A Comparative Study of Responses to Retina Questions from Either Experts, Expert-Edited Large Language Models, or Expert-Edited Large Language Models Alone.

Objective: To assess the quality, empathy, and safety of expert edited large language model (LLM), human expert created, and LLM responses to common retina patient questions. Design: Randomized, masked multicenter study. Participants: Twenty-one common retina patient questions were randomly assigned among 13 retina specialists. Methods: Each expert created a response (Expert) and then edited a LLM (ChatGPT-4)-generated response to that question (Expert + artificial intelligence [AI]), timing themselves for both tasks. Five LLMs (ChatGPT-3.5, ChatGPT-4, Claude 2, Bing, and Bard) also generated responses to each question. The original question along with anonymized and randomized Expert + AI, Expert, and LLM responses were evaluated by the other experts who did not write an expert response to the question. Evaluators judged quality and empathy (very poor, poor, acceptable, good, or very good) along with safety metrics (incorrect information, likelihood to cause harm, extent of harm, and missing content). Main Outcome: Mean quality and empathy score, proportion of responses with incorrect information, likelihood to cause harm, extent of harm, and missing content for each response type. Results: There were 4008 total grades collected (2608 for quality and empathy; 1400 for safety metrics), with significant differences in both quality and empathy (P < 0.001, P < 0.001) between LLM, Expert and Expert + AI groups. For quality, Expert + AI (3.86 ± 0.85) performed the best overall while GPT-3.5 (3.75 ± 0.79) was the top performing LLM. For empathy, GPT-3.5 (3.75 ± 0.69) had the highest mean score followed by Expert + AI (3.73 ± 0.63). By mean score, Expert placed 4 out of 7 for quality and 6 out of 7 for empathy. For both quality (P < 0.001) and empathy (P < 0.001), expert-edited LLM responses performed better than expert-created responses. There were time savings for an expert-edited LLM response versus expert-created response (P = 0.02). ChatGPT-4 performed similar to Expert for inappropriate content (P = 0.35), missing content (P = 0.001), extent of possible harm (P = 0.356), and likelihood of possible harm (P = 0.129). Conclusions: In this randomized, masked, multicenter study, LLM responses were comparable with experts in terms of quality, empathy, and safety metrics, warranting further exploration of their potential benefits in clinical settings. Financial Disclosures: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of the article.

ChatGPT enters the room: what it means for patient counseling, physician education, academics, and disease management.

PURPOSE OF REVIEW: This review seeks to provide a summary of the most recent research findings regarding the utilization of ChatGPT, an artificial intelligence (AI)-powered chatbot, in the field of ophthalmology in addition to exploring the limitations and ethical considerations associated with its application. RECENT FINDINGS: ChatGPT has gained widespread recognition and demonstrated potential in enhancing patient and physician education, boosting research productivity, and streamlining administrative tasks. In various studies examining its utility in ophthalmology, ChatGPT has exhibited fair to good accuracy, with its most recent iteration showcasing superior performance in providing ophthalmic recommendations across various ophthalmic disorders such as corneal diseases, orbital disorders, vitreoretinal diseases, uveitis, neuro-ophthalmology, and glaucoma. This proves beneficial for patients in accessing information and aids physicians in triaging as well as formulating differential diagnoses. Despite such benefits, ChatGPT has limitations that require acknowledgment including the potential risk of offering inaccurate or harmful information, dependence on outdated data, the necessity for a high level of education for data comprehension, and concerns regarding patient privacy and ethical considerations within the research domain. SUMMARY: ChatGPT is a promising new tool that could contribute to ophthalmic healthcare education and research, potentially reducing work burdens. However, its current limitations necessitate a complementary role with human expert oversight.

Enhanced durability and evolution of retreatment criteria of intravitreal antivascular endothelial growth factor agents for diabetic macular edema.

PURPOSE OF REVIEW: The increasing prevalence of diabetic macular edema (DME) necessitates an updated review of treatment modalities. While the shift from laser to anti-vascular endothelial growth factor (anti-VEGF) therapy has transformed patient outcomes, benefits of these agents are not fully realized in real-world implementation relative to the setting of controlled clinical trials. This review outlines the evolution of intravitreal anti-VEGF treatment extension protocols for DME that reflect efforts to address treatment adherence challenges while optimizing visual outcomes. RECENT FINDINGS: Recent studies highlight the efficacy of extended-interval dosing with anti-VEGF agents in managing DME. Trials such as RISE/RIDE, VISTA/VIVID, and LUCIDATE have established the foundation of these regimens by demonstrating sustained visual gains with continuous treatment. However, newer trials including PROTOCOL T, KESTREL/KITE, YOSEMITE/RHINE, and PHOTON have furthered this concept, revealing that less frequent dosing of various anti-VEGF agents can maintain similar visual acuity and anatomical outcomes to traditional monthly injections. SUMMARY: The reviewed findings suggest a paradigm shift in DME treatment toward less frequent anti-VEGF injections. This has significant implications for clinical practice, potentially leading to greater adherence to treatment regimens and sustained visual function in patients, while minimizing treatment burden and healthcare costs. Further investigation into the long-term effects of extended dosing intervals is required.

Conversion to faricimab after prior anti-vascular endothelial growth factor therapy for persistent diabetic macular oedema.

BACKGROUND: To assess the anatomical and functional outcomes in eyes with persistent diabetic macular oedema (pDME) on chronic anti-vascular endothelial growth factor therapy switched to intravitreal faricimab. METHODS: Patients with pDME on chronic anti-vascular endothelial growth factor therapy that were switched to faricimab and received at least three injections at our institution between April 2022 and May 2023 were included in this study. Patients were excluded if they had complete response to previous treatment but were switched to extend treatment intervals if they had steroid or laser treatment for DME within 6 months prior to switch. Clinical and imaging data were extracted from the electronic medical record. Central foveal thickness (CFT) and Snellen visual acuity (VA) were obtained before and after three intravitreal faricimab injections. Generalised estimating equations were used to analyse the change in CFT and VA. RESULT: During the study period, 69 eyes of 53 patients met inclusion criteria. The mean age was 68.6±9.0 years. The mean number of injections prior to switch was 18.1±16.0. Pre-switch mean logarithm of the minimal angle of resolution VA was 0.40±0.30 (Snellen equivalent 20/50) and 0.38±0.27 (Snellen equivalent 20/48) after three faricimab injections (p=0.397). Mean CFT improved from 380±155 microns to 323±147 microns (p<0.001). No ophthalmic or systemic adverse events occurred during the study period. CONCLUSIONS: Intravitreal faricimab can improve anatomic outcomes while maintaining visual acuity in eyes with pDME previously treated with anti-VEGF therapy.

Cell therapy for retinal disease.

PURPOSE OF REVIEW: This review presents an update on completed stem cell therapy trials aimed at retinal diseases. RECENT FINDINGS: In recent years, several clinical trials have been conducted examining the safety and role of cell therapy in diseases, including age-related macular degeneration, Stargardt's macular dystrophy, and retinitis pigmentosa. Studies have utilized a variety of cell lines, modes of delivery, and immunosuppressive regimens. The prevalence of fraudulent cell therapy clinics poses threats to patients. SUMMARY: Clinical trials have begun to characterize the safety of cell therapy in retinal disease. While studies have described the potential benefits of cell therapy, larger studies powered to evaluate this efficacy are required to continue progressing toward preventing retinal disease. Nonapproved cell therapy clinics require regulation and patient education to avoid patient complications.

CHARACTERISTICS OF PARTICIPANTS IN DIABETIC RETINOPATHY CLINICAL RESEARCH CLINICAL TRIALS WHO WERE LOST TO FOLLOW-UP.

PURPOSE: Our study describes the racial/ethnic composition and the broad health status/severity of disease of the cohorts of Diabetic Retinopathy Clinical Research clinical trials and compares the rates of noncompletion of the trial with the first data report (lost to follow-up). METHODS: This study is a retrospective cohort study of participants in Diabetic Retinopathy Clinical Research Retina Network clinical trials who met our inclusion criteria. The primary outcome was the lost to follow-up rates and their associated risk factors. RESULTS: Eight clinical trials with 3,492 participants met our inclusion criteria. Participants who were lost to follow-up were more likely to be younger, Hispanic or Black, had Type 1 diabetes mellitus, had worse baseline best-corrected visual acuity, had higher hemoglobin A1c, had higher blood pressure, and had proliferative diabetic retinopathy (PDR) (Diabetic Retinopathy Severity Score >66). CONCLUSION: Participants who were lost to follow-up during Diabetic Retinopathy Clinical Research clinical trials tended to be younger, Hispanic or Black, and had worse diabetes and hypertension. This study highlights the importance of ensuring that a representative population is maintained through the end of the study period.

Ophthalmology Workforce Projections in the United States, 2020 to 2035.

PURPOSE: To analyze ophthalmology workforce supply and demand projections from 2020 to 2035. DESIGN: Observational cohort study using data from the National Center for Health Workforce Analysis (NCHWA). METHODS: Data accessed from the Department of Health and Human Services, Health Resources and Services Administration (HRSA) website were compiled to analyze the workforce supply and demand projections for ophthalmologists from 2020 to 2035. MAIN OUTCOME MEASURES: Projected workforce adequacy over time. RESULTS: From 2020 to 2035, the total ophthalmology supply is projected to decrease by 2650 full-time equivalent (FTE) ophthalmologists (12% decline) and total demand is projected to increase by 5150 FTE ophthalmologists (24% increase), representing a supply and demand mismatch of 30% workforce inadequacy. The level of projected adequacy was markedly different based on rurality by year 2035 with 77% workforce adequacy versus 29% workforce adequacy in metro and nonmetro geographies, respectively. By year 2035, ophthalmology is projected to have the second worst rate of workforce adequacy (70%) of 38 medical and surgical specialties studied. CONCLUSIONS: The HRSA's Health Workforce Simulation Model forecasts a sizeable shortage of ophthalmology supply relative to demand by the year 2035, with substantial geographic disparities. Ophthalmology is one of the medical specialties with the lowest rate of projected workforce adequacy by 2035. Further dedicated workforce supply and demand research for ophthalmology and allied professionals is needed to validate these projections, which may have significant future implications for patients and providers. FINANCIAL DISCLOSURE(S): Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Google Search Trends to assess public interest in and concern about Vuity for treating presbyopia.

PURPOSE: To assess public awareness, interest, and concerns regarding Vuity (1.25% pilocarpine hydrochloride ophthalmic solution), an eye drop for the treatment of presbyopia, based on Google Trends. METHODS: We used Google Trends that provides a relative search volume for queried terms, to evaluate searches for Vuity from June 30, 2021, to June 30, 2022, in the United States. The data for this study were downloaded on June 30, 2022. Main outcome measures were changes in relative search volumes for the terms "Vuity," "Eye drops for reading," "Eye drops for near vision," "Presbyopia," "Pilocarpine," and related popular search terms, such as "Vuity side effects," and "Vuity retinal detachment". RESULTS: Since the approval of Vuity on October 29, 2021, notable increases in the relative search volumes occurred for Vuity in October 2021, December 2021, and from March to April 2022, which coincided with its approval, availability, and subsequent direct-to-consumer advertising based on positive results of clinical trials. The direct-to-consumer advertising had the greatest impact on the search volume for Vuity. Specific interests included Vuity cost, where to buy it, and its side effects. Retinal detachment was the most highly searched Vuity side effect. Geographic variation was evident, with the relative search volumes highest for "Vuity" in Wyoming, followed by North Dakota. CONCLUSION: Google Trends is a useful tool for monitoring increases in public interest in Vuity for presbyopia. Public concerns regarding side effects warrant further real-world investigations of the causal relationship between Vuity and retinal detachment.

Appropriateness and Readability of ChatGPT-4-Generated Responses for Surgical Treatment of Retinal Diseases.

OBJECTIVE: To evaluate the appropriateness and readability of the medical knowledge provided by ChatGPT-4, an artificial intelligence-powered conversational search engine, regarding common vitreoretinal surgeries for retinal detachments (RDs), macular holes (MHs), and epiretinal membranes (ERMs). DESIGN: Retrospective cross-sectional study. SUBJECTS: This study did not involve any human participants. METHODS: We created lists of common questions about the definition, prevalence, visual impact, diagnostic methods, surgical and nonsurgical treatment options, postoperative information, surgery-related complications, and visual prognosis of RD, MH, and ERM, and asked each question 3 times on the online ChatGPT-4 platform. The data for this cross-sectional study were recorded on April 25, 2023. Two independent retina specialists graded the appropriateness of the responses. Readability was assessed using Readable, an online readability tool. MAIN OUTCOME MEASURES: The "appropriateness" and "readability" of the answers generated by ChatGPT-4 bot. RESULTS: Responses were consistently appropriate in 84.6% (33/39), 92% (23/25), and 91.7% (22/24) of the questions related to RD, MH, and ERM, respectively. Answers were inappropriate at least once in 5.1% (2/39), 8% (2/25), and 8.3% (2/24) of the respective questions. The average Flesch Kincaid Grade Level and Flesch Reading Ease Score were 14.1 ± 2.6 and 32.3 ± 10.8 for RD, 14 ± 1.3 and 34.4 ± 7.7 for MH, and 14.8 ± 1.3 and 28.1 ± 7.5 for ERM. These scores indicate that the answers are difficult or very difficult to read for the average lay person and college graduation would be required to understand the material. CONCLUSIONS: Most of the answers provided by ChatGPT-4 were consistently appropriate. However, ChatGPT and other natural language models in their current form are not a source of factual information. Improving the credibility and readability of responses, especially in specialized fields, such as medicine, is a critical focus of research. Patients, physicians, and laypersons should be advised of the limitations of these tools for eye- and health-related counseling. FINANCIAL DISCLOSURE(S): Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Neurotrophic keratopathy following rhegmatogenous retinal detachment surgery.

OBJECTIVE: To describe clinical characteristics and visual outcomes of eyes developing neurotrophic keratopathy (NK) following rhegmatogenous retinal detachment (RRD) repair. METHODS: All eyes with NK at Wills Eye Hospital following RRD repair from June 1, 2011, to December 1, 2020 were included. Patients with prior ocular procedures (other than cataract surgery), herpetic keratitis, and diabetes mellitus were excluded. RESULTS: During the study period, 241 patients were diagnosed with NK, and 8179 eyes underwent RRD surgery, giving a 9-year prevalence rate of 0.1% (95% CI, 0.1%-0.2%). Mean age was 53.4 ± 16.6 years during RRD repair and 56.5 ± 13.4 years during NK diagnosis. Mean time to NK diagnosis was 3.0 ± 5.6 years (range, 6 days to 18.8 years). Mean visual acuity before NK was 1.10 ± 0.56 logMAR (20/252 Snellen), and it was 1.01 ± 0.62 logMAR (20/205 Snellen) at final visit (p = 0.75). Six eyes (54.5%) developed NK <1 year following RRD surgery. Mean final visual acuity was 1.01 ± 0.53 logMAR (20/205 Snellen) in this group versus 1.01 ± 0.78 logMAR (20/205 Snellen) in the delayed NK group (p = 1.00). CONCLUSIONS: NK may present acutely or up to several years following surgery, with severity of corneal defects ranging from stage 1 to stage 3 NK. Surgeons should be mindful of the potential for this rare complication following RRD repair.

VITRECTOMY FOR VITREOUS HEMORRHAGE ASSOCIATED WITH RETINAL VEIN OCCLUSION: Visual Outcomes, Prognostic Factors, and Sequelae.

PURPOSE: To report the outcomes of pars plana vitrectomy for vitreous hemorrhage (VH) associated with retinal vein occlusion and to identify prognostic indicators. METHODS: Interventional, retrospective consecutive case series between 2015 and 2021. RESULTS: The study included 138 eyes of 138 patients (64 female and 74 male); 81 patients had branch retinal vein occlusion and 57 had central retinal vein occlusion. The mean age was 69.8 years. The mean duration between the diagnosis of VH and surgery was 79.6 ± 115.3 (range, 1-572) days. The mean follow-up was 27.2 months. The logarithm of the minimum angle of resolution visual acuity significantly improved from 1.95 ± 0.72 (Snellen equivalent, 20/1782) to 0.99 ± 0.87 (20/195) at 6 months and to 1.06 ± 0.96 (20/230) at the final visit (both P < 0.001). The visual acuity at 6 months improved by three or more lines in 103 eyes (75%). Postoperative complications during follow-up included recurrent VH in 16 eyes (12%) (of which 8 eyes underwent reoperations), rhegmatogenous retinal detachment in six eyes (4%), and new neovascular glaucoma in three eyes (2%). Worse final visual acuity was significantly associated with older age ( P = 0.007), concurrent neovascular glaucoma ( P < 0.001), central retinal vein occlusion ( P < 0.001), worse preoperative visual acuity ( P < 0.001), postoperative new neovascular glaucoma ( P = 0.021), and postoperative retinal detachment ( P < 0.001). The duration of VH was not associated with visual outcomes ( P = 0.684). Preoperative antivascular endothelial growth factor injections and tamponade did not prevent postoperative recurrent VH. CONCLUSION: Pars plana vitrectomy is effective for VH associated with retinal vein occlusion, regardless of the duration of hemorrhage. However, pre-existing risk factors and postoperative sequelae may limit visual recovery.

Risk factors for surgical failure after primary scleral buckling for rhegmatogenous retinal detachment.

OBJECTIVE: To identify risk factors for surgical failure after scleral buckling (SB) for primary rhegmatogenous retinal detachment (RRD) repair. DESIGN: Single-centre retrospective consecutive case series. PARTICIPANTS: All patients who underwent SB for repair of primary RRD at Wills Eye Hospital between January 1, 2015, and December 31, 2018, were included. METHODS: Single-surgery anatomic success (SSAS) rate and risk factors associated with surgical failure were evaluated. A multivariable logistic regression model was completed to assess the effect of demographic, clinical, and operative variables on SSAS rate. RESULTS: A total of 499 eyes of 499 patients were included. Overall SSAS rate was 86% (n = 430 of 499). Using multivariate analysis, surgical failure was more likely in males (adjusted odds ratio [adjusted OR] = 2.98; 95% CI, 1.58-5.62; p = 0.0007) with a macula-off status on preoperative examination (adjusted OR = 2.15; 95% CI, 1.10-4.20; p = 0.03) and preoperative proliferative vitreoretinopathy (adjusted OR = 4.26; 95% CI, 1.10-16.5; p = 0.04). Time interval between initial examination and surgery (p = 0.26), distribution of buckle or band material used (p = 0.88), and distribution of tamponade used (p = 0.74) were not significantly different between eyes with and without surgical failure. CONCLUSION: Male sex, macula-off status, and preoperative proliferative vitreoretinopathy were factors with increased odds of surgical failure after primary SB for RRD repair. Operative characteristics, such as type of band or use of tamponade, were not associated with surgical failure.

Myo/Nog Cells Give Rise to Myofibroblasts During Epiretinal Membrane Formation in a Mouse Model of Proliferative Vitreoretinopathy.

Purpose: Myo/Nog cells are the source of myofibroblasts in the lens and synthesize muscle proteins in human epiretinal membranes (ERMs). In the current study, we examined the response of Myo/Nog cells during ERM formation in a mouse model of proliferative vitreoretinopathy (PVR). Methods: PVR was induced by intravitreal injections of gas and ARPE-19 cells. PVR grade was scored by fundus imaging, optical coherence tomography, and histology. Double label immunofluorescence localization was performed to quantify Myo/Nog cells, myofibroblasts, and leukocytes. Results: Myo/Nog cells, identified by co-labeling with antibodies to brain-specific angiogenesis inhibitor 1 (BAI1) and Noggin, increased throughout the eye with induction of PVR and disease progression. They were present on the inner surface of the retina in grades 1/2 PVR and were the largest subpopulation of cells in grades 3 to 6 ERMs. All α-SMA-positive (+) cells and all but one striated myosin+ cell expressed BAI1 in grades 1 to 6 PVR. Folds and areas of retinal detachment were overlain by Myo/Nog cells containing muscle proteins. Low numbers of CD18, CD68, and CD45+ leukocytes were detected throughout the eye. Small subpopulations of BAI1+ cells expressed leukocyte markers. ARPE-19 cells were found in the vitreous but were rare in ERMs. Pigmented cells lacking Myo/Nog and muscle cell markers were present in ERMs and abundant within the retina by grade 5/6. Conclusions: Myo/Nog cells differentiate into myofibroblasts that appear to contract and produce retinal folds and detachment. Targeting BAI1 for Myo/Nog cell depletion may be a pharmacological approach to preventing and treating PVR.