Artificial Intelligence and Ophthalmic Clinical Registries.

PURPOSE: The recent advances in artificial intelligence (AI) represent a promising solution to increasing clinical demand and ever limited health resources. Whilst powerful, AI models require vast amounts of representative training data to output meaningful predictions in the clinical environment. Clinical registries represent a promising source of large volume real-world data which could be used to train more accurate and widely applicable AI models. This review aims to provide an overview of the current applications of AI to ophthalmic clinical registry data. DESIGN AND METHODS: A systematic search of EMBASE, Medline, PubMed, Scopus and Web of Science for primary research articles that applied AI to ophthalmic clinical registry data was conducted in July 2024. RESULTS: Twenty-three primary research articles applying AI to ophthalmic clinic registries (n = 14) were found. Registries were primarily defined by the condition captured and the most common conditions where AI was applied were glaucoma (n = 3) and neovascular age-related macular degeneration (n = 3). Tabular clinical data was the most common form of input into AI algorithms and outputs were primarily classifiers (n = 8, 40%) and risk quantifier models (n = 7, 35%). The AI algorithms applied were almost exclusively supervised conventional machine learning models (n = 39, 85%) such as decision tree classifiers and logistic regression, with only 7 applications of deep learning or natural language processing algorithms. Significant heterogeneity was found with regards to model validation methodology and measures of performance. CONCLUSIONS: Limited applications of deep learning algorithms to clinical registry data have been reported. The lack of standardized validation methodology and heterogeneity of performance outcome reporting suggests that the application of AI to clinical registries is still in its infancy constrained by the poor accessibility of registry data and reflecting the need for a standardization of methodology and greater involvement of domain experts in the future development of clinically deployable AI.

Does Bevacizumab Improve Outcomes After Ahmed Glaucoma Valve Implantation for Refractory Glaucoma?: A Meta-Analysis.

PRCIS: Mean intraocular pressure (IOP), complete and overall success, mean IOP-lowering medications, incidence of hypertensive phase, and complications were found to be comparable between patients undergoing Ahmed glaucoma valve implantation (AGVI) with adjunctive bevacizumab versus AGVI alone. OBJECTIVE: This meta-analysis aims to assess how adjunctive bevacizumab impacts the surgical outcomes of AGVI compared with AGVI alone in all subtypes of refractory glaucoma. METHODS: A systematic search of databases for relevant randomized controlled trials (RCTs) was performed in March 2023. Primary outcomes included mean IOP and success rates. Secondary outcomes were mean IOP-lowering medications, incidence of hypertensive phase, and complications. Qualitative assessment, meta-analysis, subgroup analyses, and sensitivity analysis were performed. RESULTS: Five RCTs comprising 203 eyes were included in the quantitative analysis. Initial meta-analysis showed a strong yet nonsignificant trend (all P > 0.05) favoring adjunctive bevacizumab in all outcomes of interest. Significant heterogeneity was observed for mean IOP and success outcomes at all time points (all I2 > 50%). Subgroup analysis of the administration route revealed a reduced incidence of hyphaema in the intravitreal bevacizumab subgroup (odds ratio: 0.10; 95% CI: 0.02 to 0.59; P = 0.01) with significant heterogeneity persisting in the intravitreal bevacizumab subgroup for all measures (all I2 > 50%). Post hoc sensitivity analysis of studies without concurrent pan-retinal photocoagulation for mean IOP and success outcomes demonstrated more conservative effect sizes with a corresponding decrease in heterogeneity for all measures (all I2 < 30%). CONCLUSION: Published studies investigating the role of adjunctive bevacizumab show a strong trend to improve outcomes but contain a relatively small number of participants. This analysis underpins the need for an adequately powered RCT to explore the role of anti-vascular endothelial growth factor agents in AGVI surgery.

Recurrent post-operative endophthalmitis caused by Sphingomonas paucimobilis despite vitrectomy - a case and review of the literature.

Over the past 20 years, scattered reports have emerged about a low virulence, gram negative bacillus, Sphingomonas paucimobilis, causing unpredictable clinical presentations of endophthalmitis. Previous reports have characterised the organism as being resistant to aggressive treatment and prone to recurrence up to several months later, with few warning signs of any residual infection. We report a case of a 75 year-old male who returned 10 days after a left eye cataract surgery with an atypical, indolent endophthalmitis. He was treated with broad-spectrum intravitreal antibiotics and vitrectomy, and despite initial improvement, the patient suffered a recurrence after 2 weeks, necessitating successive rounds of intravitreal antibiotics. While our patient was able to achieve an excellent final visual acuity of 6/9, there are several cases in the literature reporting similar cases with much worse visual outcomes. Further research is required to elucidate early warning signs that may indicate a recurrence of S. paucimobilis infection, and the underlying mechanism by which it is resistant to standard endophthalmitis therapy. Alongside this case, we review and summarise the literature on postoperative endophthalmitis involving this organism.

Development, implementation, and initial results of the UC San Diego Health Moores Cancer Center Wellbeing Screening Tool.

OBJECTIVE: All accredited cancer institutions are required to screen patients for psychosocial distress. This paper describes the development, implementation, and preliminary outcomes of the University of California San Diego Health Moores Cancer Center Wellbeing Screening Program. METHOD: Essential steps learned in a formal National Cancer Institute-funded training workshop entitled "Implementing Comprehensive Biopsychosocial Screening" were followed to ensure successful program implementation. These steps included identification of stakeholders; formation of a working committee; establishment of a vision, process, and implementation timeline; creation of a screening tool; development of patient educational material; tool integration into an electronic medical record system; staff training and pilot testing of tool administration; and education about tool results and appropriate follow-up actions. Screening data were collected and analyzed retrospectively for preliminary results and rapid cycle improvement of the wellbeing screening process. RESULTS: Over an 8-month implementation and assessment period, the screening tool was administered 5,610 times of 7,664 expected administrations (73.2%.) to 2,394 unique patients. Visits in which the questionnaire was administered averaged 39.6 ± 14.8 minutes, compared with 40.3 ± 15.2 minutes for visits in which the questionnaire was not administered (t = -1.76, df = 7,662, p = 0.079). SIGNIFICANCE OF RESULTS: This program provides a process and a tool for successful implementation of distress screening in cancer centers, in a meaningful way for patients and providers, while meeting accreditation standards. Further, meaningful data about patient distress and tool performance were able to be collected and utilized.