Key result
Spectro-temporal ECG transformer detects MI with ~91% accuracy, while retinal imaging underperforms standard baselines.
Why the study?
Claims of near-perfect accuracy from multimodal fusion of retinal fundus images and ECG for cardiovascular risk assessment are frequently confounded by data leakage and evaluation on datasets lacking the stated modalities.
Does a transformer-based deep learning framework using twelve-lead ECGs and retinal fundus images improve myocardial infarction detection and carotid intima-media-thickness-defined risk prediction compared to standard machine learning and demographic baselines under a leakage-free protocol?
Observational (n=2,903)
Does a transformer-based deep learning framework using twelve-lead ECGs and retinal fundus images improve myocardial infarction detection and carotid intima-media-thickness-defined risk prediction compared to standard machine learning and demographic baselines under a leakage-free protocol?
Under a strict leakage-free evaluation protocol, deep learning via a transformer model provides strong, reproducible myocardial infarction detection from ECGs, but retinal fundus imaging fails to outperform basic demographic data (age and sex) for predicting carotid intima-media-thickness risk.
Answers are grounded in this paper and cite their sources.
ECG transformer MI detection performance does not extend to retinal models or validated multimodal fusion; leaves open clinical translation.
Background: Deep learning on retinal fundus images and on the electrocardiogram (ECG) has each shown promise for cardiovascular risk assessment, and multimodal fusion is often claimed to reach near-perfect accuracy. Such claims are frequently confounded by data leakage and by evaluation on data that does not actually contain the stated modalities. Methods: We re-implemented a trans-former-based multimodal framework and evaluated it under a strict leakage-free protocol on two public datasets: PTB-XL (21,799 twelve-lead ECGs; myocardial-infarction vs. normal, 1513 held-out test ECGs on the official patient-respecting fold) and China-Fundus-CIMT (2,903 patients, carotid intima-media-thickness-defined risk; three-fold patient-level cross-validation). A spectro-temporal ECG transformer and a retinal vision transformer were trained independently and combined by decision-level fusion; all preprocessing, model selection and thresholds were confined to training folds. Each modality was compared with strong baselines. Results: The ECG transformer reached 90.75% accuracy and an AUC of 0.970 (sensitivity 88.91%, specificity 91.80%), exceeding lo-gistic-regression, random-forest and gradient-boosting baselines (AUC 0.835-0.869). The retinal transformer reached 71.34+/-4.57% accuracy and an AUC of 0.764+/-0.016; notably, an age-plus-sex logistic baseline (AUC 0.854) outperformed the image model, indicating limited added value of the fundus images for this label. Because the two cohorts share no patients, fusion could only be studied by simulation and did not constitute per-patient validation. Conclusions: Under rigorous evalua-tion, ECG-based MI detection is strong and reproducible, whereas retinal CIMT-risk prediction is modest and does not beat demographics, and genuine multimodal fusion will be investigated in future research.
No takes yet. Share an insight, caveat, or question.
Jinkathoti et al. (2026) conducted an observational in Cardiovascular risk and myocardial infarction (n=2,903). Spectro-temporal ECG transformer and retinal vision transformer vs. Logistic-regression, random-forest, gradient-boosting, and age-plus-sex baselines was evaluated on Myocardial infarction detection (ECG) and carotid intima-media-thickness-defined risk prediction (retinal) (AUC 0.970 (ECG); AUC 0.764 (retinal)). The spectro-temporal ECG transformer achieved 90.75% accuracy for MI detection, while the retinal transformer underperformed an age-plus-sex baseline.
Synapse has enriched 10 related papers. Consider them for comparative context: