Abstract:  Currently, the medical diagnosis and treatment paradigm has transitioned from “disease-centered” to “health-centered”, and early screening and disease prevention have emerged as crucial strategies for tackling the global burden of chronic non-communicable diseases. As the sole anatomical structure in the human body enabling non-invasive and direct observation of the central nervous system and the overall circulatory status, the retina's latent value in the assessment of overall health is increasingly prominent. This article systematically reviews the research progress in quantifying retinal image features using artificial intelligence (AI) technology and elaborates on its role in the early warning, risk stratification, and auxiliary diagnosis of cardiovascular diseases, neurodegenerative diseases, metabolic diseases, kidney diseases, and hematological diseases. Based on this, this article innovatively proposes the concept of "eye routine", which is defined as a systematic health assessment system that integrates AI-driven multi-modal retinal image quantification analysis as the core and standardizes the integration of basic examinations such as visual acuity and intraocular pressure. Meanwhile, a "routine eye examination" system architecture has been established. Its implementation paths and application values in scenarios such as population screening, multi-disciplinary clinical diagnosis and treatment, and health management have been analyzed. The challenges it faces and its future development directions have also been discussed. The aim is to promote the gradual development of the "routine eye examination" into a standardized basic examination project similar to the "routine blood test", providing innovative solutions for strengthening the foundation of national health and achieving precise preventive medicine.

Key words: Eye and systemic diseases, Eye routine, Artificial intelligence, Quantitative analysis, Health management