Machine Learning-Guided microfluidic optimization of clinically inspired liposomes for nanomedicine applications

Liposomes have transformed drug delivery by enhancing the solubility, stability, and bioavailability of therapeutic agents, driving widespread clinical adoption and contributing to a rapidly expanding multi-billion-dollar market. However, despite their success, liposome development remains a highly complex, resource-intensive and time-consuming process that requires extensive experimental optimization. Recent advances in machine learning offer a powerful approach to accelerate this process by systematically analyzing and optimizing critical process parameters and quality attributes, such as particle size and polydispersity index. This study explores the application of machine learning algorithms to optimize the production of liposomes inspired by the compositions of two clinically validated formulations, Doxil® (currently approved) and Marqibo® (formerly approved), using a microfluidic production platform. More than 300 experimental conditions were examined to develop predictive models for liposome characteristics, enabling a more efficient and data-driven design process as well as unprecedented formulation-specific inference capabilities. In addition, we have developed an open-source simulation tool that enables experimentalists to virtually explore formulation spaces, design optimal experiments, and perform scenario analyses without exhaustive laboratory work. All models were rigorously validated through independent sets of wet-lab experiments, demonstrating robust performance and adaptability even under resource-constrained conditions. Our findings highlight the transformative potential of machine learning to streamline liposome development, improve reproducibility, and enable cost-effective transition from research-scale to commercial manufacturing, while aligning with regulatory quality frameworks and GMP expectations, thereby supporting the adoption of Quality by Design principles in the advancement of nanomedicine.