Carrier optimization of pulmonary powder systems with using computational intelligence tools

Abstract

Efficient delivery of drug particles to the respiratory tract determines therapeutic efficiency of pulmonary drugs. Only particles with an aerodynamic diameter (dae) below five micrometers (μm) deposit in deep lungs. However fine particles, because of their cohesive nature, frequently cause difficulties related to the manufacturing process and stability of the product. This study is focused on the optimization of carriers for pulmonary drug delivery systems with the use of empirical models. Advanced computational intelligence tools were applied to produce a mathematical formula able to predict fine particle fraction (FPF) for a particular formulation. FPF is a mass percentage of drug particles with an aerodynamic diameter below 5 μm. The best model was characterized by normalized root mean squared error (NRMSE) below 8% and R2 = 0.85. The goal of the in silico optimization was to find the possible directions of carrier modification in order to maximize FPF. Obtained results were applied to the laboratory experiments and resulted in the two new formulations with bovine serum albumin (BSA) as a model drug. Experimental results confirmed the model predictions, and the formulation composed of the carrier with higher bulk density and surface skewness resulted in greater FPF value. The presented work is an example of computational intelligence tools implementation and particles surface assessment based on SEM images in design of the decision support systems for the development of pulmonary powder formulations.