This study evaluates the ability of deep learning to improve the knowledge about structural and vascular retinal changes in diabetic patients undergoing artificial pancreas device, using optical coherence tomography angiography.

hybrid artificial pancreas systems (CSII) are able to adjust the type 1 diabetic patient's glycemic levels with insulin doses appropriate to the patient's blood glucose.

The optical coherence tomography angiography represents a novel and noninvasive diagnostic technique that allows a detailed and quantitative analysis of retinal and choriocapillary vascular features.

The study evaluates the changes in optical coherence tomography angiography features in diabetic patients undergoing artificial pancreas device, elaborating these data with artificial intelligence.

Inclusion Criteria:

• • age older than 18 years

  • diagnosis of type 1 diabetes
  • treatment with artificial pancreas device
  • absence of previous ocular surgery and congenital eye diseases.
  • absence of errors of refraction
  • absence of lens opacities
  • absence of low-quality OCT and OCTA images

Exclusion Criteria:

• age younger than 18 years
• previous ocular surgery and congenital eye diseases
• errors of refraction
• lens opacities
• low-quality OCT and OCTA images

The participans were older than 18 years with diagnosis of type 1 diabetes treated with artificial pancreas device. They did not present other ophthalmological disease