Acta Ophthalmologica最新文献

Agentic artificial intelligence (AI) represents a new generation of intelligent systems capable of autonomous goal-directed reasoning, dynamic decision-making, and coordinated action. Unlike conventional task-specific algorithms, agentic AI systems can perceive, plan, and act within complex healthcare environments with minimal supervision. Ophthalmology, an image-intensive and data-rich specialty, provides an ideal field for exploring this paradigm. This Perspective outlines the conceptual framework of agentic AI, its emerging ophthalmic applications, and the technical, ethical, and regulatory challenges that must be addressed to ensure safe and effective clinical translation. Properly developed, agentic AI has the potential to enable intelligent diagnostic assistants, adaptive surgical partners, and autonomous population health agents; thereby, it has the potential to transform future models of ophthalmic care.

The Central Nervous System (CNS), comprising the brain and the eye, is considered to have a 'privileged' mechanism for dealing with immunological challenge (immune privilege, IP). CNS IP has been revealed through experiments using foreign protein antigens and cell and tissue alloantigens (grafts), but evidence for a role for IP in modulating host-pathogen interactions in the CNS is limited. However, the low frequency of CNS infection in the face of widespread systemic exposure to CNS-tropic infectious agents, together with the high incidence of CNS infection in immunocompromised individuals, suggests that in healthy individuals, the CNS has tightly controlled regulatory mechanisms to protect against infectious agents. Although the naïve healthy brain and retina parenchyma largely lack adaptive immune cells, their border tissues (meninges, uveal tract) contain a full complement of resident immune cells, including CNS-specific regulatory T cells (Tregs), which have a fundamental role in controlling infection in the brain parenchyma. Tregs also underpin ocular IP, particularly of the neural retina. Recent studies report that Tregs are transcriptionally 'customised' to the CNS and function at a distance; that is, are located in niches/hubs around the venous sinuses of the border tissues. T cells resident in the uveal tract probably play a similar role. We propose that Tregs are key drivers of CNS IP and do so by promoting latency of infectious agents.

Congenital aniridia is a rare genetic disorder primarily caused by pathogenic variants of the PAX6 gene. It leads to various panocular anomalies, including aniridia-associated keratopathy (AAK). This review highlights recent insights into its pathogenesis, focusing on clinical staging, microstructural changes in the cornea and molecular dysregulation. We synthesized clinical and experimental findings from large European cohorts, integrating data on over 550 eyes. AAK severity correlates with iris malformation, secondary glaucoma and lens status. In vivo confocal microscopy reveals a reduction in subbasal nerve plexus density, altered keratocyte and endothelial morphology and an increase in Langerhans cell infiltration. RNA and miRNA microarrays, as well as RNA-seq studies, highlight dysregulated miRNAs (such as miR-204-5p and miR-138-5p) and altered expression of PAX6 and keratocyte markers. Limbal fibroblasts show enhanced inflammatory responses and vulnerability to oxidative stress. Advanced AAK is associated with a reduced quality of life. The progression of AAK involves intricate interactions between developmental deficits, inflammation and changes in the limbal microenvironment, suggesting molecular targets for future therapies.

Purpose: To compare retinal vessel oxygen saturations and diameters before, 1 and 3 months after trabeculectomy.

Methods: Retinal oxygen saturations and vessel diameters were obtained using retinal oximetry. The effect of the postoperative intraocular pressure (IOP) decrease on the arteriovenous difference (AV-difference) was evaluated, adjusting for mean defect (MD). The impact of trabeculectomy on AV-difference was compared between lowest (Q1) and highest (Q4) MD quartiles. Paired t-test, linear regression and Welch's t-test were applied.

Results: Of 72 enrolled participants, 64 eyes of 64 participants were included. Arteriolar oxygen saturations were unchanged (97.4 ± 4.9% preoperatively, 98.1 ± 6.0% 1 month [p = 0.053], 98.2 ± 6.8% 3 months postoperatively [p = 0.154]). Venular saturations were also unchanged (62.0 ± 9.5% preoperatively, 61.7 ± 8.0% 1 month [p = 0.505], 61.3 ± 10.6% 3 months postoperatively [p = 0.685]). No significant difference in AV-difference occurred. Arteriolar diameters were unchanged 1 month postoperatively (109.7 μm ± 16.2 vs. 110.9 μm ± 19.3 [p = 0.80]) but decreased to 106.0 μm ± 15.7 3 months postoperatively (p = 0.012). Venular diameters increased 1 month postoperatively (152.8 μm ± 21.3 vs. 147.7 μm ± 20.0 [p = 0.034]) but not 3 months postoperatively (148.6 μm ± 24.8; p = 0.893). No association between postoperative IOP decrease and postoperative MD-adjusted AV-difference was found (p = 0.380). No significant change in AV-difference between Q1 and Q4 (p = 0.18) was found.

Conclusion: Retinal oxygen saturations remained stable, whereas vessel diameter changes occurred after trabeculectomy, probably reflecting alterations in IOP and perfusion pressure.

Purpose: BEST1 variants are the third leading cause of inherited retinal dystrophies in Norway. The purpose of this study was to describe the BEST1-associated retinal dystrophy (BEST1-RD) population genetically and clinically, and to determine the prevalence of BEST1-RD in Southern and Eastern Norway.

Methods: This registry-based study used the Oslo University Hospital Inherited Retinal Disease registry for genetic data and extracted clinical data from medical records.

Results: Sixty patients were included. The prevalence of BEST1-RD in Southern and Eastern Norway was between 1:64 600 and 1:43 700. Forty-one patients were diagnosed with Best's vitelliform macular dystrophy (BVMD), 15 with autosomal recessive bestrophinopathy (ARB) and 4 with autosomal dominant vitreoretinochoroidopathy (ADVIRC). The two most common genotypes were c.403G>A and c.89A>G. These variants were associated with BVMD and a later age of onset compared with other BVMD-associated genotypes.

Conclusion: The prevalence of BEST1-RD in Southern and Eastern Norway was between 1:64 600 and 1:43 700. BVMD patients carrying c.403G>A or c.89A>G had a later age of onset than BVMD patients carrying other variants.

Purpose: Managing ocular diseases often requires frequent eye drop administration, which can challenge patient compliance. A long-acting eye drop technology using an amorphous synthetic silica composite was developed to address this. Our study aimed to assess the safety and tolerability of the Silica Eye Drop platform in healthy volunteers over 15 days.

Methods: Twelve healthy volunteers participated in a randomized, double-blinded, placebo-controlled trial of the Silica Eye Drop Product, containing no active substance. Participants received one drop in one eye daily for 13 days, with the other eye serving as an untreated control. Safety and tolerability were evaluated through various examinations at multiple time points. Ocular discomfort was assessed with a questionnaire at these times, and additional evaluations of lens, vitreous body, visual acuity (BCVA), ocular protection index (OPI), and intraocular pressure (IOP) were performed at t0 and D15.

Results: No significant differences in ocular metrics between treated and untreated eyes were observed after 15 days of Silica Eye Drop application (IOP: control 14.1 ± 2.02 mmHg, treated 13.6 ± 1.75 mmHg; BCVA: control 1.22 ± 0.16, treated 1.24 ± 0.15; OPI: control 1.66 ± 0.43, treated 1.63 ± 0.40). Questionnaire responses indicated that 68% of volunteers experienced mild discomfort during the product application, while 32% noted moderate discomfort. The average pleasantness score was 4.9 ± 1.83 using a 10-point scale, indicating acceptable tolerability of Silica Eye Drops.

Conclusions: The findings suggest that Silica Eye Drops are safe and well-tolerated by study subjects when used once daily. This supports further developing sustained release topical ocular products for delivering pharmacological treatments in various eye conditions.

Objective: To assess the adherence of glaucoma surgical and laser studies to WGA guidelines for reporting glaucoma surgery studies, analyse trends in adherence over time and explore associations between adherence and study characteristics.

Methods: Systematic review (PROSPERO:CRD42023394477) of glaucoma surgical and laser studies published between 2010 and 2023 in PubMed/MEDLINE and EMBASE. Eligible studies included RCTs, non-randomized comparative and prospective observational designs (>100 eyes). Two reviewers independently extracted data across five domains: Methodology, Definition of success, Ethics, Postoperative complications and Statistical reporting. Temporal trends and associations with study features were analysed using linear regression.

Results: Two hundred and fifty-six studies were included, 75% of which were published in Q1-Q3 journals. Mean overall adherence was 47% ± 9.2%. Domain-level adherence was highest in Ethics (61% ± 20%), followed by Postoperative complications (50% ± 22%), Statistical reporting (48% ± 18%), Methodology (44% ± 12%) and Definition of success (30% ± 13%). No significant differences (p > 0.06) were observed in overall adherence for studies from Europe, Asia, Oceania or the Middle East. Studies involving cataract surgery for angle-closure disease (est. = -10% [-19%, -2.2%], p = 0.014) and laser trabeculoplasty (est. = -7.1% [-11%, -3.5%], p < 0.001) had lower adherence compared with trabeculectomy, while MIGS studies showed no difference (p = 0.45). Visual field progression was reported in only 3% of studies, while various anatomical outcomes (e.g. bleb morphology) were reported in 0%-24% of studies.

Conclusion: Current literature shows poor adherence to WGA guidelines across both traditional and newer glaucoma surgeries, reflecting inadequate reporting and outdated recommendations. Evidence-based updates, broader consensus and stronger implementation are needed to ensure standardized and meaningful reporting.

Purpose: Norwegian guidelines designate off-label Avastin as the first-line intravitreal therapy for neovascular age-related macular degeneration (nAMD) because of its well-documented clinical efficacy and low price. However, this overlooks non-drug costs, which increase with injection frequency. We evaluated whether newer, longer-acting agents offer greater long-term cost-efficiency by reducing total costs despite higher drug prices in a Norwegian setting.

Methods: We developed a 2-year cost-minimization model that included pharmaceutical, consultation and administrative and patient-related costs in Norway; the pharmaceutical cost component incorporated the routine practice of splitting vials into prefilled syringes in hospital pharmacies. The model compared four nAMD monotherapies, Avastin, Eylea 2 mg, Eylea 8 mg and Vabysmo, as well as the common practice of switching treatment-resistant patients from Avastin to Eylea 2 mg. We derived injection frequencies from clinical trials (for monotherapies) or observational data (for switching) and conducted one-way sensitivity analyses to identify key cost drivers.

Results: Over 2 years, the switching regimen had the highest per-patient cost (146 722 NOK), followed by Eylea 2 mg (100 481 NOK), Vabysmo (93 207 NOK), Avastin (86 262 NOK) and Eylea 8 mg (68 738 NOK). Avastin had the lowest drug cost, but its high injection frequency increased non-drug costs. Sensitivity analyses showed that injection frequency strongly influenced total costs for high-priced drugs, while patient time had a substantial impact for Avastin.

Conclusion: In our model, longer-acting agents reduced injection frequency and decreased overall treatment costs. These findings suggest that adopting longer-acting monotherapy could improve cost-efficiency in long-term nAMD management in Norway.

Purpose: To assess the validity of the HelpMeSee Manual Small Incision Cataract Surgery (MSICS) module as a virtual reality training tool for technical skills and stress management in ophthalmology.

Methods: This prospective study enrolled 47 volunteer surgeons from five groups: four groups of eye surgeons with increasing experience (novice, junior, senior and expert) and a fifth group of experts from other specialties. Participants completed two standardized MSICS training runs on the HelpMeSee simulator. Performance scores, penalties and completion time were recorded. Ergonomics were assessed via the Rapid Upper Limb Assessment (RULA) score, and stress was evaluated subjectively and objectively using the State-Trait Anxiety Inventory-Y (STAI-Y) and the Analgesia Nociception Index (ANI) score. Data from the two runs were analysed and compared across groups.

Results: Overall scores increased significantly from novice residents (32.4 ± 10.7 out of 72) to the expert ophthalmic surgeons (50.1 ± 9.41) (p < 0.001). Non-ophthalmic experts had a lower mean score (16.8 ± 18.0). Total penalties, particularly in the second run, decreased with experience among eye surgeons, while experts from other specialties incurred the highest penalties. Time analysis did not differ between groups, as for RULA or STAI-Y scores. The mean ANI score decreased with experience, suggesting higher stress levels in more experienced participants.

Conclusions: The HelpMeSee MSICS module effectively differentiates surgical experience levels, confirming its validity as a tool for technical skills training. The ANI score demonstrated modified behaviour in expert surgeons, suggesting the simulator's potential for assessing non-technical skills. These findings support the use of this virtual reality simulator for objective, skills-based surgical education.

Conventional optical coherence tomography (OCT) has a floor effect in patients with severe visual field loss, such as seen in advanced primary open-angle glaucoma (POAG). OCT angiography (OCTA) does not suffer from such a floor effect. However, which OCTA parameters are most useful for monitoring longitudinal progression is unclear. We conducted a systematic review and meta-analysis to investigate the clinical use of OCTA in monitoring the progression of intermediate and advanced POAG by searching four databases (Medline, Embase, Web of Science and Cochrane Database of Systematic Reviews) for longitudinal studies on POAG and OCTA. For each meta-analysed OCTA parameter, we calculated a pooled effect estimate with a 95% confidence interval (95% CI). Parameters that could not be meta-analysed were compared narratively. A total of 18 studies were included in the systematic review and seven in the meta-analyses. In the meta-analyses, a lower baseline peripapillary vessel density (VD) significantly increased the risk of visual field (VF) progression (HR [95% CI]: 1.05 [1.02, 1.07] for each percentage decrease in peripapillary VD per year). Baseline parafoveal VD and risk of VF progression showed no significant association. The inferior hemifield foveal avascular zone parameters and the presence of peripapillary choroidal microvascular dropout were significantly associated with the risk of VF progression in the systematic review. Peripapillary VD may be a useful predictor of VF progression in intermediate and advanced glaucoma patients. Although FAZ and MvD also seem to be potential predictors, longitudinal studies on advanced POAG are limited and heterogeneous, highlighting the need for more consistent and comprehensive research.