
Glaucoma, a multifaceted and frequently asymptomatic disease, remains a leading cause of irreversible blindness worldwide. Despite technological advancements in ophthalmic diagnostics, the nuanced onset and progression of glaucoma pose considerable detection and management challenges. My research leverages artificial intelligence (AI) to innovate these paradigms, enhancing both diagnosis and therapeutic strategies.
AI in Early Glaucoma Detection and Diagnosis
Prompt detection is crucial in mitigating glaucoma-induced vision loss. AI’s robust analytical capabilities are transformative, particularly in interpreting multiple imaging modalities such as optical coherence tomography (OCT), fundus photography, and visual field tests. AI algorithms excel in identifying subtle glaucomatous changes from different modalities and merging that are often challenging to clinicians.Â
Most of our research is focused on combining information from multiple imaging and data sources to identify complex and multifactorial glaucoma and its progression.1 A key aspect of my research involves visual field analysis. Visual field tests, which assess central and peripheral vision, can reveal early glaucomatous damage.2 AI models can analyze large pools of visual fields, quantifying visual field deficits and patterns of loss to integrate this information with structural analyses from fundus or OCT, thereby offering more comprehensive diagnostic and prognostic profiles. This integrated approach aids prediction of glaucoma onset and progression, potentially years before traditional clinical diagnosis can achieve.Â
My research group at the University of Tennessee, along with others, have developed AI models that enhance the interpretation visual field data, fundus, and OCT, focusing on the early identification and stratification of glaucoma risk. These models utilize advanced machine learning techniques to detect and predict the anatomical and functional changes characteristic of glaucoma, thus facilitating timely and targeted interventions.3
Case Studies of AI for Glaucoma
Recent clinical implementations of AI underscore its potential to revolutionize glaucoma screening and diagnosis. To that end, the Centers for Disease Control and Prevention (CDC) has initiated efforts to perform comparative effectiveness trials to investigate the feasibility, effectiveness, and cost-effectiveness of telehealth-based interventions to detect and manage glaucoma among high-risk populations. In a large study, AI algorithms were employed to interpret visual field data, achieving diagnostic accuracies comparable to those of glaucoma specialists while significantly reducing evaluation times.4 These findings suggest that AI can substantially increase diagnostic throughput without compromising accuracy, essential in settings with high patient volumes.5
AI in Treatment Planning and Management
AI’s utility extends beyond screening and diagnostic applications to encompass personalized treatment planning. By longitudinally analyzing patient data, AI algorithms can forecast disease trajectory and response to various therapeutic options. This prognostic capability enables clinicians to tailor treatments to individual patient profiles, optimizing therapeutic outcomes and potentially slowing disease progression.3
Future Prospects
One of the potential applications of AI in glaucoma lies in home monitoring given the fact that glaucoma is chronic and slowly progressing. The integration of AI with at home intraocular pressure (IOP) monitoring, including connected wearable technology, are poised to further personalize glaucoma management. Given the volumes of data generated, AI can help clinicians interpreting data and managing these patients more effectively. These advancements promise to enhance the precision of interventions and improve prognostic outcomes for patients.
References
- Huang X, Sun J, Gupta K, Montesano G, Crabb DP, Garway-Heath DF, Brusini P, Lanzetta P, Oddone F, Turpin A, McKendrick AM, Johnson CA, et al. Detecting glaucoma from multi-modal data using probabilistic deep learning. Front Med (Lausanne). 2022;9:923096.PMC9556968
- Yousefi S, Elze T, Pasquale LR, Saeedi O, Wang M, Shen LQ, Wellik SR, De Moraes CG, Myers JS, Boland MV. Monitoring Glaucomatous Functional Loss Using an Artificial Intelligence-Enabled Dashboard. Ophthalmology. 2020;127(9):1170-1178.PMC7483368
- Huang X, Poursoroush A, Sun J, Boland MV, Johnson CA, Yousefi S. Identifying Factors Associated With Fast Visual Field Progression in Patients With Ocular Hypertension Based on Unsupervised Machine Learning. Journal of glaucoma. 2024;33(11):815-822.PMC11534539
- Yousefi S, Pasquale LR, Boland MV, Johnson CA. Machine-Identified Patterns of Visual Field Loss and an Association with Rapid Progression in the Ocular Hypertension Treatment Study. Ophthalmology. 2022;129(12):1402-1411.PMC9691587
- Lee T, Jammal AA, Mariottoni EB, Medeiros FA. Predicting Glaucoma Development With Longitudinal Deep Learning Predictions From Fundus Photographs. Am J Ophthalmol. 2021;225:86-94.PMC8239478
