1. Background
Glaucoma filtration surgery is used to treat medically uncontrolled glaucoma, but it has a high failure rate of 50% after 5 years of follow-up due to postoperative fibrosis. Monoclonal antibodies (mAbs) offer a promising alternative due to their high target specificity and minimal systemic toxicity. A critical aspect of developing a mAb sustained-release implant is the attainment of an appropriate drug release profile, which is essential to ensure the desired therapeutic efficacy.
2. Novelty & Importance
Traditionally, the development of the optimal sustained-release implants has relied on empirical design-of-experiment methodologies, which require extensive experimental iterations to fine-tune formulation parameters. Machine learning (ML) models provide a paradigm shift by enabling predictive modelling of drug release profiles and assisting in the rational design of implant formulations. Unlike conventional methods, ML algorithms can efficiently process complex, non-linear relationships between material composition, implant structure, and release kinetics, leading to a significant reduction in experimental workload and cost. Moreover, explainable AI techniques, such as SHapley Additive exPlanations (SHAP) analysis, can reveal how key formulation parameters influence drug release kinetics, enabling the optimisation of sustained-release implants.
3. Aims & Objectives
Our aim is to address two critical challenges in the development of ocular drug-eluting implants by integrating machine learning and AI with polymer science.
Objective 1: The interplay between implant fabrication parameters, such as polymer composition, porosity, and implant dimensions, leads to complex and non-linear effects on drug release kinetics. We will integrate machine learning with polymer science to systematically optimise polycaprolactone – polyethylene glycol (PCL-PEG) implant design.
Objective 2: There remains an unmet need for long-acting anti-fibrotic implants capable of modulating conjunctival scarring in glaucoma surgery, requiring sustained-release technology. We will test our machine Learning-designed antibody-eluting implants in vitro and in an animal model of glaucoma filtration surgery.