Priego Lucia, Mora Maria Teresa, Llopis-Lorente Jordi, Finsberg Henrik, Daversin-Catty Cecile, Van Herck Ilsbeth, Wall Samuel, Arevalo Hermenegild, Saiz Francisco Javier, Trenor Beatriz
Centro de Innovación e Investigación en Bioingeniería, Universitat Politècnica de València, Valencia, España.
Simula Research Laboratory, Oslo, Norway.
Comput Methods Programs Biomed. 2025 Sep;269:108896. doi: 10.1016/j.cmpb.2025.108896. Epub 2025 Jun 4.
Cardiac drug safety assessment is essential to detect molecules with potential adverse effects, particularly those increasing the risk of arrhythmias such as Torsade de Pointes (TdP). The challenge lies in achieving early predictions with minimal experimental studies while maintaining high sensitivity and specificity. Traditional approaches rely primarily on electrophysiological (EP) biomarkers; however, mechanical effects of drugs on cardiac contractility remain underexplored. This study aims to integrate electrophysiological and mechanical biomarkers to improve cardiac risk assessment models.
The present study investigated the integration of electrophysiological and mechanical biomarkers using a cellular and three-dimensional in silico population approach. We evaluated the effects on different EP and mechanical relevant biomarkers to assess both, the proarrhythmic and inotropic risks of 39 compounds, including CiPA compounds and calcium channel blockers (CCBs). Classification models were developed using EP biomarkers alone and in combination with mechanical biomarkers to evaluate their predictive capabilities.
Classification models based solely on EP biomarkers demonstrated robust prediction on CiPA torsadogenic risk. The inclusion of mechanical biomarkers did not enhance classification accuracy for TdP risk. However, mechanical metrics revealed significant contractile changes induced by CCBs and other negative inotropic compounds, such as mavacamten. Drugs induced a range of fractional shortening values, correlated with ejection fraction variations, highlighting clinically relevant contractile effects.
EP-based assessments remain reliable for predicting torsadogenic risk, but mechanical biomarkers provide crucial insights into drug-induced cardiac contractile effects. Future studies should focus on increasing experimental data availability and incorporating more complex cardiac geometries to enhance translational applicability of in silico models in comprehensive drug safety evaluation.
心脏药物安全性评估对于检测具有潜在不良反应的分子至关重要，尤其是那些增加心律失常风险的分子，如尖端扭转型室速（TdP）。挑战在于在进行最少的实验研究的同时实现早期预测，同时保持高灵敏度和特异性。传统方法主要依赖电生理（EP）生物标志物；然而，药物对心脏收缩性的机械作用仍未得到充分探索。本研究旨在整合电生理和机械生物标志物，以改进心脏风险评估模型。
本研究使用细胞和三维计算机模拟群体方法研究电生理和机械生物标志物的整合。我们评估了对不同的EP和机械相关生物标志物的影响，以评估39种化合物（包括CiPA化合物和钙通道阻滞剂（CCB））的促心律失常和变力风险。使用单独的EP生物标志物以及与机械生物标志物相结合的方式开发分类模型，以评估它们的预测能力。
仅基于EP生物标志物的分类模型对CiPA致TdP风险表现出强大的预测能力。纳入机械生物标志物并未提高TdP风险的分类准确性。然而，机械指标显示CCB和其他负性变力化合物（如mavacamten）引起了显著的收缩变化。药物诱导了一系列的缩短分数值，与射血分数变化相关，突出了临床上相关的收缩效应。
基于EP的评估对于预测致TdP风险仍然可靠，但机械生物标志物为药物诱导的心脏收缩效应提供了关键见解。未来的研究应专注于增加实验数据的可用性，并纳入更复杂的心脏几何形状，以提高计算机模拟模型在综合药物安全性评估中的转化适用性。
