Online Program

Decision Tree–Based Approaches for Classification with Application to Ex Vivo Study (302399)

Qianyu Dang, US Food and Drug Administration 
*Yu-yi Hsu, FDA/CDER/OB/DB6 
Dalong Patrick Huang, FDA/CDER  
Nan Xi, UCLA 

Keywords: classification, ordinal data, BART

A type of ex vivo study, rabbit ventricular wedge assay (RVWA), was proposed to be used to inform the early development decision and assist the evaluation of potential risk. Multiple ECG intervals and an ordinal variable representing the early afterdepolarization related incidence are recorded from each sample in a blinded validation study. This research project uses a decision tree based model to predict proarrhythmia based from RVWA results.A Bayesian additive regression tree (BART) model was used to analyze data from the RVWA validation study. Each drug is assigned to a risk category based on the predicted probabilities. The model uses a regularization prior to summarize information from multiple binary prediction trees.