD diagnosis, as determined by International Classification of Disease codes. They performed an evaluation using Area-Under-the-Curve (AUC) metrics, derived by testing against ECGs from patients at five hospitals within the Mount Sinai Health System who represented a demographically diverse patient population in New York City. They examined data from 2006 to 2023 within the GE MUSE system that exports electrocardiograms as individual XML files containing raw waveforms. The experts also used ECGs from patients at another hospital and ECGs of patients with COPD within the UK BioBank to expand the cohort and validate the analysis.
Findings: Mount Sinai researchers analyzed more than 208,000 ECGs from over 18,000 COPD cases, matched to more than 49,000 controls by age, sex, and race. The model exhibited robust performance across diverse populations with an AUC of 0⋅80 (0⋅80–0⋅80) in internal testing, 0⋅82 (0⋅81–0⋅82) in external validation, and 0⋅75 (0⋅71–0⋅78) in the UK cohort. Subsequent analyses linked ECG-derived model predictions with spirometry data, and model explainability highlighted P-wave changes, an indication of COPD.
Why the Research Is Interesting: Experts at Mount Sinai found that AI-powered ECG analysis offers a promising path for early COPD detection, potentially facilitating earlier and more effective management of the condition. Implementing such machine learning tools in clinical settings could significantly enhance COPD screenings and diagnostic accuracy, thereby improving patient outcomes and addressing global health burdens of the disease.
Said Mount Sinai's Dr. Monica Kraft of the research:
"Our study is the first to demonstrate that deep learning models applied to standard 10-second, 12-lead ECGs can accurately detect Chronic Obstructive Pulmonary Disease (COPD) across large, real-world patient cohorts. Using the model known as a convolutional neural network, we show that ECGs—a low-cost and widely available tool—can capture COPD-related physiological changes, including those that precede formal clinical diagnosis.
Unlike previous exploratory work, our analysis includes external validation across distinct cohorts over various times and locations, as well as analysis in the subgroup categories of irregular heartbeat and smoking exposure."
Said Mount Sinai's Dr. Girish Nadkarni of the research:
"By demonstrating that AI can enhance the diagnostic utility of ECGs for COPD, a pathway is opened for earlier intervention and management of this disease, potentially reducing the severity of its progression and associated financial cost burdens. The use of such AI-enhanced diagnostic tools can be expanded to remote or under-resourced areas where access to specialized diagnostic facilities might be limited.
Additionally, this study lays the groundwork for future research into the integration of AI technologies with other routine diagnostic tools—possibly improving the diagnostic accuracy and timeliness for a range of chronic conditions and ultimately enhancing prevention and early intervention."
This study was supported by the Scientific Computing and Data at the Icahn School of Medicine at Mount Sinai; the Clinical and Translational Science Awards (UL1TR004419) from the National Center for Advancing Translational Sciences; and the National Heart, Lung, and Blood Institute (R01HL167050-02).
For more Mount Sinai artificial intelligence news, visit: https://icahn.mssm.edu/about/artificial-intelligence .
