Formulation, Characterization, and Machine Learning Prediction of Poly(lactic-co-glycolic) acid Nanoparticles for Oncological Pregnant Women Treatment.

Cancer treatments during pregnancy pose serious risks, particularly chemotherapy, due to the potential transplacental passage of drugs that may attack the fetus. Such exposure can lead to malformations or, in severe cases, miscarriage. Despite these concerns, there is a clinical need for effective and safe cancer therapy for pregnant patients. The objective of this thesis is to develop a nanoparticle-based drug delivery system capable of reducing fetal exposure to chemotherapeutic agents, while simultaneously introducing a novel machine learning (ML)-driven strategy to optimize the nanoparticle design process. Specifically, polymeric nanoparticles were synthesized using Poly-(Lactic-co-Glycolic Acid) (PLGA) to encapsulate doxorubicin and minimize its transplacental transfer. Nanoparticles were prepared via a double emulsion method (water-in-oil-in-water, W/O/W), testing over 50 formulations by varying the proportions of aqueous phase, oil phase, and surfactant. Each formulation was systematically evaluated in terms of particle size, polydispersity index (PDI), zeta potential, and drug encapsulation efficiency (EE%). A key novelty of this work lies in the integration of machine learning techniques to guide and accelerate formulation development. Supervised ML algorithms were trained to predict both the likelihood of stable nanoemulsion formation and the physicochemical properties of the resulting nanoparticles. This approach not only reduced experimental workload and material waste but also achieved predictive accuracies above 80%, demonstrating its potential as a powerful tool for formulation design. By combining experimental nanoparticle synthesis with computational machine learning modeling, this thesis provides a dual strategy for creating safer and more efficient drug delivery systems for use in pregnancy; addressing both the medical challenge of maternal cancer treatment and the methodological innovation of applying ML to nanomedicine.