Reut Shor’s research while affiliated with University of Toronto and other places

Importance Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are extensively used in treating diabetes and obesity, yet little is known about the long-term ocular effects of systemic prolonged exposure. Objective To evaluate the risk of developing neovascular age-related macular degeneration (nAMD) associated with the use of GLP-1 RAs in patients with diabetes. Design, Setting, and Participants This population-based, retrospective cohort study was conducted from January 2020 to November 2023, with a follow-up period of 3 years. Data analysis was performed from August 2024 to October 2024. The investigators used comprehensive administrative health and demographic data from patients in Ontario, Canada, which were collected by the Institute for Clinical Evaluative Sciences in the context of a universal public health care system. Inclusion criteria were patients aged 66 years or older with a diagnosis of diabetes and a minimum follow-up period of 12 months following initial diabetes diagnosis. Patients with incomplete Ontario Health Insurance Plan or Ontario Drug Benefit data or patients exposed to GLP-1 RA for less than 6 months were excluded. Of 1 119 517 eligible patients, a 1:2 matched cohort of 139 002 patients was created, including 46 334 patients who were exposed to GLP-1 RAs and 92 668 unexposed matched patients. Systemic comorbidities that were associated with any kind of AMD and socioeconomic status were used to calculate propensity scores. Exposure GLP-1 RA use for 6 months or longer. Main Outcomes and Measures The primary outcome was the incidence and time to event of nAMD during the follow-up period. Results Among 139 002 matched patients, mean (SD) patient age was 66.2 (7.5) years, and 64 775 patients (46.6%) were women. The incidence of nAMD was higher among the exposed cohort than among the unexposed cohort. Cox proportional hazard models, both unadjusted (crude) and adjusted, estimated hazard ratios for nAMD development of greater than 2.0 among patients exposed to GLP-1 RAs (exposed, 0.2% vs unexposed, 0.1%; difference, 0.1%; crude: HR, 2.11; 95% CI, 1.58-2.82; adjusted: HR, 2.21; 95% CI, 1.65-2.96). Conclusions and Relevance In this cohort study, the use of GLP-1 RAs among patients with diabetes was associated with a 2-fold higher risk of incident nAMD development than among similar patients with diabetes who did not receive a GLP-1 RA. Further research is needed to elucidate the exact pathophysiological mechanisms involved and to understand the trade-offs between the benefits and risks of GLP-1 RAs.

Topic Epiretinal membrane (ERM) can impair central vision by forming a pre-retinal fibrous layer on the inner retina. Artificial intelligence (AI)–based tools may streamline ERM diagnosis, but their overall performance and factors affecting accuracy require evaluation. Clinical Relevance With an aging population, ERM prevalence is expected to rise, placing increased demands on clinical resources. Early detection via AI models could expedite diagnosis, reduce subjective errors, and guide timely surgical intervention. This systematic review and meta-analysis evaluates the pooled diagnostic performance of AI models for detecting ERM and identifies study- and model-level factors influencing their performance. Design Systematic Review and Meta-Analysis Methods Comprehensive searches were conducted in Medline, Embase, Cochrane Library, Web of Science, and preprint databases from inception to June 2024. Included studies evaluated AI models for ERM diagnosis. Study quality and risk of bias were assessed using the Quality Assessment for Diagnostic Accuracy Studies 2 (QUADAS-2) tool. A random-effects model was applied to pool diagnostic accuracy, sensitivity, specificity, and diagnostic odds ratio. Subgroup analyses explored factors affecting model performance. The study protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO - CRD42024563571). Results Of 379 articles screened, 26 met inclusion criteria, and 19 contributed to the meta-analysis. Study settings were predominantly hospital-based (76.9%), with some studies from academic computer and biomedical science departments (15.4%) and community centers (7.7%). Quality assessments suggested low or unclear risk of bias and applicability concerns in 95% of studies. The pooled sensitivity was 90.1% (95% CI: 85.8–93.2), and the pooled specificity was 95.7% (95% CI: 88.8–95.2). Subgroup analysis showed higher specificity (97.1%, 95% CI: 96.0–97.9) in AI models using color fundus photographs than optical coherence tomography scans, which had a specificity of 92.6% (95% CI: 88.8–95.2). External validation was performed in 26.9% of studies. All included studies used expert human grading as the reference standard, of which 25 (96.2%) were based on the same imaging modality as the AI input. The proportion of ERM cases in development datasets varied across studies, particularly between single-disease and multiclass models. Conclusions AI models demonstrate high diagnostic performance for ERM. However, limited external validation and variability in AI development methodologies limits direct comparison between models and real-world applicability. Future work should standardize model development and reporting practices, improve data interoperability, and develop prediction models to track disease progression and determine optimal surgical timing.

Background and Objective This study evaluated optometrists' referral patterns for epiretinal membrane (ERM) patients in Ontario, Canada, and their attitudes towards an artificial intelligence (AI) tool for improving referral accuracy. An anonymous online survey with 11 questions was conducted. Patients and Methods The survey targeted optometrists across Ontario, Canada. The survey aimed to understand optometrists' reasons for referring ERM patients to retina specialists, their expectations of the specialists' management, and their openness to using an AI tool for triage. To prevent bias, the survey described an AI tool as an online consultation feature limited to predefined questions without directly mentioning “AI.” The main objective was to assess if this AI tool could decrease unnecessary ERM referrals to retina specialists. Results A total of 135 optometrists participated. They reported seeing an average of eight ERM cases monthly, referring every fourth case to a specialist. The primary referral reason (84.3%) was to evaluate for surgery. In terms of referral confidence, 34.3% felt fully confident (5/5), and 47.8% slightly less so (4/5). They anticipated that 20% of patients would have a change in management post-consultation with a specialist. When introduced to the concept of an online consultation tool for patient screening, optometrists believed it could reduce their ERM referrals by 40%. Conclusions Optometrists often refer ERM patients to retina specialists. An AI tool for screening ERM referrals, based on presenting vision and OCT images, could significantly lower the number of unnecessary referrals, offering clinical guidance to optometrists. [Ophthalmic Surg Lasers Imaging Retina 2025;56:166–169.]

Aim To assess the primary reattachment rate (PARR) in pneumatic retinopexy (PnR) versus pars plana vitrectomy (PPV) for rhegmatogenous retinal detachment (RRD) meeting the Pneumatic Retinopexy versus Vitrectomy for the Management of Primary Rhegmatogenous Retinal Detachment Outcomes Randomised Trial (PIVOT) criteria with a single break in detached retina. Methods A post hoc analysis of two clinical trials. To be included, patients with primary RRD had to meet PIVOT criteria but could have only one break in the detached retina. Patients with additional pathology in the attached retina were included in a secondary analysis. The primary outcome was PARR following PnR versus PPV at 1-year postoperatively. Results 162 patients were included. 53% (86/162) underwent PnR and 47% (76/162) had a PPV. 99% (85/86) and 86.8% (66/76) completed the 1-year follow-up visits in the PnR and PPV groups, respectively. PARR was 88.2% (75/85) in the PnR group and 90.9% (60/66) in the PPV group (p=0.6) with a mean postoperative logMAR best-corrected visual acuity of 0.19±0.25 versus 0.34±0.37 (Snellen 20/30 vs 20/44) (p=0.01) each in the PnR and PPV groups, respectively. In an additional analysis of patients who were also allowed to have any pathology in the attached retina, the PARR was 85% (91/107) and 91.6% (66/72) in the PnR and PPV groups, respectively (p=0.18). Conclusions PnR and PPV provide similar long-term PARR in a substantial proportion of patients meeting PIVOT criteria with only a single break in the detached retina. Therefore, in patients meeting these specific criteria, PnR is an appropriate first-line therapy as it offers superior functional outcomes without compromising PARR.

Purpose To assess the performance of Chat Generative Pre-Trained Transformer-4 in providing accurate diagnoses to retina teaching cases from OCTCases. Design Cross-sectional study. Subjects Retina teaching cases from OCTCases. Methods We prompted a custom chatbot with 69 retina cases containing multimodal ophthalmic images, asking it to provide the most likely diagnosis. In a sensitivity analysis, we inputted increasing amounts of clinical information pertaining to each case until the chatbot achieved a correct diagnosis. We performed multivariable logistic regressions on Stata v17.0 (StataCorp LLC) to investigate associations between the amount of text-based information inputted per prompt and the odds of the chatbot achieving a correct diagnosis, adjusting for the laterality of cases, number of ophthalmic images inputted, and imaging modalities. Main Outcome Measures Our primary outcome was the proportion of cases for which the chatbot was able to provide a correct diagnosis. Our secondary outcome was the chatbot’s performance in relation to the amount of text-based information accompanying ophthalmic images. Results Across 69 retina cases collectively containing 139 ophthalmic images, the chatbot was able to provide a definitive, correct diagnosis for 35 (50.7%) cases. The chatbot needed variable amounts of clinical information to achieve a correct diagnosis, where the entire patient description as presented by OCTCases was required for a majority of correctly diagnosed cases (23 of 35 cases, 65.7%). Relative to when the chatbot was only prompted with a patient’s age and sex, the chatbot achieved a higher odds of a correct diagnosis when prompted with an entire patient description (odds ratio = 10.1, 95% confidence interval = 3.3–30.3, P < 0.01). Despite providing an incorrect diagnosis for 34 (49.3%) cases, the chatbot listed the correct diagnosis within its differential diagnosis for 7 (20.6%) of these incorrectly answered cases. Conclusions This custom chatbot was able to accurately diagnose approximately half of the retina cases requiring multimodal input, albeit relying heavily on text-based contextual information that accompanied ophthalmic images. The diagnostic ability of the chatbot in interpretation of multimodal imaging without text-based information is currently limited. The appropriate use of the chatbot in this setting is of utmost importance, given bioethical concerns. Financial Disclosure(s) Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Importance Ophthalmology is reliant on effective interpretation of multimodal imaging to ensure diagnostic accuracy. The new ability of ChatGPT-4 (OpenAI) to interpret ophthalmic images has not yet been explored. Objective To evaluate the performance of the novel release of an artificial intelligence chatbot that is capable of processing imaging data. Design, Setting, and Participants This cross-sectional study used a publicly available dataset of ophthalmic cases from OCTCases, a medical education platform based out of the Department of Ophthalmology and Vision Sciences at the University of Toronto, with accompanying clinical multimodal imaging and multiple-choice questions. Across 137 available cases, 136 contained multiple-choice questions (99%). Exposures The chatbot answered questions requiring multimodal input from October 16 to October 23, 2023. Main Outcomes and Measures The primary outcome was the accuracy of the chatbot in answering multiple-choice questions pertaining to image recognition in ophthalmic cases, measured as the proportion of correct responses. χ ² Tests were conducted to compare the proportion of correct responses across different ophthalmic subspecialties. Results A total of 429 multiple-choice questions from 136 ophthalmic cases and 448 images were included in the analysis. The chatbot answered 299 of multiple-choice questions correctly across all cases (70%). The chatbot’s performance was better on retina questions than neuro-ophthalmology questions (77% vs 58%; difference = 18%; 95% CI, 7.5%-29.4%; χ ² 1 = 11.4; P < .001). The chatbot achieved a better performance on nonimage–based questions compared with image-based questions (82% vs 65%; difference = 17%; 95% CI, 7.8%-25.1%; χ ² 1 = 12.2; P < .001).The chatbot performed best on questions in the retina category (77% correct) and poorest in the neuro-ophthalmology category (58% correct). The chatbot demonstrated intermediate performance on questions from the ocular oncology (72% correct), pediatric ophthalmology (68% correct), uveitis (67% correct), and glaucoma (61% correct) categories. Conclusions and Relevance In this study, the recent version of the chatbot accurately responded to approximately two-thirds of multiple-choice questions pertaining to ophthalmic cases based on imaging interpretation. The multimodal chatbot performed better on questions that did not rely on the interpretation of imaging modalities. As the use of multimodal chatbots becomes increasingly widespread, it is imperative to stress their appropriate integration within medical contexts.

Purpose: To assess real-world primary anatomic reattachment rate and postoperative visual acuity outcomes in patients aged 75 years or older who underwent pneumatic retinopexy (PnR) for primary rhegmatogenous retinal detachment (RRD). Design: A retrospective cohort study. Subjects: Patients aged 75 years or older with primary RRD undergoing PnR. Methods: This study evaluates real-world outcomes from Oct 1, 2010, to December 31, 2022, of eligible patients with a minimum of 3 months follow-up. There were no limitations with respect to the number, size or location of retinal breaks. Exclusion criteria included significant proliferative vitreoretinopathy or previous retinal detachment repair in the index eye, inability to maintain the postoperative posturing requirements (such as physical disability), very inferior breaks below 5 or 7 o'clock) or inability to carry out adequate examination of the peripheral retina due to media opacity. Lens status did not impact decision to offer PnR. Main outcome measures: PnR primary anatomic reattachment rate and postoperative visual acuity (VA) at 3 months. Results: 80 patients with a mean age of 80.6±4.6 years were included in this study. 35% (28) were phakic, and 34% (27) presented with a fovea-on RRD. The mean number of breaks in the detached retina was 1.52±1.13, and the mean number of quadrants of detached retina was 2.35±0.93. The primary anatomic reattachment rate at 3 months following PnR was 78.8% (63/80), while the remaining 21.2% (17/80) failed PnR and underwent an operating room procedure. There was a statistically significant improvement in LogMAR VA from baseline to 3 months (1.29±0.94 and 0.69±0.67, respectively, p<0.001). A subgroup analysis that only included patients above the age of 80 was also performed, with a total of 39 patients with a mean age of 84.4±3.5 years. Primary anatomic reattachment rate with PnR in this subgroup was 74.4% (29/39) with a statistically significant improvement in LogMAR VA from baseline to 3 months (1.4±1.05 and 0.77±0.70, respectively, p=0.004). Conclusion: Patients aged 75 years and above or aged 80 years and above treated with PnR for primary RRD had primary anatomic reattachment rates of 78.8% and 74.4%, respectively. These are relatively comparable primary reattachment rates in this age group with other surgical techniques such as pars plana vitrectomy (PPV), scleral buckle (SB) or PPV/SB. Pneumatic retinopexy is an effective, minimally invasive office-based procedure which may be desirable for some elderly patients.