Arrhythmias & electrophysiology

Enhancing origin prediction: deep learning model for diagnosing premature ventricular contractions with dual-rhythm analysis focused on cardiac rotation

EuropacePublished Diagnostic accuracy study

Authors' conclusion

Our deep learning-based model, incorporating both PVC and SR morphologies, resulted in a higher prediction accuracy for PVC origin, considering SR is particularly important for predicting right-sided origin in patients with counterclockwise rotation.

Study context

Population
Patients with inferior axis premature ventricular contractions (PVCs), 11 centres (Japan, Germany; Belgium external validation)100 external validation
Participants
593 patients
Comparison
Dual-rhythm model vs PVC model
Outcome
Prediction of PVC origin (right vs left side) on 12-lead ECG
Methodological detailsmulticentre, external validation, Hybrid Resnet50-based CNN and transformer model
  • Multicentre
  • External validation
  • Hybrid Resnet50-based CNN and transformer model

Performance

As reported in the abstract

PVC origin prediction, external validation data set
Test or modelAUCSpecificityAccuracyPrecisionF1-score
Dual-rhythm model0.910.870.840.730.77
PVC model0.860.680.740.550.68

AUC

Dual-rhythm model0.91
PVC model0.86

Specificity

Dual-rhythm model0.87
PVC model0.68

Accuracy

Dual-rhythm model0.84
PVC model0.74

Precision

Dual-rhythm model0.73
PVC model0.55

F1-score

Dual-rhythm model0.77
PVC model0.68
  • Dual-rhythm model vs PVC model: P < 0.01 — accuracy
  • Dual-rhythm model vs PVC model: P < 0.01 — precision
  • Dual-rhythm model vs PVC model: P < 0.01 — specificity
  • Dual-rhythm model vs PVC model: P = 0.03 — AUC
  • Dual-rhythm model vs PVC model: P = 0.03 — F1-score

Further findings

Contribution to PVC origin prediction

PVC contribution
77.3%
Sinus rhythm (SR) contribution
22.7%
In counterclockwise rotation, SR had greater contribution in predicting right-sided PVC origin

Original publication

Enhancing origin prediction: deep learning model for diagnosing premature ventricular contractions with dual-rhythm analysis focused on cardiac rotation. Europace. Published 13 September 2024. doi:10.1093/europace/euae240

Composed by AI from the published abstract and checked against it. How Briefs are made →

Get them every week

New studies on this topic in My Week and in your Monday e-mail.

We e-mail you a sign-in link and a code. No password, no spam.