Automated Spectral Optimization of Selective Trabeculoplasty (SLT) Laser Parameters via Adaptive Bayesian Reinforcement Learning

**Abstract:** This paper presents a novel system for automated optimization of Selective Trabeculoplasty (SLT) laser treatment parameters. Leveraging adaptive Bayesian Reinforcement Learning (ABRL) integrated with real-time optical coherence tomography (OCT) feedback, our system dynamically adjusts laser spot size, power, and pulse duration to achieve superior intraocular pressure (IOP) reduction and minimal collateral damage. The method minimizes trial-and-error in clinical settings, accelerating treatment and improving patient outcomes within the established safety profile of SLT. We demonstrate the viability and potential improvement over existing manual procedures through simulation and retrospective data analysis, projecting significant long-term benefits for glaucoma treatment.