Application of Response Surface and Artificial Neural Network Simulation to Improve Dissolution of Poorly Soluble Methotrexate Via Microwave Assisted Solid Dispersions

Methotrexate, a poorly water-soluble BCS Class II NSAID, exhibits limited oral bioavailability due to its low solubility. This study aimed to enhance its solubility and dissolution rate by formulating solid dispersions using hydrophilic polymers via a microwave-assisted fusion technique.

An initial screening of polymers like Hydroxypropyl methylcellulose E5 (HPMC E5), Eudragit EPO (poly(methacrylate) copolymer), Polyethene glycol 6000 (PEG 6000), Polyvinylpyrrolidone (PVP), and Soluplus was conducted. PEG 6000 and Soluplus were selected for further investigation based on their low melting points and high wettability, which are suitable for the fusion method. Solid dispersions were prepared using both conventional and microwave fusion techniques at various drug-to-polymer ratios. Phase solubility studies in 0.1 N HCl (pH 1.2) were performed, and the drug, physical mixture and formulations were characterized by FTIR, DSC, and XRD. A Central composite design(CCD) was employed to optimize microwave power and exposure time for the MW solid dispersion of methotrexate with Soluplus for the formulation of an immediate release methotrexate tablet. Comparative modelling using Artificial Neural Networks (ANN) and Design of Experiments (DoE) was conducted for dissolution prediction.

The optimal 1:2 drug-to-Soluplus ratio achieved a 3.07-fold increase in solubility, compared to a 2.55-fold enhancement with PEG 6000. Characterization studies confirmed uniform drug dispersion and partial conversion to the amorphous form. The prepared 13 batches via CCD of Microwave solid dispersion was compressed into tablet form and Final optimised batch demonstrated significantly improved drug release (59.91% at 15 min, 88.66% at 45 min, and 97.34% at 60 min) with no significant changes during stability studies. The ANN model exhibited superior predictive accuracy for dissolution profiles with lower RMSE values (Y1: 0.0004, Y2: 0.0007, Y3: 0.0003) compared to DoE.

Microwave-assisted fusion using hydrophilic polymers, particularly Soluplus, markedly improved the dissolution rate and potential bioavailability of methotrexate. The optimized solid dispersion, when formulated as a tablet, demonstrated excellent drug release properties. ANN modeling proved to be an effective tool for predictive formulation optimization.