Design and optimization of indomethacin-loaded solid dispersion orally dissolving films: Machine learning and molecular dynamics study

Indomethacin (IND) is a nonsteroidal anti-inflammatory drug (NSAID) with anti-inflammatory, analgesic, and antipyretic properties, and is widely used for the treatment of diverse inflammatory diseases. However, its poor aqueous solubility severely limits its clinical application. An orally dissolving film (ODF) loaded with an indomethacin solid dispersion (IND-SD) was developed to enhance dissolution. Based on Hansen solubility parameters (HSP), four polymers were selected to prepare the IND-SD, and the optimal carrier and drug to polymer ratio were identified by in vitro dissolution testing. Molecular docking and molecular dynamics (MD) simulations were employed to elucidate drug-polymer interactions at the molecular level. Under the guidance of Quality by Design (QbD), an optimization framework integrating a Box-Behnken design (BBD) and an artificial neural network (ANN) was established to design and optimize the ODF formulation. Multiple statistical metrics were used to assess the Box-Behnken design response surface methodology (BBD-RSM) model and the ANN model, with the ANN model demonstrating superior predictive accuracy in predicting the film critical quality attributes (CQAs). PXRD and DSC analyses confirmed that IND existed in an amorphous state in both the IND-SD and the ODF. In vitro dissolution experiments demonstrated that the cumulative drug release from the ODF in simulated saliva within 1 min was significantly higher than that of pure IND and IND-SD.