• Diabetic retinopathy has no symptoms in its early stages, and by the time it is diagnosed, it is often so advanced that vision loss is inevitable.
• The pilot, successfully implemented in the Mexican state of Jalisco, lasted eight months and took place in three first-level health clinics in the Metropolitan Area of Guadalajara, the state’s capital.
• The combination of emerging technologies and medical procedures ensured that patients received timely diagnoses and effective treatment.

Monterrey, Nuevo León, México.- Retina-IA, a pilot screening project designed for the early detection of diabetic retinopathy that combines artificial intelligence (AI) with telemedicine, was successfully implemented in first-level clinics in Guadalajara, the capital of the Mexican state of Jalisco. This initiative helped reduce complications of this condition that can lead to blindness.

Now part of the state Cardiometabolic Program, Retina-IA was promoted by fAIr LAC Jalisco, an initiative in which Tecnologico de Monterrey, the Inter-American Development Bank (IDB), the Government of Jalisco and C Minds collaborate, with the purpose of using AI in a responsible and ethical manner for social good.

“Diabetic retinopathy is a significant concern across Mexico due to its chronic and progressive nature. The disease has no symptoms in its early stages, and when diagnosed, is often so advanced that vision loss is inevitable. Although it cannot be cured, its progression can be controlled if detected in time,” says Enrique Cortes-Rello, leader of the Artificial Intelligence Initiative at the School of Engineering and Sciences of Tecnologico de Monterrey, the number one university in engineering and technology in Mexico, according to the QS World Ranking 2024.

Early detection of this condition is not typically available on a mass scale in the country due to various factors, the most relevant being the inability to perform a diagnosis in the first-level clinics where diabetic patients undergo routine checkups. When vision problems do appear and patients are referred to a specialized third-level unit where retinal ophthalmologists treat them, diabetic retinopathy is often in its advanced stages.

The times it takes for a diagnosed patient to move from first-level care to second-level care can be at least three months, while referral to the third-level care can take up to six months. These delays in care are the reason why more than 40% of cases are not treated on time.

Finding the focus of attention
The degenerative nature of diabetic retinopathy places it among public health priorities both in the public sector and in society in general. The decision to focus on this disease for the Retina-IA project was made through consensus between participating organizations, the Government of Jalisco and state health authorities.

Cortes-Rello highlights the discussion with specialists from the state health system to focus on type II diabetes mellitus and on diabetic retinopathy. “With the help of doctors, we refined the project’s focus. The spectrum of diseases related to this condition is very broad, but we determined that early detection of this degenerative eye condition would benefit many people,” he says.

“This project was characterized by its multidisciplinary nature, which allowed us to have a comprehensive perspective of the disorder, something that is key from an engineering standpoint,” adds the researcher.

The magnitude of the diabetes problem in Mexico is alarming. According to the National Health and Nutrition Survey 2022, 18.3% of the Mexican population suffers from diabetes, amounting to approximately 14.6 million people. This disease is the second leading cause of death and the leading cause of disability in the country. Additionally, statistics from the International Diabetes Federation indicate that by 2030, at least 17.2 million people in Mexico will have diabetes, a figure equivalent to twice the total population of the state of Jalisco. This highlights the relevance of projects like Retina IA.

The screening process
Retina-IA starts at the Diabetic Retinopathy Assessment and Care Unit (UVARD) in first-level health centers. Images of the back of the patient's eye were captured and stored on a telemedicine platform. The Medical College of Ophthalmologists of Jalisco was responsible for reviewing these images and issuing a diagnosis, supported by a deep learning AI algorithm.

If diabetic retinopathy was at a level of R3 and R4, which responds to a higher degree of specialized care, the patient was referred to a second or third level clinic. Otherwise, the patient continued regular care at their local clinic.

The clinics hosting the project
The Retina-IA pilot project was implemented in three clinics located in marginalized areas of Guadalajara between August 2021 and March 2022, serving 1,053 people, of whom 54 were diagnosed with referable diabetic retinopathy. Thanks to the collaboration with the Medical College of Ophthalmologists of Jalisco, diagnostic delivery times were reduced from 14 days to a maximum of two days.

To prepare the models used in the Retina-IA project, exhaustive research was conducted on diabetic retinopathy-focused models from other countries. It was decided to use models approved by the United States Food and Drug Administration (FDA). “We considered that the FDA-approved model was ideal, and we used it in clinics,” adds Cortes-Rello.

The combination of cutting-edge technologies and better-defined medical procedures ensured that patients received fast and accurate diagnoses, significantly reducing waiting times and increasing the chances of effective treatment.

Benefits and impact
The cost-benefit analysis of the Retina-IA project, conducted in collaboration with the Inter-American Development Bank, demonstrated that this approach is not only feasible, but also highly cost-effective. According to the economic evaluation, for each dollar invested in the project (technology, infrastructure, and people), between 3 and 10 dollars are generated in benefits for society (patients, their families, and the health system).

The project presents a solid proposal for scalability and replicability in other regions of Mexico and other emerging countries.

The main economic benefits include improvements in labor productivity and significant savings for the public health system. By preventing blindness and other severe complications of diabetic retinopathy, patients can maintain their work capacity and actively contribute to the economy, while the health system reduces costs associated with advanced treatments and long-term care.

Retina-IA also has a positive impact on patients' quality of life. Early detection and appropriate treatment not only prevent blindness but also improve the overall well-being of patients and their families. This preventive approach contributes to a healthier and more productive society, with fewer people suffering from visual disabilities that could have been prevented with an early diagnosis.

Currently, several states in Mexico have expressed interest in replicating the Retina-IA pilot project, says Enrique Cortes-Rello. The use of AI in Retina-AI not only improves the accuracy and speed of diagnoses but it also sets a new standard for the application of advanced technologies in public health.

###

About Tecnologico de Monterrey

Tecnologico de Monterrey (http://www.tec.mx) is a private, non-profit university recognized for its academic excellence, educational innovation, and global vision. It was founded in 1943 and currently has a presence in 33 municipalities across 20 states in Mexico, with an enrollment of 62,000 students at the undergraduate and graduate levels, as well as more than 27,000 high school students. Accredited by SACSCOC since 1950, it ranks 184th in the QS World University Rankings 2024 and 4th in Latin America according to the THE Latin America University Rankings 2023. It also stands out in global employability and entrepreneurship programs, being part of international networks such as APRU, U21, and WUN.