AG13736 (Axitinib), an inhibitor of vascular endothelial growth factor (VEGF) under investigation as an oncology drug, is currently manufactured via a three-step process that utilizes two palladium-mediated cross-couplings. Historically, removal of residual heavy metals from the active pharmaceutical ingredient has been a persistent issue. The development of a much improved process for palladium removal and a useful screening technique developed to rapidly identify the most efficient reagents for this purpose are outlined. The performance of the new endgame process in pilot-plant scale-up is also discussed.

A simple and efficient screening methodology for scavenger reagent identification was carried out to overcome the high levels of residual palladium present in the Axitinib process. The purification strategy relied on crystallization of the desired product from the crude reaction mixture leaving the residual palladium in the mother liquor by using Pdchelating agents. Theoretical models based on naïve Bayesian classifier were also performed to support the screening team. Using the best model, a virtual screening of commercially available compounds was conducted and new theoretical hits were generated. Some of these theoretical ligands were then tested experimentally, confirming the validity of our model.