Uncertainty assessment of proarrhythmia predictions derived from multi-level in silico models.

In silico methods can be used for an early assessment of arrhythmogenic properties of drug candidates. However, their use for decision-making is conditioned by the possibility to estimate the predictions' uncertainty. This work describes our efforts to develop uncertainty quantification methods for the predictions produced by multi-level proarrhythmia models. In silico models used in this field usually start with experimental or predicted IC50 values that describe drug-induced ion channel blockade. Using such inputs, an electrophysiological model computes how the ion channel inhibition, exerted by a drug in a certain concentration, translates to an altered shape and duration of the action potential in cardiac cells, which can be represented as arrhythmogenic risk biomarkers such as the APD90. Using this framework, we identify the main sources of aleatory and epistemic uncertainties and propose a method based on probabilistic simulations that replaces single-point estimates predicted using multiple input values, including the IC50s and the electrophysiological parameters, by distributions of values. Two selected variability types associated with these inputs are then propagated through the multi-level model to estimate their impact on the uncertainty levels in the output, expressed by means of intervals. The proposed approach yields single predictions of arrhythmogenic risk biomarkers together with value intervals, providing a more comprehensive and realistic description of drug effects on a human population. The methodology was tested by predicting arrhythmogenic biomarkers on a series of twelve well-characterised marketed drugs, belonging to different arrhythmogenic risk classes.

Application of machine learning to improve the efficiency of electrophysiological simulations used for the prediction of drug-induced ventricular arrhythmia.

BACKGROUND AND OBJECTIVE: In silico prediction of drug-induced ventricular arrhythmia often requires computationally intensive simulations, making its application tedious and non-interactive. This inconvenience can be mitigated using matrices of precomputed simulation results, allowing instantaneous computation of biomarkers such as action potential duration at 90% of the repolarisation (APD90). However, preparing such matrices can be computationally intensive for the method developers, limiting the range of simulated conditions. In this work, we aim to optimise the generation of these matrices so that they can be obtained with less effort and for a broader range of input values. METHODS: Machine learning methods were applied, building models trained with only a small fraction of the originally simulated results. The predictive performances of the models were assessed by comparing their predicted values with the actual simulation results, using percentual mean absolute error and mean relative error, as well as the percentage of data with a relative error below 5%. RESULTS: Our method obtained highly accurate estimations of the original values, leading to a nearly one hundred-fold decrease in computation time. This method also allows precomputing more complex matrices, describing the effect of more ion channels on the APD90. The best results were obtained by applying Support Vector Machine models, which yielded errors below 1% in most cases. This approach was further validated by predicting the APD90 of a set of 12 CiPA compounds and exporting the optimal settings for predicting APD90 using a different set of ion channels, always with satisfactory results. CONCLUSIONS: The proposed method effectively reduces the computational effort required to generate matrices of precomputed electrophysiological simulation values. The same approach can be applied in other fields where computationally costly simulations are applied repeatedly using slightly different input values.