Starr, Matthew R.

ImportanceSpecialized ophthalmology terminology limits comprehension for nonophthalmology clinicians and professionals, hindering interdisciplinary communication and patient care. The clinical implementation of large language models (LLMs) into practice has to date been relatively unexplored.ObjectiveTo evaluate LLM–generated plain language summaries (PLSs) integrated into standard ophthalmology notes (SONs) in improving diagnostic understanding, satisfaction, and clarity.Design, Setting, and ParticipantsRandomized quality improvement study conducted from February 1, 2024, to May 31, 2024, including data from inpatient and outpatient encounters in a single tertiary academic center. Participants were nonophthalmology clinicians and professionals and ophthalmologists. The single inclusion criterion was any encounter note generated by an ophthalmologist during the study dates. Exclusion criteria were (1) lack of established nonophthalmology clinicians and professionals for outpatient encounters and (2) procedure-only patient encounters.InterventionAddition of LLM-generated plain language summaries to ophthalmology notes.Main Outcomes and MeasuresThe primary outcome was survey responses from nonophthalmology clinicians and professionals assessing understanding, satisfaction, and clarity of ophthalmology notes. Secondary outcomes were survey responses from ophthalmologists evaluating PLS in terms of clinical workflow and accuracy, objective measures of semantic quality, and safety analysis.ResultsA total of 362 (85%) nonophthalmology clinicians and professionals (33.0% response rate) preferred the PLS to SON. Demographic data on age, race and ethnicity, and sex were not collected. Nonophthalmology clinicians and professionals reported enhanced diagnostic understanding (percentage point increase, 9.0; 95% CI, 0.3-18.2; P = .01), increased note detail satisfaction (percentage point increase, 21.5; 95% CI, 11.4-31.5; P < .001), and improved explanation clarity (percentage point increase, 23.0; 95% CI, 12.0-33.1; P < .001) for notes containing a PLS. The addition of a PLS was associated with reduced comprehension gaps between clinicians who were comfortable and uncomfortable with ophthalmology terminology (from 26.1% [95% CI, 13.7%-38.6%; P < .001] to 14.4% [95% CI, 4.3%-24.6%; P > .06]). PLS semantic analysis found high meaning preservation (bidirectional encoder representations from transformers score mean F1 score: 0.85) with greater readability than SONs (Flesch Reading Ease: 51.8 vs 43.6; Flesch-Kincaid Grade Level: 10.7 vs 11.9). Ophthalmologists (n = 489; 84% response rate) reported high PLS accuracy (90% [320 of 355] a great deal) with minimal review time burden (94.9% [464 of 489] ≤1 minute). PLS error rate on ophthalmologist review was 26% (126 of 489). A total of 83.9% (104 of 126) of errors were deemed low risk for harm and none had a risk of severe harm or death.Conclusions and RelevanceIn this study, use of LLM-generated PLSs was associated with enhanced comprehension and satisfaction among nonophthalmology clinicians and professionals, which might aid interdisciplinary communication. Careful implementation and safety monitoring are recommended for clinical integration given the persistence of errors despite physician review.

Purpose: To report the anatomic and visual outcomes of eyes developing delayed rhegmatogenous retinal detachment (RRD) after macular surgery. Methods: This was a retrospective, multicentric study including eyes undergoing RRD repair at least 3 months after macular surgery for either full-thickness macular hole (FTMH) or epiretinal membrane (ERM). Anatomic and functional assessment, i.e., pre- and postoperative best-corrected visual acuity (BCVA), indications and details of macular and RRD surgery, time interval from macular surgery to RRD, intraocular tamponade, location, extent of RRD, and nature of retinal break, was performed. Results: Sixteen out of 2,906 eyes (0.55%) undergoing macular surgery developed RRD after 8.4 ± 5.8 months (mean ± SD). Intraoperative peripheral retina evaluation showed missed/de novo break (11 eyes; 68.75%), no identifiable break i.e., probable microbreak (two eyes; 12.5%), giant retinal tear (two eyes; 12.5%), and retinoschisis related break (one eye; 6.25%). Anatomical success, i.e., attached retina after one surgery, was seen in 14 eyes (87.5%). BCVA worsened from 0.6 ± 0.3 logMAR (Snellen equivalent 20/80) at baseline to 0.9 ± 0.8 logMAR (20/160) through the last visit (P < 0.001). Conclusion: RRD after macular surgery occurs predominantly via missed/de novo retinal breaks. A significant percentage (50%) of eyes showed worse visual acuity compared with baseline despite successful retinal reattachment.

Purpose of review Large language models (LLMs) are rapidly entering the landscape of medicine in areas from patient interaction to clinical decision-making. This review discusses the evolving role of LLMs in ophthalmology, focusing on their current applications and future potential in enhancing ophthalmic care. Recent findings LLMs in ophthalmology have demonstrated potential in improving patient communication and aiding preliminary diagnostics because of their ability to process complex language and generate human-like domain-specific interactions. However, some studies have shown potential for harm and there have been no prospective real-world studies evaluating the safety and efficacy of LLMs in practice. Summary While current applications are largely theoretical and require rigorous safety testing before implementation, LLMs exhibit promise in augmenting patient care quality and efficiency. Challenges such as data privacy and user acceptance must be overcome before LLMs can be fully integrated into clinical practice.

Background: Retrospective cohort study of 561 adult patients undergoing secondary intraocular lens (IOL) implantation by vitreoretinal surgeons at a single institution from April 2015 to December 2020. Methods: Patient historical factors, intraoperative/postoperative complications, and outcomes of IOL type (anterior chamber IOL versus scleral sutured IOL versus scleral fixated IOL versus. sulcus) were assessed. Primary outcomes were rates of postoperative retinal tears and rhegmatogenous retinal detachment. Secondary outcomes were rates of intraoperative endolaser, intraoperative retinal tear, and further IOL surgery. Results: The incidence of intraoperative retinal tears was 7.3% and not significantly different between techniques. Rates of intraoperative endolaser use were 17.5% among all techniques and not significantly different between techniques. Rates of postoperative retinal tear were low (0%–2.7%). Rates of postoperative rhegmatogenous retinal detachment were not significantly different between techniques (anterior chamber IOL 9/198 [4.5%], SFIOL 1/54 [1.9%], scleral sutured IOL 14/274 [5.1%], sulcus 2/35 [5.7%], total 26/561 [4.6%], P = 0.79). Rates of repeat IOL surgery trended higher in sulcus lenses (anterior chamber IOL 5/198 [2.5%], SFIOL 4/54 [7.4%], scleral sutured IOL 16/274 [5.8%], sulcus 5/35 [14.3%], total 30/561 [5.3%], P = 0.12). Conclusion: Intraoperative endolaser use and intraoperative retinal tear are not uncommon in secondary IOL surgery and underscore the importance of careful vitreoretinal management among these patients.

Purpose: To evaluate anatomic outcomes and surgeon response following the use of microserrated (Sharkskin, Alcon, Forth Worth, TX) internal limiting membrane (ILM) forceps compared with conventional (Grieshaber; Alcon) ILM forceps for peeling of the ILM. Methods: Patients were prospectively assigned in a 1:1 randomized fashion to undergo ILM peeling using microserrated forceps or conventional forceps. Rates of retinal hemorrhages, deep retinal grasps, ILM regrasping, time to ILM removal, and surgeon questionnaire comparing the use of microserrated and conventional ILM forceps were analyzed. Results: A total of 90 eyes of 90 patients were included in this study. The mean number of deep retinal grasps was higher in the conventional forceps group (1.51 ± 1.70 vs. 0.33 ± 0.56, respectively [P < 0.0001]). The mean number of failed ILM grasps was higher with conventional forceps (6.62 ± 3.51 vs. 5.18 ± 2.06 [P = 0.019]). Microserrated forceps provided more comfortability (lower number) in initiating the ILM flap (2.16 ± 0.85 vs. 1.56 ± 0.76, P < 0.001), comfortability in regrasping the ILM flap (2.51 ± 1.01 vs. 1.98 ± 0.89, P = 0.01), and comfortability in completing the ILM flap (2.42 ± 1.03 vs. 1.84 ± 1.02, P = 0.01). Conclusion: Surgeons utilizing the microserrated forceps experienced fewer deep retina grasps and fewer failed ILM grasps compared with conventional ILM forceps. The microserrated forceps was also a more favorable experience subjectively among the surgeons.

Purpose of review To highlight the recent work published from the Primary Retinal Detachment Outcomes (PRO) Study Group. Recent findings The PRO database was a large dataset made up of patients with primary rhegmatogenous retinal detachments (RRD) who had surgical repair during 2015. The database was constituted of nearly 3000 eyes from 6 centers across the United States and included 61 vitreoretinal surgeons. Nearly 250 metrics were collected for each patient, creating one of the richest datasets of patients with primary rhegmatogenous detachments and their outcomes. The importance of scleral buckling was demonstrated, particularly for phakic eyes, elderly patients, and those with inferior breaks. 360° laser may result in poorer outcomes. Cystoid macular edema was common, and risk factors were identified. We also found risk factors for vision loss in eyes presenting with good vision. A PRO Score was devised, to predict outcomes based on presenting clinical characteristics. We also identified characteristics of surgeons with the highest single surgery success rates. Overall, there were no major outcome differences between viewing systems, gauges, buckles sutured vs. scleral tunnels, drainage method, and techniques to address proliferative vitreoretinopathy. All incisional techniques were found to be very cost-effective treatment modalities. Summary Numerous studies resulted from the PRO database that significantly added to the literature regarding the repair of primary RRDs in the current era of vitreoretinal surgery.

Purpose: In this study, the authors aimed to identify the demographics of retinal detachment in children in the United States and to report the treatment modalities chosen by vitreoretinal surgeons and associated outcomes. Methods: This was a multicenter cohort analysis of the Vestrum Health LLC Database (Naperville, IL). Children (1–17 years) with rhegmatogenous retinal detachment, as identified by ICD10 and CPT codes, between January 2015 and August 2021 were enrolled. Demographic, comorbidity, visual acuity, and treatment modality data were analyzed. Results: A total of 168,152 RRDs were identified, of which 2,200 (1.3%) were aged 1 to 17 years. The mean age was 12.7 years, and 821 (37%) were women. The prevalence of rhegmatogenous retinal detachment increased with age (P = 0.009). Associated comorbidities included myopia (17.3%), ocular trauma (7.5%), and history of prematurity (5.7%). Laser retinopexy alone was used as the initial treatment modality in 19%, primary vitrectomy in 23%, primary scleral buckle in 25%, and vitrectomy with scleral buckle in 33%. The single surgery success rate for all procedures was 73.3%. Of the incisional surgical modalities, primary scleral buckling had the best single surgery success rate (79.0%) compared with vitrectomy alone (64.5%) and vitrectomy with scleral buckle (67.2%) (P < 0.001 and P = 0.004, respectively). Younger age resulted in worse SSSR overall (coefficient = 0.151, R2 = 0.746, P = 0.027). Conclusion: Rhegmatogenous retinal detachments in children increased with age. Myopia, trauma, and history of retinopathy of prematurity were common risk factors. Treatment techniques varied, but primary scleral buckling had the best anatomical and visual outcomes.

Purpose: To evaluate the impact of anti–vascular endothelial growth factor therapy on the refractive error in eyes with diabetic macular edema. Methods: Post hoc analysis of Diabetic Retinopathy Clinical Research protocol T. Spherical equivalent (SE) was calculated for study and fellow eyes at baseline, 1-year, and 2-year visits. The SE change of the eyes with edema was compared with those with resolved edema. The correlation between refractive error changes and central subfield thickness was evaluated. Results: Among 543 study eyes, SE changed from −0.17 (2.04) D at baseline visit to −0.16 (2.14) D at 2 years giving a hyperopic shift of 0.04 (0.82) D (P = 0.022). Among fellow eyes, mean (SD) SE changed from −0.19 (2.1) D at baseline to −0.11 (2.1) D at 2 years, giving a hyperopic shift of 0.12 (0.84) D (P = 0.001). No significant difference in SE shift was found between eyes with and without edema at 2 years in phakic (0.12D and 0.08 D, P = 0.87) and pseudophakic eyes (−0.24D and −0.08D, P = 0.30). The SE shift was not correlated with central subfield thickness change at the end of the second year (r = 0.02, P = 0.62). Conclusion: Diabetic macular edema patients have minimal changes in refractive error. The correction of refractive error may be considered during treatment, regardless of the presence of edema.