Artificial neural network-based pore size prediction of alginate gel scaffold for targeted drug delivery

Pore size distribution of polymeric gel scaffold and the corresponding porosity are key elements influencing their drug release profile in several targeted drug delivery applications. This research aimed to build a network model that can successfully determine the pore size distribution of the alginate gel scaffold fabricated using microfluidic techniques. Scaffold properties controlling the average pore diameter, porosity, and yield stress identified as sodium alginate concentration and surface activity were computed by measuring the viscosity, surface tension, and contact angle, respectively. A mathematical model was established using an artificial neural network (ANN) comprising alginate concentration, viscosity, surface tension, and contact angle as input and average pore diameter, porosity, and yield stress as output. The proposed model successfully determined the pore size distribution of gel scaffold covering a range of 70-400 mm, fabricated by various combinations of sodium alginate and pluronic F-127 and compared with random forest (RF), support vector regression (SVR) and k-nearest neighbor (kNN) models. The surface activity of gel scaffold was estimated to have an utmost impact on pore size followed by viscosity and finally contact angle. The results of this work efficaciously emphasized the scaffold properties affecting pore size distribution and established the efficacy of ANN in predicting the average pore diameter, porosity, and yield stress of polymeric gel scaffolds.