A cardiologist-like computer-aided interpretation framework to improve arrhythmia diagnosis from imbalanced training datasets

Arrhythmias can pose a significant threat to cardiac health, potentially leading to serious consequences such as stroke, heart failure, cardiac arrest, shock, and sudden death. In computer-aided electrocardiogram interpretation systems, the inclusion of certain classes of arrhythmias, which we term “aggressive” or “bullying,” can lead to the underdiagnosis of other “vulnerable” classes. To address this issue, a method for arrhythmia diagnosis is proposed in this study. This method combines morphological-characteristic-based waveform clustering with Bayesian theory, drawing inspiration from the diagnostic reasoning of experienced cardiologists. The proposed method achieved optimal performance in macro-recall and macro-precision through hyperparameter optimization, including spliced heartbeats and clusters. In addition, with increasing bullying by aggressive arrhythmias, our model obtained the highest average recall and the lowest average drop in recall on the nine vulnerable arrhythmias. Furthermore, the maximum cluster characteristics were found to be consistent with established arrhythmia diagnostic criteria, lending interpretability to the proposed method. [Display omitted] •Bullying from aggressive arrhythmias against vulnerable arrhythmias was observed•A cardiologist-like computer-aided interpretation framework of ECG is proposed•Morphological-characteristic-based waveform clustering was used for ECG data encoding•The diagnostic experience of cardiologists was simulated by Bayesian theory Medical professionals are increasingly looking to AI-based methods to help interpret medical data and provide more accurate or timely diagnoses for patients. Current AI methods, however, may not always perform consistently for different patient sub-groups or disease sub-types. Underlying dataset imbalances and inhomogeneous inter-class similarity often cause these issues. Underrepresented classes may be less well learned by the AI model, and, in some cases, the presence of particular classes in a dataset can actually interfere with the AI’s ability to learn other classes. In this paper, the authors study this issue in detail in the context of diagnosing heart arrhythmias, a common and sometimes life-threatening cardiac disorder, and show that an AI framework that mimics the reasoning of experienced cardiologists can better diagnose arrhythmia sub-types that are sensitive to this kind of interference. The computer-aided interpretation system of ECG is an important tool