Canadian journal of ophthalmology. Journal canadien d'ophtalmologie最新文献

Objective: 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.

Design: Prospective qualitative study.

Participants: Microsoft Copilot, GPT-4 (GPT-4o and GPT-4o mini), and Google Gemini (Gemini and Gemini Advanced).

Methods: 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).

Results: 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.

Conclusions: 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.

Objective: Assess safety and effectiveness of subretinal gene replacement therapy at 18 months post treatment.

Design: Retrospective, longitudinal study conducted at the Hospital for Sick Children in Toronto, Canada.

Participants: Patients with bi-allelic RPE65 variants, early onset retinal degeneration, and residual viable retina who underwent voretigene neparvovec r-zyl gene replacement therapy.

Methods: Data collected included demographic information, molecular genetic results, and comprehensive ocular assessment results from preoperataive and postoperative visits up to 18 months. To assess the treatment's efficacy, postoperative best corrected visual acuity, full-field stimulus test (FST), visual field (VF) area, optical coherence tomography, and global satisfaction were compared to preoperative findings.

Results: The procedures were safe with no complications. There was no significant improvement in visual acuity. Three eyes showed a slight reduction in VF area. All showed a reduction in ellipsoid zone thickness and area, but the outer nuclear layer thickness and area were stable. All had a significant improvement in retinal sensitivity, as per FST, allowing better navigation in a dim environment. All 3 patients reported being "very satisfied".

Conclusion: Following a safe gene replacement therapy, the 3 first Canadian cases had an improvement in retinal sensitivity as per FST, and our patients described their experience as positively life changing.

Objective: Biallelic RPE65 pathogenic variants may cause Leber congenital amaurosis (LCA). Voretigene neparvovec-rzyl (VN, Luxturna) is the only approved subretinal gene therapy that demonstrated benefit and safety. The eligibility criteria are vague and variable between centres. This is the first comprehensive outcome report of RPE65-LCA patients with World Health Organization blindness criteria vision treated with VN.

Design: Multicentre retrospective case series.

Participants: Patients meeting the treatment criteria for VN who had best-corrected visual acuity (BCVA) <20/400 or visual field (VF)III4e isopter <10°.

Methods: Patients were followed for a mean of 11.1 ± 4.7 months. Age, sex, BCVA, central retinal thickness (CRT), retinal atrophy, VF, full-field stimulus testing (FST), and subjective impressions were assessed.

Results: Nine patients met the inclusion criteria (mean: BCVA 1.89 LogMAR, range: 1.4 - 2.7 LogMAR, mean age: 28.7-years-old, range: 17-59 years). Though VF area did not improve, FST improved in patients with better baseline FST (-8.83 dB vs -0.56 dB; p = 0.010), and better VFV4e (7245 vs 341^o2; p < 0.001) and III4e (596.1 vs 24.8^o2; p = 0.011) area. VA improved in younger (20 vs 32 years; p = 0.011) patients with thinner CRT^1mm (155 vs 193 µm; p = 0.038). VFV4e loss occurred in older (38 vs 19 years; p = 0.001) patients with worse baseline V4e area (1728 vs 8159^o2; p < 0.001). Subjective improvement in dim light navigation skills occurred in younger patients (20.3 vs 45.3 years; p < 0.001).

Conclusions: Blindness is not a contraindication to VN treatment for RPE65-LCA. Superior results correlated with greater baseline FST but not with CRT^1mm, provided that measurable outer retinal structures persist.

Objective: To conduct a pilot study to evaluate and compare the 24-hour habitual intraocular pressure (IOP) and ocular perfusion pressure (OPP) fluctuation in glaucoma patients treated with medical therapy, selective laser trabeculoplasty (SLT) or trabeculectomy.

Design: Pilot study.

Participants: Criteria for inclusion were patients aged 18 years or older with well-controlled IOP with either maximum tolerated medical therapy, previous SLT, or previous trabeculectomy.

Methods: Recruited patients were admitted to the sleep lab for 24-hour serial habitual IOP and blood pressure measurements. IOP and OPP fluctuation among the 3 treatment groups were compared.

Results: Thirty three (33) eyes from 33 patients were recruited in this study, including 11 patients in the medical therapy group, 11 patients in the SLT group, and 11 patients in the trabeculectomy group. The medical therapy group was found to have significantly higher 24-hour IOP fluctuation (8.3 ± 1.6 mmHg) than the SLT (3.5 ± 1.9 mmHg) and trabeculectomy (4.3 ± 1.3 mmHg) groups (P < 0.001). Mean 24-hour OPP fluctuation was also significantly higher in the medical therapy group (18.5 ± 4.0 mmHg) than the SLT (11.9 ± 7.3 mmHg) and trabeculectomy (14.1 ± 3.9 mmHg) groups (P < 0.05). No difference in IOP or OPP fluctuation was found between SLT and trabeculectomy groups (P > 0.05).

Conclusions: Both SLT and trabeculectomy may be more effective in reducing 24-hour IOP and OPP fluctuation than medical therapy alone. IOP and OPP fluctuation was comparable between SLT and trabeculectomy cohorts. Future studies are warranted to investigate this further.