Cardiovascular Safety Pharmacology using Collaborative Biosignal Analytics

ExEcutivE Summary Cardiotoxicity screening is in a state of flux. It is evident that screening for hERG block in immortalized cell lines and QT interval prolongation in animals as recommended by the ICH S7B guidelines is limited in predicting a drug's risk of causing the potentially fatal arrhythmia, Torsades de Pointes (TdP). The " Thorough QT study, " required by the FDA/ICH E14 safety guidelines to resolve discrepancies between in vitro and in vivo studies, is also under scrutiny. Most importantly, the pharmaceutical industry and regulatory agencies are considering human stem cell-derived cardiomyocytes (hSC-CMs) as a standard preparation for preclinical safety screening. Microelectrode arrays (MEAs) seem ideal for testing drug effects on the hSC-CM field potential duration (FPD), the in vitro correlate of the QT interval, and thus predicting the clinical QT interval's response to a drug. Significantly, low-impedance MEAs' enhanced sensitivity enables screening for actual proarrhythmic phenomena such as early afterdepolarizations (EADs), the initiating mechanism of TdP. MEAs produce large volumes of heterogeneous data, especially with medium-and high-throughput platforms. Analysis can be extremely time consuming and subject to observer bias. In addition, drug development requires preclinical in vivo studies in non-rodent animals as well as human testing. These studies can be undertaken using different hardware and software platforms resulting in large amounts of data in different file formats, although the biosignals are all similar. A scalable, automated software platform with reusable, customizable analysis templates reduces subjectivity and increases throughput to make the most of these state-of-the-art technologies and facilitate harmonization of drug development within a pharmaceutical enterprise. Neural ID's Intelligent Waveform Service (IWS) pattern-recognition technology combines human expertise with machine-learning to recognize waveforms and their nuances, including EADs. Via a user-friendly, show-and-go interface, users " train " the software to recognize example patterns. Training creates a knowledge base, which then can be run against multiple file sets for dose-response, repeated-dose, and long-term experiments. Built-in queries (SQL) can be used to generate reports for parameters of interest such as beating frequency, QT interval, and FPD. Analysis templates are shareable between researchers and groups, allowing for accurate unbiased auditing of data and quality control. Notably, drug programs discarded because of a hERG signal can be revisited and repurposed, now that we have a better understanding of the mechanism of drug-induced arrhythmia and TdP. This white paper describes how IWS can maximize the productivity and predictivity of MEA/hSC-CM data analysis in an …