The personalized titration and optimization of insulin regimens for treatment of type 2 diabetes (T2D) are resource-demanding healthcare tasks. Here we propose a model-based reinforcement learning (RL) framework (called RL-DITR), which learns the optimal insulin regimen by analyzing glycemic state rewards through patient model interactions. When evaluated during the development phase for managing hospitalized patients with T2D, RL-DITR achieved superior insulin titration optimization (mean absolute error (MAE) of 1.10 ± 0.03 U) compared to other deep learning models and standard clinical methods. We performed a stepwise clinical validation of the artificial intelligence system from simulation to deployment, demonstrating better performance in glycemic control in inpatients compared to junior and intermediate-level physicians through quantitative (MAE of 1.18 ± 0.09 U) and qualitative metrics from a blinded review. Additionally, we conducted a single-arm, patient-blinded, proof-of-concept feasibility trial in 16 patients with T2D. The primary outcome was difference in mean daily capillary blood glucose during the trial, which decreased from 11.1 (±3.6) to 8.6 (±2.4) mmol L⁻¹ (P < 0.01), meeting the pre-specified endpoint. No episodes of severe hypoglycemia or hyperglycemia with ketosis occurred. These preliminary results warrant further investigation in larger, more diverse clinical studies. ClinicalTrials.gov registration: NCT05409391.

To report the initial efficacy results of the Retina Implant Alpha AMS (Retina Implant AG, Reutlingen, Germany) for partial restoration of vision in end-stage retinitis pigmentosa (RP). Prospective, single-arm, investigator-sponsored interventional clinical trial. Within-participant control comprising residual vision with the retinal implant switched ON versus OFF in the implanted eye. The Retina Implant Alpha AMS was implanted into the worse-seeing eye of 6 participants with end-stage RP and no useful perception of light vision. Eligibility criteria included previous normal vision for ≥12 years and no significant ocular or systemic comorbidity. Vision assessments were scheduled at 1, 2, 3, 6, 9, and 12 months postimplantation. They comprised tabletop object recognition tasks, a self-assessment mobility questionnaire, and screen-based tests including Basic Light and Motion (BaLM), grating acuity, and greyscale contrast discrimination. A full-field stimulus test (FST) was also performed. Improvement in activities of daily living, recognition tasks, and assessments of light perception with the implant ON compared with OFF. All 6 participants underwent successful implantation. Light perception and temporal resolution with the implant ON were achieved in all participants. Light localization was achieved with the implant ON in all but 1 participant (P4) in whom the chip was not functioning optimally because of a combination of iatrogenic intraoperative implant damage and incorrect implantation. Implant ON correct grating detections (which were at chance level with implant OFF) were recorded in the other 5 participants, ranging from 0.1 to 3.33 cycles/degree on 1 occasion. The ability to locate high-contrast tabletop objects not seen with the implant OFF was partially restored with the implant ON in all but 1 participant (P4). There were 2 incidents of conjunctival erosion and 1 inferotemporal macula-on retinal detachment, which were successfully repaired, and 2 incidents of inadvertent damage to the implant during surgery (P3 and P4). The Alpha AMS subretinal implant improved visual performance in 5 of 6 participants and has exhibited ongoing function for up to 24 months. Although implantation surgery remains challenging, new developments such as OCT microscope guidance added refinements to the surgical technique.

Two recent clinical reports of retinal gene therapy with adeno-associated virus (AAV) vectors in patients with Leber’s congenital amaurosis showed initial gains in visual function that subsequently declined.^1,2 We previously reported early improvement in visual acuity in two of six patients who received retinal gene therapy in one eye (the study eye) to treat choroideremia,³ a disease that is characterized by atrophy of the choriocapillaris and retinal pigment epithelium and involves vision loss that leads to blindness.