ChatGPT’s Limitations in Athlete ECG Interpretation: Evidence from a Multicenter Diagnostic Study
Journal of Cardiovascular Development and Disease · 29.04.2026 · 13 (5) · 191
DOI: 10.3390/jcdd13050191 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Cardiovascular Development and Disease, online erschienen am 29.04.2026. Verfasst von Stefano Palermi und 40 weiteren Autor:innen. Beteiligte Einrichtung: Department of Medicine and Surgery, UniCamillus Saint Camillus International University of Health Sciences, 00131 Rome, Italy. Die Arbeit wurde bisher 8-mal zitiert (Crossref) und nennt 24 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Background
Artificial intelligence (AI) has shown promise in the interpretation of electrocardiograms (ECGs) using signal-based deep learning models. In parallel, large language models (LLMs) have gained increasing visibility in clinical practice, including exploratory applications in ECG analysis. Whether a general-purpose LLM can meaningfully discriminate cardiovascular disease from athlete ECGs during PPS remains unknown. We aimed to evaluate the diagnostic performance of a general-purpose LLM for this task.
Methods
In this multicentre diagnostic accuracy study, we evaluated a commercially available LLM (ChatGPT, version 5) in 2950 competitive athletes undergoing PPS. All athletes underwent resting 12-lead ECG, with second- and third-line investigations performed when clinically indicated. The reference outcome was confirmed cardiovascular disease after full diagnostic work-up (n = 450, 15.3%). For each ECG, the LLM generated a numeric score (0–100) representing the inferred likelihood of underlying disease using a standardized prompt and without task-specific fine-tuning. Discriminative performance was assessed using receiver operating characteristic (ROC) analysis. Misclassification patterns were analysed according to International ECG Criteria.
Results
GPT-derived scores demonstrated a marked floor effect, with a median value of 0 (IQR 0–2) in both diseased and non-diseased athletes and substantial overlap between groups. The area under the ROC curve was 0.52 (95% CI 0.49–0.55), indicating performance close to random classification. At the Youden-derived threshold, 79% of athletes with confirmed disease were incorrectly classified as negative. False-negative cases were predominantly characterized by borderline ECG patterns (82%), and a substantial number of red-flag ECG abnormalities were also missed.
Conclusions
In this PPS cohort, a general-purpose LLM used in a naïve configuration showed no clinically meaningful ability to discriminate between cardiovascular disease and athlete ECGs. Without task-specific training or domain adaptation, such models should not be used for diagnostic triage in athlete screening.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Medicine and Surgery, UniCamillus Saint Camillus International University of Health Sciences, 00131 Rome, Italy.
- Sports and Exercise Medicine Division, Department of Medicine, University of Padova, 35122 Padova, Italy.
- Division of Cardiology, Cardiovascular and Thoracic Department, Città della Salute e della Scienza di Torino Hospital, 10126 Turin, Italy.
- Department of Medical Sciences, University of Turin, 10124 Turin, Italy.
- Bambino Gesù Children's Hospital, IRCCS, 00165 Rome, Italy.
- Med-Ex, Medicine & Exercise, Medical Partner Scuderia Ferrari, 00187 Rome, Italy.
ORCID-Profile: Stefano Palermi, Marco Vecchiato, Matteo Anselmino, Rachele Adorisio, Francesco Borrelli, Nicoletta Cantarutti, Elena Cavarretta, Mattia Cominacini, Marco Corsi, Flavio D'Ascenzi, Giuseppe Di Gioia, Gianluigi Dorelli, Giulia Foccardi, Sabina Gallina, Francesca Graziano, Elisa Lodi, Viviana Maestrini, Davide Mansour, Maria Grazia Modena, Daniel Neunhaeuserer, Andrea Palermi, Alessio Pellegrino, Filippo Maria Quattrini, Fabrizio Ricci, Alessandro Zorzi, Andrea Saglietto
Zitieren
APA
Palermi, S., Vecchiato, M., Iacovone, T. R., Anselmino, M., Adorisio, R., Biffi, A., Borrelli, F., Brugin, E., Cantarutti, N., Cavarretta, E., Cominacini, M., Corsi, M., D'Ascenzi, F., De Feo, V., Di Gioia, G., Dorelli, G., Foccardi, G., Gallina, S., Giangrandi, S., …, & Saglietto, A. (2026). ChatGPT’s Limitations in Athlete ECG Interpretation: Evidence from a Multicenter Diagnostic Study. Journal of Cardiovascular Development and Disease, 13(5), 191. https://doi.org/10.3390/jcdd13050191
Vancouver
Palermi S, Vecchiato M, Iacovone TR, Anselmino M, Adorisio R, Biffi A, et al. ChatGPT’s Limitations in Athlete ECG Interpretation: Evidence from a Multicenter Diagnostic Study. Journal of Cardiovascular Development and Disease. 2026;13(5):191. doi:10.3390/jcdd13050191
BibTeX
@article{Palermi2026,
author = {Palermi, Stefano and Vecchiato, Marco and Iacovone, Tommaso Remo and Anselmino, Matteo and Adorisio, Rachele and Biffi, Alessandro and Borrelli, Francesco and Brugin, Erica and Cantarutti, Nicoletta and Cavarretta, Elena and Cominacini, Mattia and Corsi, Marco and D'Ascenzi, Flavio and De Feo, Vittorio and Di Gioia, Giuseppe and Dorelli, Gianluigi and Foccardi, Giulia and Gallina, Sabina and Giangrandi, Silvia and Graziano, Francesca and Lodi, Elisa and Livio, Alberto and Maestrini, Viviana and Manfredi, Guglielmo Leonardo and Mansour, Davide and Modena, Maria Grazia and Neunhaeuserer, Daniel and Nigro, Antonia and Palermi, Andrea and Pellegrino, Alessio and Pelliccia, Antonio and Quattrini, Filippo Maria and Ricci, Fabrizio and Scarzella, Fiammetta and Squeo, Maria Rosaria and Tonelli, Riccardo and Zanardo, Emanuele and Zorzi, Alessandro and D'Ascenzo, Fabrizio and De Ferrari, Gaetano Maria and Saglietto, Andrea},
title = {ChatGPT’s Limitations in Athlete ECG Interpretation: Evidence from a Multicenter Diagnostic Study},
journal = {Journal of Cardiovascular Development and Disease},
year = {2026},
volume = {13},
number = {5},
pages = {191},
doi = {10.3390/jcdd13050191}
}
EUVIMED