(607f) Predicting Drug Release from Intraocular Implants Using a Virtual Eye Model

We develop an in-silico model and data processing pipeline to predict the performance of intraocular drug delivery implants in real ocular settings, by simulating the flow of Aqueous Humor (AH) and drug delivery from an implant to the Trabecular Meshwork (TM) in realistic eye geometries. This model accurately predicts the physiological values of Intraocular Pressure (IOP) for both healthy individuals and glaucoma patients, as reported in the literature. Further, it allows to estimate the bioavailability time period (BTP) for a given amount of loaded drug, design and material properties of implant. Moreover, the model may also predict the in-vivo behavior of the implant with a known in-vitro release profile as the input. We specifically show the results for (i) intracameral implants placed in anterior chamber, and (ii) intraocular implants attached to the haptics of intraocular lens placed in the posterior chamber. Results reveal that the effective diffusivity of the drug within the implant is the critical parameter that can alter the BTP from a few days to months. Intuitively, BTP should increase as effective diffusivity decreases. However, we discovered that with lower levels of initial drug loading, BTP declines when effective diffusivity falls below a specific threshold. Our findings further reveal that, while AH flow has a minimal effect on the drug release profile at the implant site, it significantly impacts drug availability at the TM. Further, we demonstrate that both the optimal positioning of the drug reservoir within the implant and the implant design itself can independently extend the operational time period by up to 75%. A simple geometric design rule is proposed to guide implant shaping for maximizing the bioavailability time period of drug at Trabecular Meshwork. Furthermore, the use of in-vitro drug release profiles as input instead of the incorporation of the explicit drug distribution and drug release mechanism in the implant can predict drug availability at the trabecular meshwork with less than 5% error. This approach may therefore be used for any implant with known in vitro drug release profile, even in cases when the physiochemical and mechanistic aspects of the implant is not provided.