A simple approach discriminating cardio-safe drugs from toxic ones

Effects of Trastuzumab Emtansine (T-DM1) on QT Interval and Safety of Pertuzumab Plus T-DM1 in Patients With Previously Treated Human Epidermal Growth Factor Receptor 2-Positive Metastatic Breast Cancer

Trastuzumab emtansine (T-DM1) is an antibody-drug conjugate in development for human epidermal growth factor receptor 2 (HER2)-positive cancer. Drugs in development are generally tested for their effects on QT interval, prolongation of which is associated with the potentially fatal arrhythmia torsades de pointes. In addition, an association between left ventricular dysfunction and other HER2-directed agents has been documented. This multicenter, phase 2 study, TDM4688g, assessed the safety and pharmacokinetic characteristics of T-DM1 (3.6 mg/kg every 3 weeks) in patients with previously treated HER2-positive metastatic breast cancer, and the safety of pertuzumab plus T-DM1, an anti-HER2 extracellular domain antibody, in patients with early disease progression on T-DM1 alone. The primary end point was the change in QTc interval from baseline to each postbaseline time point, adjusted for heart rate using Fridericia's correction. T-DM1 had no clinically relevant effect on QTc interval. The observed upper limit of the one-sided 95% confidence interval was below the 10-millisecond threshold of safety concern. The safety and efficacy of single-agent T-DM1 was consistent with that observed in previous studies. Pertuzumab plus T-DM1 was generally well tolerated with no new safety signals. These results support further investigation of T-DM1 as a single agent and with pertuzumab.

Multiple Kernel Learning for Drug Discovery

The support vector machine (SVM) methodology has become a popular and well-used component of present chemometric analysis. We assess a relatively recent development of the algorithm, multiple kernel learning (MKL), on published structure-property relationship (SPR) data. The MKL algorithm learns a weighting across multiple kernel-based representations of the data during supervised classifier creation and, thereby, may be used to describe the influence of distinct groups of structural descriptors upon a single structure-property classifier without explicitly omitting any of them. We observe a statistically significant performance improvement over a conventional, single kernel SVM on all three SPR data sets analysed. Furthermore, MKL output is observed to provide useful information regarding the relative influence of five distinct descriptor subsets present in each data set.