Update on the structure-function relationship in glaucoma

The relationship between structural and functional damage in glaucoma, the structure-function relationship, forms the cornerstone of disease assessment, monitoring, and prognosis. We provide an updated synthesis of current knowledge on the structure-function relationship, emphasizing recent advances in imaging, analytical methodologies, and artificial intelligence (AI)-driven modelling. We summarize merging evidence from optical coherence tomography (OCT), OCT angiography, and related imaging technologies, highlighting their integration with functional measures such as standard automated perimetry. We also evaluate classical, mechanistic, statistical, and hybrid models that link retinal microstructure and visual function and discusses the role of AI-based approaches in predicting visual field loss and improving clinical interpretation. The structure-function relationship in glaucoma is predominantly nonlinear for currently used clinical metrics, reflecting biological redundancy, psychophysical variability, and measurement constraints with current tools. Stage-dependent changes reveal strong correlations during moderate disease but increasing divergence in early and advanced stages due to floor effects and functional compensation. Emerging frameworks, such as anatomically compensated mapping, multimodal fusion, and deep learning, enhance accuracy and allow individualized structure-function relationship modelling. Data quality, myopia-related anatomical variability, and cross-device harmonization remain key limitations. Integrating multimodal imaging, AI-based harmonization, and longitudinal data will advance structure-function relationship modelling beyond static correlations toward dynamic, personalized biomarkers of neurovascular and neurofunctional health. These developments promise earlier detection, more precise monitoring, and individualized glaucoma management through a deeper mechanistic understanding of the link between retinal structure and visual function.