Diabetic Retinopathy Screening through Telemedicine and Artificial Intelligence in the Municipality of Dumont

• Grant number: 24/13690-0
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: October 01, 2024
• End date: September 30, 2025
• Field of knowledge: Health Sciences - Medicine - Surgery
• Principal Investigator: Rodrigo Jorge
• Grantee: Fillipe Davis Monte Mór Martins de Christo
• Host Institution: Hospital das Clínicas da Faculdade de Medicina de Ribeirão Preto da USP (HCMRP). Secretaria da Saúde (São Paulo - Estado). Ribeirão Preto , SP, Brazil

Abstract

Diabetic Retinopathy (DR) is a serious, chronic, and progressive complication of Diabetes Mellitus (DM), leading to significant vision loss and blindness in economically active adults in both developed and developing countries. The gold standard for DR diagnosis involves imaging with traditional fundus cameras, which are costly, especially for the public health service of small municipalities, and require experienced ophthalmologists, posing logistical challenges.Portable fundus cameras offer a lightweight, compact, and robust alternative, maintaining calibration during transport. These devices allow trained paramedics to capture fundus images, which can then be analyzed by ophthalmologists remotely. In the context of telemedicine, this approach provides a low-cost alternative for early DR screening and diagnosis.This study aims to identify new cases of diabetic retinopathy in Dumont-SP by utilizing fundus images captured by a portable fundus camera operated by non-medical professionals. These images will be analyzed using an artificial intelligence (AI) software (SELENA ®) and telemedicine, with results compared to the degree of DR assessed by an ophthalmologist through Indirect Binocular Ophthalmoscopy (IBO). Additionally, the study will evaluate the implementation costs of the screening program and the savings generated by local patient evaluations. The expectation is that this system will facilitate, expedite, and democratize early DR screening and diagnosis within this population.Diabetic patients will be invited to the Basic Health Unit of Dumont, where a nursing technician, previously trained by the HCFMRP-USP team, will administer a brief questionnaire and perform the fundus photography exam. The photographic documentation will be sent via the internet using a sequence of randomly generated numbers, ensuring patient anonymity and compliance with the General Data Protection Law. The images will be analyzed by a team of retina specialists from HCFMRP-USP with at least three years of residency experience, who will provide a report on the presence or absence of DR, as well as its severity according to the International Clinical Classification for Diabetic Retinopathy. Patients will then be referred through the Regional Health Department (DRS) for specialized evaluation and treatment. The project will gather information on:1. Accuracy of telemedicine using fundus photos taken by non-medical professionals for DR screening and classification.2. Accuracy of AI software analysis for DR screening using fundus photos taken by non-medical professionals.3. New DR diagnoses facilitated by remote medical evaluations (telemedicine) in a municipality with existing ophthalmological evaluation services.4. New DR diagnoses facilitated by AI screening (SELENA® software) in a municipality with existing ophthalmological evaluation services.5. Difference in the occurrence of new diagnoses made by telemedicine (remote evaluation of fundus photographs by doctors) versus AI evaluation (SELENA® software).6. Cost of DR diagnosis by on-site medical professionals (using devices such as indirect binocular ophthalmoscopes and medical labor).7. Cost of DR diagnosis via telemedicine (cost of portable fundus photography exams and estimated report values).8. Cost of DR diagnosis via AI (cost of portable fundus photography exams and software usage).9. Reduction in costs related to on-site examinations compared to those performed via telemedicine and AI.