AI-Driven Optimization of PCL/PEG Electrospun Scaffolds for Enhanced In Vivo Wound Healing

Abstract

Here, an artificial intelligence (AI)-based approach was employed to optimize the production of electrospun scaffolds for in vivo wound healing applications. By combining polycaprolactone (PCL) and poly(ethylene glycol) (PEG) in various concentration ratios, dissolved in chloroform (CHCl₃) and dimethylformamide (DMF), 125 different polymer combinations were created. From these polymer combinations, electrospun nanofiber meshes were produced and characterized structurally and mechanically via microscopic techniques, including chemical composition and fiber diameter determination. Subsequently, these data were used to train a neural network, creating an AI model to predict the optimal scaffold production solution. Guided by the predictions and experimental outcomes of the AI model, the most promising scaffold for further in vitro analyses was identified. Moreover, we enriched this selected polymer combination by incorporating antibiotics, aiming to develop electrospun nanofiber scaffolds tailored for in vivo wound healing applications. Our study underscores three noteworthy conclusions: (i) the application of AI is pivotal in the fields of material and biomedical sciences, (ii) our methodology provides an effective blueprint for the initial screening of biomedical materials, and (iii) electrospun PCL/PEG antibiotic-bearing scaffolds exhibit outstanding results in promoting neoangiogenesis and facilitating in vivo wound treatment.