Pediatric Artificial Intelligence for Retinopathy Screening in Children With Type 1 Diabetes

The proposed project is a clinical intervention trial testing the feasibility and effectiveness of diabetic retinopathy screening evaluated by artificial intelligence (AI) based software in children with type 1 diabetes (CwD). Another novel method, the confocal microscopy of the retina will be used to assess the early stages of diabetic neuropathy. In parallel, we aim to assess the prevalence of diabetic retinopathy and neuropathy in a well-controlled population of CwD at a tertiary diabetes care center. Each participant will undergo an examination of diabetic retinopathy using the non-mydriatic fundus camera. The resulting photography will be evaluated by AI driven software. The participant will then follow this examinaton with fundus ophtalmoscopy in arteficial mydriasis as a standard method of diabetic retinopathy assessment. Another method, the optic coherence tomography (OCT), which is considered as the most sensitive method for diabetic retinopathy assessment, will be performed after that. The results of these methods will be compared to assess the sensitivity of each. The examination-satisfaction questionnaire will be given to the participants. In subjects over 18 years, a confocal microscopy of the retina examination will be performed to assess the status of the corneal sub-basal nerve plexus and the presence of diabetic neuropathy will be noted. The prevalence of diabetic retinopathy and neuropathy in this group of children with diabetes will be assessed based on the results.

Inclusion Criteria: * Type 1 diabetes diagnosed according to the ADA criteria * Age over 11.0 years or duration of diabetes over 2 years * Signed written informed consent by both the CwD and their parent/caregiver Exclusion Criteria: * Unwillingness to sign a written informed consent by both the CwD and their parent/caregiver