AI-enhanced X-ray microscopy for non-destructive detection, quantification, and particle size analysis of crystalline miconazole in amorphous solid dispersions

Abstract

In the pharmaceutical industry, amorphous solid dispersions (ASDs) are commonly utilized to enhance the solubility and bioavailability of poorly soluble active pharmaceutical ingredients (APIs). Polymers such as copovidone are also frequently employed as stabilizing agents in ASDs due to their ability to reduce the thermodynamical instability of the amorphous form compared to their crystalline form. Detecting and measuring any crystalline API particles present in the ASD formulation is critical for ensuring drug efficacy and stability. In this study, ASD tablets miconazole in a copovidone matrix spiked with known concentration of crystalline miconazole were characterized by X-ray microscopy (XRM). We demonstrate how XRM combined with AI-assisted image segmentation can provide quantitative characterization of crystalline particles, including detailed particle size distribution (PSD) information. The AI-assisted image processing model was found to be robust, reducing human error associated with traditional threshold selection. Additionally, we evaluated particle size changes from the blend to the final drug product to understand how the API and excipient behave during tablet manufacturing. This approach has broader applications in drug product development, particularly in monitoring recrystallization and evaluating process-induced changes during blending and compression.