A population-based in silico platform for arrhythmia prediction

项目总结

PROJECT SUMMARY Ventricular arrhythmias are the leading cause of sudden cardiac death accounting for ~300,000 deaths per year in the US alone. However, no pharmacological or biological therapy has yet emerged with comparable efficacy to the implantable cardioverter-defibrillator. The major hurdles are: 1) at the individual scale, arrhythmias have multiple and multiscale causes and mechanisms. Drugs target entities at the molecular scale but arrhythmias are fundamentally tissue-scale phenomena, with no simple one-to-one relationships due to complex multiscale nonlinear interactions. An antiarrhythmic drug may suppress one particular arrhythmia mechanism but potentiate another mechanism, unexpectedly increasing rather than decreasing mortality as shown in large clinical trials; and 2) at the population scale, a drug may be antiarrhythmic for one individual but proarrhythmic for another due to inter-individual variability/diversity and complex environmental differences, which may also account for the failure of current antiarrhythmic drug therapies. Therefore, for antiarrhythmic drug discovery, one must evaluate the effects of a molecular intervention or a drug on not just a single arrhythmia mechanism, but all possible arrhythmia mechanisms. Additionally, one must take into account inter-individual variability and complex environmental stresses. An equally important and crucial problem is effective proarrhythmia risk (cardiotoxicity) screening for drug safety. In the past, ~30% of the drugs removed from the market were due to their proarrhythmia risk. Owing to the extreme complexity of the problem, computer modeling and simulation will be required to evaluate a drug's antiarrhythmic and proarrhythmic effects. Recently, the Cardiac Safety Research Consortium and FDA have recommended computer simulation as a complementary approach for proarrhythmia drug screening. However, traditional modeling approaches are limited and population-based modeling approaches are required. Moreover, the model populations need to accurately account for the inter-individual variability and arrhythmia mechanisms for accurate prediction of a drug's antiarrhythmic and proarrhythmic effects. This project proposes to develop a novel in silico platform which includes multiscale normal and diseased model populations emulating the inter-individual variability of human populations. The model populations will be filtered and validated against clinical data under normal and diseased conditions. “Virtual clinical trials” will be then performed for antiarrhythmic drug discovery and drug safety screening. The specific aims are: 1) to develop and validate an in silico platform incorporating model populations that emulate the inter-individual variability of human populations under normal and disease conditions; 2) to utilize the in silico human model populations as a platform for novel antiarrhythmic drug discovery and cardiotoxicity screening. This is a data-driven in silico approach which integrates computational modeling, experimental human heart data, and clinical data to translate computational modeling to clinical medicine.