The use of autonomous artificial intelligence for the detection of early signs of retinal damage and association with long-term fluctuations in glucose levels in children and youth with type 1 diabetes

Type 1 diabetes (T1D) is one of the most common chronic conditions of children and adults all over the world. Its incidence is increasing worldwide with an estimated overall annual rate of up to 3.4%, the incidence worldwide remains the highest in the youngest children. The most common long‐term complication of diabetes is diabetic retinopathy (DR), the leading causes of blindness in developed countries. It is estimated that approximately 0.9 million individuals aged 50 years and older have blindness and additional 2.9 million has moderate or serious visual impairment due to DR. Screening for early signs of retinal damage is therefore imperative for early intervention and possible loss of sight prevention. Current international guidelines recommend that screening in individuals with T1D is initiated within 3–5 years of diagnosis, with yearly (every 2 years) follow‐up exams thereafter. However, data show that only 35–72% of youth with diabetes undergo recommended ophthalmic exams in accordance with clinical practice guidelines. To improve accessibility and screening adherence, digital fundus photography using nonmydriatic cameras has been implemented in adult and pediatric clinics, further improved with fully autonomous artificial intelligence (AI)–based systems for detection of DR and diabetic macular oedema that has already demonstrated reduced costs and better adherence of individuals with diabetes to the programme. The main objectives of our study are: (i) to determine the incidence of diabetic retinopathy in children and youth with type 1 diabetes in Slovenia using autonomous AI analysis of fundus and OCT / OCTA (optic coherence tomography (angiography) digital images. (ii) We will cooperate in the development of autonomous AI for the analysis of OCT / OCT‐A retinal imaging and the introduction of the method in youth with T1D (iii) To evaluate the association between long‐term fluctuations in glucose levels and the prevalence of early signs of retinal damage; including a one‐year follow‐up with detailed information on glucose control (iv) To evaluate these observations compared to negative control group (non‐diabetic firstdegree relatives of persons with T1D). (v) To evaluate additional risk factors for the development of early retinal damage (genetic factors, indicators of mild inflammation, plasma lipids, copeptin, kidney function). (vi) Additionally, we will analyse data from already isolated DNA material to profile genotype–phenotype of affected individuals with early signs of retinal changes; therefore, we will build on information obtained with ARRS research project J7-1820, enabling patient-centered or personalised clinical care. Participants We aim to invite all youth aged between 10 and 21years with T1D for at least 3 years (case) in Slovenia recruited through pediatric diabetes outpatient clinic. We anticipate that around 350 individuals are eligible and with a predicted 90% response rate we aim to include at least 300 participants. Additionally, we aim to include approximately 100 healthy first‐degree relatives of individuals with T1D aged between 10 and 21 years through pediatric and adult outpatient clinic (negative control) at the participating clinical site. For the negative control group, we will invite participants, who are already or are going to be included in other ongoing clinical trials. They must not have two or more positive diabetes specific auto‐antibodies (anti GAD65, anti IA‐2, anti ZnT8, their HbA1c value must be below 5,7% The ophthalmological examination will include retinal examination will digital fundus imaging and OCT/OCT‐A using TOPCON 3D OCT‐1 Maestro2 with integrated nonmydriatic Retinal Camera TOPCON TRC‐NW400 and SS‐OCTA. The obtained digital image will be analysed with autonomous artificial intelligence algorithm. All participants in the case group will be invited to perform a follow‐up visit with all listed procedures after approximately 12 months.