By Dedong Wu, Ph.D.
Advanced Drug Delivery, PharmSci, R&D,
AstraZeneca, Boston US
dedong.wu@astrazeneca.com

It is challenging and laborious to identify a crystallization process to access an optimal drug substance form, usually the most thermodynamically stable crystalline form. By the conventional experimental approach, there is always a risk to miss a more stable form early on and discover it during further crystallization experiments. In silico solid-state modeling provides an opportunity to mitigate the risk in drug substance form development. By employing emerging technology of crystal structure prediction (CSP), a “virtual polymorph screening” approach allows the scientists to confirm whether the most stable crystalline form is identified in crystallization experiments. When the predicted most stable form is not observed experimentally, predictions of ‘conformer weighting in solution’ based on COSMO-RS method (Conductor like Screening Model for Real Solvents) highlighted solvent systems to increase the likelihood of the most stable form’s conformation, making it possible for a “targeted crystallization” process to produce the desired solid form. The presentation will provide a showcase of a rapid form selection and crystallization process development by using a minimal amount of sample within a shortened time period. Furthermore, it will provide scientific rational to ensure the most stable form is selected in early stage of pharmaceutical R&D in order to lower the risk for future pharmaceutical development.