Artificial Intelligence and the Stewardship Gap in Glaucoma Care: Will Earlier Detection Translate into Less Blindness?

follow-up, delayed treatment escalation, medication adherence challenges, affordability constraints, and fragmented care pathways that fail to support patients through a lifelong disease.[8][9][10] The American Academy of Ophthalmology Preferred Practice Pattern for Primary Open-Angle Glaucoma explicitly frames the condition as requiring continuous, risk-based monitoring rather than episodic intervention.11 Patient-centered care, shared decision-making, and quality-of-life concerns are now recognized as inseparable from clinical management.12,13 AI technologies are well positioned to address the first failure.Their ability to influence the second will determine their true impact on blindness prevention. ArtificiAl intelligence A n d t h e Promise o f eArlier detectionCurrent AI applications in glaucoma have understandably focused on improving diagnostic performance.Deep learning algorithms trained on large imaging datasets demonstrate strong performance in detecting glaucomatous optic neuropathy, identifying retinal nerve fiber layer progression patterns, and supporting population-level screening.2,3 A systematic review and meta-analysis of 48 studies found pooled sensitivity and specificity of 0.92 and 0.93, respectively, for fundus photographybased deep learning systems, with an area under the receiver operating characteristic curve (AUROC) of 0.90, performance comparable to expert clinicians.4 For OCT-based detection, a separate meta-analysis of 51 models reported similarly high pooled diagnostic accuracy.5 If effectively implemented, these technologies may meaningfully shorten the interval between disease onset and diagnosis, an important contributor to preventable vision loss globally.AI may also extend beyond detection into longitudinal care.Potential applications include automated recall systems, riskadjusted follow-up scheduling, progression alerts, and predictive