Computerized electrocardiogram data transformation enables effective algorithmic differentiation of wide QRS complex...

Area of Science:

• Cardiology
• Medical Informatics
• Artificial Intelligence in Medicine

Background:

• Wide QRS complex tachycardia (WCT) requires accurate differentiation into ventricular tachycardia (VT) and supraventricular wide complex tachycardia (SWCT).
• Computerized electrocardiogram (ECG) analysis offers a potential solution for automated WCT classification.

Purpose of the Study:

• To develop and validate novel approaches for differentiating WCT using ECG data.
• To assess the efficacy of WCT differentiation with and without a corresponding baseline ECG.

Main Methods:

• Five classification models (logistic regression, artificial neural network, Random Forests, support vector machine, ensemble learning) were evaluated on a derivation cohort.
• Two logistic regression models were prospectively validated: one using WCT ECG alone (Solo Model) and another using paired WCT and baseline ECGs (Paired Model).

Main Results:

• All models in the derivation cohort demonstrated high performance (AUC > 0.96).
• In the validation cohort, the Solo Model achieved AUCs of 0.87 and 0.84, while the Paired Model achieved AUCs of 0.95 for both patient groups (with and without baseline ECGs).

Conclusions:

• Accurate WCT differentiation is achievable using computerized ECG data.
• Both WCT ECG analysis alone and paired WCT/baseline ECG analysis provide effective methods for WCT classification.