Fine-tuning AI-ECG models on homogeneous cardiology department data achieved high AF prediction AUCs (0.829-0.885), outperforming models trained on heterogeneous data.
Does fine-tuning AI-enhanced ECG models on homogeneous data improve the prediction of paroxysmal atrial fibrillation from sinus rhythm across different institutions and devices?
191,783 sinus rhythm ECGs from Tokai University (n=172,613) and The Cardiovascular Institute (n=19,170)
AI-enhanced ECG model fine-tuned on homogeneous datasets from cardiology departments (Model F2)
Models trained from scratch or fine-tuned on heterogeneous datasets from all departments (Model F1)
Area under the receiver operating characteristic curve (AUC) for detecting paroxysmal atrial fibrillation from sinus rhythm ECGssurrogate
Fine-tuning AI-enhanced ECG models on homogeneous data from cardiology departments improves performance and generalizability for predicting paroxysmal atrial fibrillation from sinus rhythm across different institutions and devices.
ABSTRACT Background Artificial intelligence (AI)‐enhanced electrocardiography (ECG) has been developed to detect paroxysmal atrial fibrillation (AF) from sinus rhythm ECGs (SR‐ECGs). For broader applicability, model development across institutions and ECG systems is essential. Methods We developed an AI‐enhanced ECG model using SR‐ECGs from Tokai University ( n = 172,613; Nihon Kohden NK system) and The Cardiovascular Institute ( n = 19,170; GE MUSE system). AF‐labeled SR‐ECGs were defined as recordings within 31 days of an AF episode, while SR‐labeled SR‐ECGs were those with ≥ 1095 days of AF‐free follow‐up. Three datasets were constructed: Dataset 1 (Tokai University, all departments, NK), Dataset 2 (Tokai University, Cardiology Department, NK), and Dataset 3 (The Cardiovascular Institute, Cardiology Department, MUSE). We developed five models: scratch models (S1–S3) trained on Datasets 1–3, and fine‐tuned models (F1, F2) trained on Datasets 1 and 2 after pretraining on Dataset 3. Models were evaluated using A1–A3 (same as Datasets 1–3) and B1–B3, which differed in ECG resolution and compression (B1: original MUSE, B2: MUSE‐NK intermediate, B3: NK‐converted). Results Model F2, fine‐tuned on homogeneous datasets from cardiology departments, showed consistently high performance (AUC: A1 = 0.885, A2 = 0.829, A3 = 0.845). Model F1, fine‐tuned on heterogeneous datasets, demonstrated lower performance (AUC: A1 = 0.837, A2 = 0.726, A3 = 0.660). Model performance was consistent across different ECG format variants (B1–B3). Conclusion Fine‐tuning on homogeneous data improved performance and generalizability, whereas heterogeneous data led to reduced performance. ECG system format differences had minimal impact on model accuracy.
Building similarity graph...
Analyzing shared references across papers
Loading...
Shinya Suzuki
Mari Amino
Nobumoto Moriai
Annals of Noninvasive Electrocardiology
Tokai University
Cardiovascular Institute Hospital
Nihon Kohden (Japan)
Building similarity graph...
Analyzing shared references across papers
Loading...
Suzuki et al. (Mon,) reported a other. Fine-tuning AI-ECG models on homogeneous cardiology department data achieved high AF prediction AUCs (0.829-0.885), outperforming models trained on heterogeneous data.
www.synapsesocial.com/papers/698c1c73267fb587c655ef71 — DOI: https://doi.org/10.1111/anec.70159