Ophthalmic imaging and artificial intelligence in early diagnosis of Alzheimer disease

Alzheimer disease represents a major medical and social challenge owning to the complexity of its diagnosis, especially at the preclinical stage. This underscores the need for noninvasive and cost-effective screening methods. This review summarizes current evidence on using ophthalmic imaging modalities and artificial intelligence technologies for early detection of Alzheimer disease. The diagnostic potential of retinal biomarkers identified with optical coherence tomography and optical coherence tomography angiography is detailed, including thinning of the peripapillary retinal nerve fiber layer, reduced macular and choroidal thickness, decreased capillary perfusion density, and deposition of pathological proteins. These changes correlate with cerebral condition and can be detected at preclinical stages of Alzheimer disease. The article describes the roles of machine-learning algorithms and neural networks in automated image analysis, demonstrating their ability to identify complex imaging patterns and substantially improve diagnostic accuracy (AUC > 0.9). It also addresses methodological limitations and implementation challenges—including variability of results, insufficient specificity, and the black-box nature of artificial intelligence. We also highlight the high potential of multimodal approaches that combine retinal imaging with MRI and positron emission tomography. The evidence presented supports the feasibility of developing standardized protocols for the use of retinal biomarkers and artificial intelligence technologies as tools for large-scale screening of at-risk populations, enabling their integration into clinical practice for earlier initiation of therapy.