Design and methods of the AUTOMATED-WCT trial: evaluating machine learning-based ECG support for WCT interpretation.

Area of Science:

• Cardiology
• Medical Technology
• Artificial Intelligence

Background:

• Differentiating wide complex tachycardia (WCT) into ventricular tachycardia (VT) or supraventricular WCT (SWCT) is crucial but difficult.
• Current computerized ECG interpretation systems often provide limited diagnostic information for WCT.
• Novel machine learning (ML) models offer potential for improved WCT classification but require clinical validation.

Purpose of the Study:

• To assess the diagnostic performance of ML-based ECG interpretation tools for WCT.
• To compare ML models against standard clinical practice in a randomized trial.
• To evaluate the impact of ML decision support on physician diagnostic accuracy and confidence.

Main Methods:

• A prospective, multicenter, four-arm randomized reader trial.
• Physicians will be randomized to interpret WCT ECGs with or without ML model outputs (Solo Model, Paired Model).
• Participants will classify WCT, rate confidence, perceived usefulness, and intended management.

Main Results:

• The primary endpoint is WCT classification accuracy.
• Secondary endpoints include sensitivity, specificity, positive predictive value, negative predictive value, F1 score, time to diagnosis, confidence, perceived usefulness, and intended management.
• This trial will generate the first randomized, multicenter evidence on ML-based ECG decision support for WCT differentiation.

Conclusions:

• The AUTOMATED-WCT Trial is the first randomized trial to evaluate ML-based ECG decision support for WCT.
• Findings will inform the integration of AI tools into clinical workflows for managing WCT.
• This research aims to improve the accuracy and efficiency of diagnosing potentially life-threatening cardiac arrhythmias.