A novel practical algorithm using machine learning to differentiate outflow tract ventricular arrhythmia origins

概要

Diagnosis of the origin of ventricular tachycardia (VT) and premature ventricular contractions
(PVCs) from electrocardiograms (ECGs) is crucial in catheter ablation for idiopathic VT and
PVCs. In particular, the outflow tract ventricular arrhythmia (OTVA) with a precordial
transition in lead V3 (V3TZ) has been recognized as one of the most difficult QRS
morphologies to differentiate the origin.1 Although several algorithms for the diagnosis of the
origin of OTVA with V3TZ have been developed so far,2-14 the accuracy is not definitive with
limited performance in each criterion.15,16 To address this challenge, we aimed to propose a
new algorithm to differentiate the left ventricular outflow tract (LVOT) from the right
ventricular outflow tract (RVOT) of OTVA origin with V3TZ using artificial intelligence (AI).
The importance of AI has been recognized in various fields of clinical medicine, and the
arrhythmia treatment is no exception. However, the results and process of the AI algorithm are
sometimes difficult to interpret, and are also difficult to use in the clinical setting. In this study,
we analyzed several ECG parameters including previously reported criteria and evaluated
useful indicators in the clinical setting. We performed a decision tree analysis in machine
learning, which is highly useful and comprehensive for interpreting the decision process and
results, to create the best algorithm for the diagnosis of the origin of OTVA with V3TZ. ...

この論文で使われている画像