Abstract
Purpose :
Testing, on a large sample of fundus images, the lens age estimation using the color of the central vessels of the optic nerve as a pattern. This reports the increased absorption of the crystalline lens to short wavelength radiation.
Methods :
Fundus images were obtained from 426.764 eyes with a Topcon TRC-NW400 fundus camera (Topcon, Japan). Image quality, laterality of the eye, nerve position and nerve and vessel segmentation (Fig. 1) were automatically assessed by using the Deep Learning training for the Laguna ONhE program which is mainly used for glaucoma (1-4). A multiple regression equation was calculated based on the RGB components of the vessels so as to deduce the biological age of normal subjects.
Results :
The biological age of the subjects was estimated with a standard error (SE) of 6.91 years (R=0.901, P<0.0001) in 80% of the cases (central band of the figure 2) and with a standard error (SE) of 8.31 years (R=0.856, P<0.0001) in 5% of the cases, in each of the two intermediate bands of the image. The remaining cases represented in the upper left zone correspond to 5% of prematurely aged crystalline lenses and the 5% represented in the lower right zone to young-for-age crystalline lenses.
Conclusions :
The ageing of the crystalline lens can be estimated by observing the color changes of the central retinal vessels as they pass through the optic disc. These data confirm previous normal results obtained with the DEC-200 fundus camera (MiiS, Taiwan)(5). Prematurely aged crystalline lenses may correspond, for example, to diabetic patients, as we have described previously (6), and too young crystalline lenses to cataract surgery patients who have had a pseudophaco lens implanted, as we have found in some cases.
References:
Gonzalez de la Rosa M et al. IOVS 2013;54:482-489.
Gonzalez-Hernandez M et al. J Ophthalmol 2017;2340236.
Gonzalez-Hernandez D et al. J Clin Exp Opthamol 2018;9:5 DOI: 10.4172/2155-9570.1000760.
Mendez-Hernandez C et al. Br J Ophthalmol 2020;doi:10.1136/bjophthalmol-2020-316455.
This abstract was presented at the 2023 ARVO Imaging in the Eye Conference, held in New Orleans, LA, April 21-22, 2023.