Effects of myopia-related parameters on retinal vessel calibers measurement results on AI fundus photographs

Background： There has been a paucity of information on the efects of myopia-related parameters factors on the retinal vessel caliber (RVC). Information on this relationship should be useful in studying possible roles of refractive status in the development and progression of ocular circulation. Methods： Subjects were from the Health Medicine Department of Jinling Hospital, Afliated Hospital of Medical School, Nanjing University in China aged 20 to 60 years. All participants underwent fundus photography examination. Parameters assessed were central retinal artery equivalent (CRAE), central retinal vein equivalent (CRVE), Arteriovenous ratio (AVR), cup/disc ratio, optic disc size, cup area,cup/disc ratio, arc spot to optic disc area ratio, and average density of leopard print spots. Normal subjects and eyes were defined according to standardized systemic and ocular examinations. The central retinal artery and vein equivalents (CRAE, CRVE, Arc spot to optic disc area ratio, Average density of leopard print spots and other indicators) were determined on fundus photographs by correcting the magnifcation using the refractive error, corneal curvature, and axial length (AL) of an individual eye and paraxial ray tracing. Results： In a total of 4252 right eyes (4252 normal subjects), aged 37.27±7.29 (standard deviation) years with gradable photographs and planimetric results, CRAE averaged 0.02±0.01μm and CRVE 0.03±0.01 μm. After adjustment for the efects of confounding factors in multivariate analysis, the horizontal cup area ( P ＜0.0001), vertical cup area( P ＜0.0001), vertical cup/disc ratio ( P ＝0.036) and arc spot to optic disc area ratio ( P ＜0.0001) positively with CRAE and the Age ( P ＜0.0001), mean blood pressure ( P ＝0.011 ), body mass index ( P ＝0.003 ), horizontal optic disc size ( P ＝0.005 ), horizontal cup area ( P ＜0.0001 ) and average density of leopard print spots ( P ＜0.0001 ) correlated negatively with CRAE; the horizontal cup area ( P ＝0.001), vertical cup area( P ＜0.0001), vertical cup/disc ratio ( P ＝0.009) and arc spot to optic disc area ratio ( P ＜0.0001) positively with CRVE and the age ( P ＜0.0001), horizontal optic disc size ( P ＝0.021 ), and average density of leopard print spots ( P ＜0.0001 ) correlated negatively with CRVE. There were differences in CRAE, CRVE, AVR based on the quartile of the ratio of arc point to optic disc area (all P ＜0.0001), and there were also differences in CRAE, CRVE, AVR based on the quartile of the average density of leopard point ( P ＝0.035, P ＝0.030, P ＜0.0001). Conclusions： The current study showed signifcant efects of arc spot to optic disc area ratio, average density of leopard print spots on the photographs on the RVC measurement results. Further, it showed a necessity to incorporate the myopia related parameters as co-variables to correctly estimate the efects of various factors on the RVC.