Development and In Vivo evaluation of liposomal fentanyl nanocarriers using thin-film hydration and AI-based characterization for enhanced analgesic efficacy in anesthesia

The aim of this study was to develop and evaluate a liposomal formulation of fentanyl to improve its analgesic efficacy and safety profile in anesthesia, using artificial intelligence (AI) techniques to characterize and optimize the formulation. The liposomes were prepared using the thin-film hydration method, resulting in vesicles with an average diameter of 120.4 ± 10.2 nm and a polydispersity index of 0.23 ± 0.05. The encapsulation efficiency of fentanyl was determined by high performance liquid chromatography (HPLC) and was 85.0 ± 3.2%. A convolutional neural network (CNN) implemented in TensorFlow/Keras (Python) was used for automated analysis of scanning electron microscopy (SEM) images to evaluate the morphology of liposomes. In vitro drug release was studied over 12 hours using a dialysis method. The analgesic efficacy and safety of the liposomal formulation was evaluated in a rat pain model by comparing the paw withdrawal latency (PWL) and mechanical thresholds of Frey with standard fentanyl. A Gaussian Process Regression (GPR) model developed using scikit-learn (Python) was used to predict liposome properties based on formulation parameters. CNN analysis confirmed predominantly spherical liposomes with 3.2% aggregates and 1.8% broken particles. In vitro release studies showed sustained release of fentanyl with an initial burst of approximately 10%. In vivo, the liposomal fentanyl group showed significantly prolonged analgesia (PWL: 12.4 ± 2.1 seconds) compared to standard fentanyl (8.5 ± 1.5 seconds, p < 0.05), with no observed respiratory depression or sedation. The GPR model demonstrated strong predictive performance (R² > 0.85) for key formulation characteristics. The novelty of this work lies in the integration of AI-driven characterization (CNN for SEM) and predictive modeling (GPR) into the development and optimization of liposomal fentanyl nanocarriers, enabling objective morphological analysis and data-driven formulation design. The developed liposomal fentanyl formulation provides sustained drug release, increased analgesic efficacy and improved safety compared to standard fentanyl. This AI-enhanced approach offers a promising strategy for optimizing liposomal drug delivery systems in anesthesia and pain management.