Assessment of Drug Proarrhythmicity Using Artificial Neural Networks With in silico Deterministic Model Outputs

Abstract

As part of the Comprehensive in vitro Proarrhythmia Assay initiative, methodologies for predicting the occurrence of drug-induced torsade de pointes via computer simulations have been developed and verified recently. However, their predictive performance still requires improvement. Herein, we propose an artificial neural networks (ANN) model that uses nine multiple input features, considering the action potential morphology, calcium transient morphology, and charge features to further improve the performance of drug toxicity evaluation. The voltage clamp experimental data for 28 drugs were augmented to 2,000 data entries using an uncertainty quantification technique. By applying these data to the modified O’Hara Rudy in silico model, nine features (dVm/dtmax, APresting, APD90, APD50, Caresting, CaD90, CaD50, qNet, and qInward) were calculated. These nine features were used as inputs to an ANN model to classify drug toxicity into high-risk, intermediate-risk, and low-risk groups. The model was trained with data from 12 drugs and tested using the data of the remaining 16 drugs. The proposed ANN model demonstrated an AUC of 0.92 in the high-risk group, 0.83 in the intermediate-risk group, and 0.98 in the low-risk group. This was higher than the classification performance of the method proposed in previous studies.

Proarrhythmic risk assessment of drugs by dVm/dt shapes using the convolutional neural network

“…The proposed deep CNN model can achieve better classification performance than the classical models that evaluate drug toxicity by focusing on the ion channel change caused by drugs. In addition, it showed excellent results, exceeding the performance of our other study—Yoo et al developed the ANN model using nine in silico features considering the morphological information of the AP trace and transient calcium trace, including the qNet and qInward, which are the ion net charge features 30 …”

“…In addition, it showed excellent results, exceeding the performance of our other study-Yoo et al developed the ANN model using nine in silico features considering the morphological information of the AP trace and transient calcium trace, including the qNet and qInward, which are the ion net charge features. 30 In several studies, the dV m /dt was identified to be helpful to detect the occurrence of TdP or the drug effects. 31,32 Because when drugs make a blockage in the ion channels, the dV m /dt in the depolarization phase is reduced; Passini et al observed the main effect of lidocaine and mexiletine to the peak sodium through a decreased maximal dV m /dt.…”

“…In addition, it showed excellent results, exceeding the performance of our other study—Yoo et al developed the ANN model using nine in silico features considering the morphological information of the AP trace and transient calcium trace, including the qNet and qInward, which are the ion net charge features. 30 …”

Action Potential Morphology Accurately Predicts Proarrhythmic Risk for Drugs With Potential to Prolong Cardiac Repolarization

“…A general issue with in silico based classifiers, such as that described here and those published as part of CiPA 16 and in other studies, 19,20 is the significant cost, time and technical burden of in vitro data collection required to inform the models-that is, the measures of potency and binding kinetics against multiple cardiac ion channels. The measurement of kinetics, in particular, has proven difficult to implement, 47 particularly on high-throughput patch clamp platforms.…”

Application of Machine Learning to Improve the Efficiency of Electrophysiological Simulations Used for the Prediction of Drug-Induced Ventricular Arrhythmia

“…Several works have been published on the implementation of the CiPA based in silico simulations for the prediction of ventricular arrhythmia and TdP biomarkers using predicted or experimentally determined drug-induced ion channel inhibition data [11][12][13][14][15][16][17][18][19] . Computational models of human and animal electrophysiology operate at different biological levels, ranging from a single channel to whole tissue simulations and vary in terms of the degree of complexity and abstraction, the underlying mathematical approaches, and physiological parameters [11] .…”