Investigation of the effects of process variables on derived properties of spray dried solid-dispersions using polymer based response surface model and ensemble artificial neural network models.

The objective of this study was to use different statistical tools to understand and optimize the spray drying process to prepare solid dispersions. In this study we investigated the relationship between input variables (inlet temperature, feed concentration, flow rate, solvent and atomization parameters) and quality attributes (yield, outlet temperature and mean particle size) of spray dried solid dispersions (SSDs) using response surface model and ensemble artificial neural network. The Box Behnken design was developed to investigate the effect of various input variables on quality attributes of final products. Moreover, Pearson correlation analysis, self organizing map, contour plots and response surface plot were used to illustrate the relationship between input variables and quality attributes. The influence of different physicochemical properties of solvent on the quality attributes of spray dried products was also investigated. Final validation of prepared models was done using binary SSDs of six model drugs with PVP. Results demonstrated the effectiveness of proposed PVP based model which can help scientists to gain detailed understanding of spray drying process of solid dispersion using minimal resources and time during early formulation development stage. It will also help them to ensure consistent quality of SSDs using broad range of input variables.