To evaluate the performance of large language models (LLMs), specifically Microsoft Copilot, GPT-4 (GPT-4o and GPT-4o mini), and Google Gemini (Gemini and Gemini Advanced), in answering ophthalmological questions and assessing the impact of prompting techniques on their accuracy. Prospective qualitative study. Microsoft Copilot, GPT-4 (GPT-4o and GPT-4o mini), and Google Gemini (Gemini and Gemini Advanced). A total of 300 ophthalmological questions from StatPearls were tested, covering a range of subspecialties and image-based tasks. Each question was evaluated using 2 prompting techniques: zero-shot forced prompting (prompt 1) and combined role-based and zero-shot plan-and-solve+ prompting (prompt 2). With zero-shot forced prompting, GPT-4o demonstrated significantly superior overall performance, correctly answering 72.3% of questions and outperforming all other models, including Copilot (53.7%), GPT-4o mini (62.0%), Gemini (54.3%), and Gemini Advanced (62.0%) (p < 0.0001). Both Copilot and GPT-4o showed notable improvements with Prompt 2 over Prompt 1, elevating Copilot's accuracy from the lowest (53.7%) to the second highest (72.3%) among the evaluated LLMs. While newer iterations of LLMs, such as GPT-4o and Gemini Advanced, outperformed their less advanced counterparts (GPT-4o mini and Gemini), this study emphasizes the need for caution in clinical applications of these models. The choice of prompting techniques significantly influences performance, highlighting the necessity for further research to refine LLMs capabilities, particularly in visual data interpretation, to ensure their safe integration into medical practice.

Purpose: To compare the performance and cost-effectiveness of DeepSeek-R1 with OpenAI o1 in diagnosing and managing ophthalmology clinical cases. Study Design: Cross-sectional evaluation. Methods: A total of 300 clinical cases spanning 10 different ophthalmology subspecialties were collected from StatPearls. Each case presented a multiple-choice question regarding the diagnosis or management of the clinical case. DeepSeek-R1 was accessed through its public chat-based interface, while OpenAI o1 was queried via an Application Program Interface (API) with a standardized temperature setting of 0.3. Both models were prompted using the Plan-and-Solve+ (PS+) prompt engineering method, instructing them to answer multiple choice questions for each case. Performance was calculated as the proportion of correctly answered multiple choice questions. McNemar’s test was employed to compare the two models’ performance on paired data. Inter-model agreement for correct diagnoses was evaluated via Cohen’s kappa. A token-based cost analysis was performed to estimate the comparative expenditures of running each model at scale, accounting for both input prompts and model-generated output. Results: DeepSeek-R1 and OpenAI o1 both achieved an identical overall performance of 82.0% (n=246/300; 95% CI: 77.3-85.9). Subspeciality-specific analysis revealed numerical variation in performance, though none of these comparisons reached statistical significance (p>0.05). Agreement in performance between the models was moderate overall (κ=0.503, p<0.001), with substantial agreement in Refractive Management/Intervention (κ=0.698, p<0.001) and moderate agreement in Retina/Vitreous (κ=0.561, p<0.001) and Ocular Pathology/Oncology (κ=0.495, p<0.01) cases. Cost analysis indicated an approximately 15-fold reduction in per-query, token-related expenses when using DeepSeek-R1 compared with OpenAI o1 for the same workload. Conclusions: DeepSeek-R1 demonstrates robust diagnostic reasoning and management decision-making capabilities, performing comparably to OpenAI o1 across a range of ophthalmic subspecialty cases, while also offering a substantial reduction in usage costs. These findings highlight the feasibility of utilizing open-weight, reinforcement learning-augmented LLMs as an accessible, cost-effective alternative to proprietary models.

Huntington's disease (HD) is a monogenic disorder with autosomal dominant inheritance. In HD patients, neurons in the striatum and cortex degenerate, leading to motor, psychiatric and cognitive disorders. Dysregulated synaptic function and calcium handling are common in many neurodegenerative diseases, including HD. N‐methyl‐D‐aspartate (NMDA) receptor function is enhanced in HD at extrasynaptic sites, altering the balance of calcium‐dependent neuronal survival versus death signalling pathways. Endoplasmic reticulum (ER) calcium handling is also abnormal in HD. The ER, which is continuous with the nuclear envelope, is purportedly involved in nuclear calcium signalling; based on this, we hypothesised that nuclear calcium signalling is altered in HD. We explored this hypothesis with calcium imaging techniques, including simultaneous epifluorescent imaging of cytosolic and nuclear calcium using jRCaMP1b and GCaMP3 sensors, respectively, in striatal spiny projection neurons in cortical–striatal co‐cultures from the YAC128 mouse model of HD. Our data show contributions from a variety of calcium channels to nuclear calcium signalling. NMDA receptors (NMDARs) play an essential role in initiating action potential‐dependent calcium signalling to the nucleus, and ryanodine receptors (RyR) contribute to both cytosolic and nuclear calcium signals. Unlike previous reports in glutamatergic hippocampal and cortical neurons, we found that in GABAergic striatal neurons, L‐type voltage‐gated calcium channels (CaV) contribute to cytosolic, but not nuclear calcium signalling. Calcium imaging also suggests impairments in nuclear calcium signalling in HD striatal neurons, where spontaneous action potential‐dependent calcium transients in the nucleus were smaller in YAC128 striatal neurons compared to those of wild‐type (WT). Our results elucidate mechanisms involved in action potential‐dependent nuclear calcium signalling in GABAergic striatal neurons, and have revealed a clear deficit in this signalling in HD. image

Boston keratoprosthesis (KPro) plays an important role in treating complex corneal disease but is associated with many complications. This study addresses correlations and trends found in KPro patients, to understand the evolution and optimize postoperative care to minimize complications. A retrospective case series analyzing outcomes and evolution over 15 years. 157 eyes of 136 patients who underwent KPro surgery between 2008 and 2022 in Montreal, Quebec. Mean follow-up of 8.44 ± 4.70 years. Parameters followed included initial graft indication, age, pre-and-post-operative best corrected visual acuity (BCVA), intraocular pressure, cup-to-disc ratio, glaucoma interventions, and postoperative complications. Aniridia was the most common initial graft indication. Fifty-four percent of patients had at least one prior graft failure. The most significant decrease in vision occurred in the first 3 years, followed by a gradual decrease, with 33% of patients retaining BCVA better than 20/200 by 15 years. The best BCVA was significantly better than the initial BCVA. Panocular conditions, such as aniridia, showed quicker visual decline than purely anterior segment pathologies. Retroprosthetic membranes, hypotony, and cystoid macular edema are common complications. Certain complications were seen to be significantly more prevalent in younger populations, although visual acuity and de novo glaucoma development did not differ significantly with age. De novo glaucoma occurred in 63.64%, diagnosed most frequently in the first 2-3 years, with one-third of patients requiring surgical intervention. KPro offers a prolongation of functional vision, but patients must be monitored closely to detect and treat associated complications, especially in the first 5 years postoperatively.

Purpose: To investigate if patients had a subjectively inferior experience during second-eye surgery vs first-eye surgery during immediate sequential bilateral cataract surgery (ISBCS) as both operations were performed within the same surgical visit. Setting: Multisurgeon ophthalmic surgical center. Design: Survey-based prospective design. Methods: Patients of the center undergoing routine cataract surgery who enrolled in the study completed a questionnaire immediately after surgery to describe their surgical experience for each eye. Survey questions measured patient pain, discomfort or pressure, comfort, relaxation, estimated length of surgery, and predicted visual outcome. Results: Patient-reported level of pain during the second-eye surgery was significantly greater than for the first eye ( P < .001). Furthermore, the order of surgery was found to be a strong predictor of patient-reported pain ( P < .001), more so than surgical length ( P < .008), additional anesthesia ( P < .35), patient age ( P < .44), or patient sex ( P < .88). Overall, surgery during the first eye was reported as more comfortable ( P < .001) and shorter in duration ( P < .001), while discomfort and pressure were reported as worse for the second eye ( P < .001). Conclusions: Understanding differential experiences between eyes can help surgeons when counselling patients regarding expectations for ISBCS and related visual outcomes.