Use of a human image analyser algorithm to characterise retinal vasculature in cats with hyperthyroidism, diabetes mellitus and chronic kidney disease

To investigate the feasibility of using a computer vision algorithm to analyse images of the feline fundus. Animal studied : Cat Procedures:An algorithm designed to analyse human fundus images was applied to feline images to compare retinal vascular parameters in healthy cats and cats with conditions that could cause hypertension-associated retinopathy including diabetes mellitus, hyperthyroidism and renal failure. Results Cats with renal failure were significantly different to healthy cats in six out of 17 parameters, and diabetic cats were significantly different in two. Conclusions: The human software algorithm (Junction Mark-up) could be used to analyse feline images and significant differences were found between healthy cats and those with renal failure or diabetes mellitus.As diabetic cats rarely develop cataracts this species would be an excellent model for comparative longitudinal retinal imaging studies investigating diabetic retinopathy