The Use of ROC Analysis for the Qualitative Prediction of Human Oral Bioavailability from Animal Data

To develop and evaluate a tool for the qualitative prediction of human oral bioavailability (F-human) from animal oral bioavailability (F-animal) data employing ROC analysis and to identify the optimal thresholds for such predictions. A dataset of 184 compounds with known F-human and F-animal in at least one species (mouse, rat, dog and non-human primates (NHP)) was employed. A binary classification model for F-human was built by setting a threshold for high/low F-human at 50%. The thresholds for high/low F-animal were varied from 0 to 100 to generate the ROC curves. Optimal thresholds were derived from 'cost analysis' and the outcomes with respect to false negative and false positive predictions were analyzed against the BDDCS class distributions. We successfully built ROC curves for the combined dataset and per individual species. Optimal F-animal thresholds were found to be 67% (mouse), 22% (rat), 58% (dog), 35% (NHP) and 47% (combined dataset). No significant trends were observed when sub-categorizing the outcomes by the BDDCS. F-animal can predict high/low F-human with adequate sensitivity and specificity. This methodology and associated thresholds can be employed as part of decisions related to planning necessary studies during development of new drug candidates and lead selection.