Integrating machine-learning and nanotechnology to quantify pH-modulated oxaliplatin release.

The purpose of this work was to formulate and characterize pH-sensitive, surfactant-based nanomicelles for the targeted delivery of Oxaliplatin to breast cancer cells. A secondary aim was to utilize machine learning (ML) models to interpolate and dissect the sophisticated, pH-dependent drug release kinetics. Nanomicelles of Oxaliplatin were prepared by Pluronic F-127 through a thin-film hydration technique. The nanomicelles were characterized for size, morphology, encapsulation efficiency, and drug release profile in physiological (pH 7.4) and acidic, tumor-mimicking (pH 5.4) media. The MTT assay was used to test cytotoxicity against L929 normal fibroblasts and MCF-7 breast cancer cells. ML models (Random Forest, Gradient Boosting, SVR) were trained on experimental release data to anticipate crucial release phase changes using SHAP analysis. Prepared nanomicelles were monodisperse and spherical with hydrodynamic diameter 290.3 nm and encapsulation efficiency 40.2%. They had good pH-responsive release with cumulative release 77.5% and 43.5% at pH 5.4 and 7.4, respectively, in 96 h. Kinetic modeling revealed a shift from Fickian diffusion at pH 7.4 to anomalous transport at pH 5.4. ML models showed great interpolation performance (R² > 0.97), and SHAP analysis showed remarkable release transitions. The cytotoxicity assays were different from free Oxaliplatin with improved activity against MCF-7 cells and lower toxicity against L929 cells. Surfactant-based nanomicelles are an effective delivery platform for pH-directed delivery of Oxaliplatin to enhance its therapeutic index. Nanomedicine formulation design is enabled by ML through comprehensive release kinetics analysis.