Introduction

Traditional toxicokinetic (TK) models rely heavily on in vivo data, necessitating animal testing. At the same time, the scientific toolbox is expanding with new approach methodologies (NAMs) that do not rely on TK studies. One promising strategy (functional similarity1 , TK-equivalence, or read-across2) involves leveraging known in vivo data of a reference compound to develop and validate a TK model for a target compound without additional animal data. Of course, this assumes the absence of activity cliffs. Here, we compare three methods (structure, TK inputs, and TK outputs) to find a reference compound (nearest neighbor (NN)) for the target compound. The similarity of the Cp-time curves of each pair, defined as a distance, is used to judge success.

By Michael Lawless, Rafał A. Bachorz, and Priyata Kalra

SOT 64th Annual Meeting and ToxExpo, March 16-20, 2025, Orlando, Florida