Smartphone-based fundus imaging and computer vision analysis for monitoring retinopathy of prematurity in neonatal rats with and without lutein treatment

Abstract

Retinopathy of prematurity (ROP), caused by oxidative stress leading to abnormal retinal vessel growth, is often studied in rodent models. However, current methods rely on in vitro analysis, preventing the monitoring of disease progression. Fundus imaging is commonly used in clinical ophthalmology and could provide a method for in vivo imaging in rodents with ROP. Recently, a smartphone and condensing lens have been shown to be effective in humans and mice. The aim of this study was to develop a smartphone-based fundus camera system that can be used to monitor in vivo changes caused by ROP and determine its effectiveness upon introducing an intervention of lutein. KRN 633 was used to induce ROP and fundus imaging was conducted using a condensing lens and a smartphone. In vitro analysis of the retina was conducted for comparison using immunohistochemistry. The tortuosity was analyzed from fundus images manually using ImageJ and by a newly developed computer vision-based method. ROP induced tortuous arteries, which was improved by the administration of lutein in the fundus and microscopy images of the retina. The two imaging techniques showed a strong positive correlation. Similarly, the computer vision-based tortuosity analysis displayed a strong linear relationship with ImageJ analysis. This smartphone-based fundus camera system offers a cheap and accessible method for in vivo imaging to supplement traditional histological analysis. With future development, the image processing technique described can be enhanced to include other measures of visual function.