Derivation of the Stochastic Model for Predicting Mechanical Performance Using Electrical Patterns During Ventricular Tachy-arrhythmia

Da Un Jeong and Ki Moo Lim
Kumoh National Institute of Technology


Aims: During ventricular tachyarrhythmia, it is difficult to predict the mechanical performance. Because the mechanical behavior of the myocardial tissue at this time is very irregular and contracts asynchronously at a high frequency. In this study, we derived the stochastic models for predicting mechanical performance using representative electrical phenomenological parameters during tachyarrhythmia. Methods: We implemented 96 kinds of various ventricular tachyarrhythmia by using four reentry generation methods and by changing the conduction characteristics of potassium channel. The mechanical response due to complex electrical patterns were obtained from a deterministic model of excitation-contraction coupling simulation. We derived the stochastic model for predicting mechanical response using single- and multi-regression analysis. We used action potential duration (APD), dominant frequency (DF), phase singularity (PS) and filament as the independent variables representing the electrical phenomenon, and stroke volume (SV) and the amplitude of myocardial tension (ampTens) as the dependent variables representing the mechanical response. Results: In single-regression model, the accuracies of stochastically predicted SV and ampTens were the highest at 68% when using APD and at 88% when using DF, respectively. It is impossible to predict the mechanical response using the multi-regression model with four electrical independent variables due to multi-collinearity of APD and DF. Therefore, we performed multi-regression analysis using three electrical parameters except for APD, which has the highest variance inflation factor. As a results, the accuracy of the stochastically predicted SV was 65%, which is lower than that of single-regression model. However, the accuracy of predicted ampTens was 89.3%, which is higher than that of single-regression model. Conclusion: We showed the possibility of predicting the mechanical performance during ventricular tachyarrhythmia using the derived stochastic model from deterministic simulation. More realistic predictions of mechanical behavior are possible given the variety of phenomena that affect the electrical activity of the ventricular tissue.