More information
Description
Diabetic retinopathy is the leading cause of blindness and reduced vision in the developed world. Many countries in Europe, including Estonia, Latvia and Lithuania, have no national screening programs for diabetic retinopathy, while Norway just recently implemented such a program, which is not yet personalized. Currently, the monitoring interval for patients with diabetic retinopathy set according to the international guidelines is “rigid”. It may, however, vary from patient to patient, as it depends on various risk factors. The cost of regular annual screening of diabetic retinopathy is enormous, while the rising number of patients with diabetes mellitus surpasses the capacity of ophthalmologists. The aim of the project is to implement a new personalized diabetic retinopathy screening and monitoring program using artificial intelligence (AI) for future applications in the integrated care of patients with diabetes.
Objectives of this project include: 1. Evaluation of the current diabetic retinopathy status and risk factors in patients with diabetes in partner countries; 2. Improvement and implementation of a personalized risk-stratification algorithm in the daily diabetes eye screening; 3. Utilization of fast data extraction methods from medical electronic records and AI to detect novel risk factors for diabetic retinopathy; 4. Evaluation of the cost-efficacy of running an AI-based diabetic retinopathy monitoring program in the partner countries; 5. Initiation of a sustainable screening and monitoring programs for diabetic retinopathy in the partner countries as part of their ongoing or future eHealth initiatives.
The project will be promoted by the University of Latvia (Riga, Latvia). Other participants will be Tartu University (Tartu, Estonia), University of Oslo (Oslo, Norway) and Lithuanian University of Health Sciences (Kaunas, Lithuania).
Summary of project results
Modern medicine is unthinkable without science and technology. However, sometimes the interaction of science, technology and medicine leads to challenging problems. The international project PerDiRe, in which Latvia was a leading partner, was dedicated to the early recognition of diabetic retinopathy and the development of new approaches to patient monitoring. This project, in which 500 patients from Latvia, Lithuania, Estonia and Norway participated, was dedicated not only to improving eye health, but also to the application of new technologies and data analysis in medicine.
Diabetic retinopathy (or diabetic eye disease) is one of the leading causes of vision loss in developed countries. The aim of the project was to improve the diagnosis and prognosis of diabetic retinopathy by applying a new artificial intelligence-based screening and monitoring program for diabetic retinopathy, which introduced machine learning and image analysis.
The project’s activities included monitoring patients for one year, where clinical factors were monitored and analysed with the help of artificial intelligence, as well as an innovative approach in the analysis of eye images. The cost-effectiveness of this solution was evaluated. When observing patients, in addition to the standard ophthalmologist examination, as part of diabetic retinopathy screening, we also collected data on diabetes history, comorbidities, and some clinical factors (e.g., blood pressure). As one of the new and promising biomarkers for determining the risk of progression of diabetic retinopathy, we measured sugar metabolism products in the skin of the patients. In addition, genetic risk factors for diabetic retinopathy were also searched within the project. Currently, the analysis of project data is still ongoing. However, we can already report the first results. First, experience at Oslo University Hospital shows that AI-based fundus image analysis for monitoring diabetic retinopathy is cost-effective. Second, we managed to develop a new approach for early automated diagnosis of diabetic retinopathy using image segmentation method. Second, we managed to develop a new approach for early automated diagnosis of diabetic retinopathy using image segmentation method. Thirdly, the results of the project clearly indicate that screening for diabetic retinopathy should not be limited to an eye doctor’s examination. To effectively observe the patient and reduce the risk of vision loss, the cooperation of several specialists (ophthalmologist, endocrinologist, general practitioner) and the observation and correction of clinical factors in close connection with the treatment of eye changes are required. For example, the project data demonstrate that many diabetic patients have a blood pressure that is outside the recommended norms, and they are not prescribed appropriate treatment or are not explained the importance of taking medication regularly to prevent the development of diabetes complications. Fourth, the new biomarkers studied in the project (detection of end products of sugar metabolism in the skin and genetic factors) may improve the determination of the severity and risk of progression of diabetic retinopathy.
Results of this project were disseminated via 3 scientific publications and served as a basis for 3 joint applications. During the project, the training of young researchers, including PhD students, took place. The bilateral- and multilateral cooperation between Latvia, Estonia, Lithuania, and Norway was strengthened. The novel approach piloted for diabetic retinopathy screening and monitoring, which integrates advanced AI methods and comprehensive clinical data, enhances early detection and personalized treatment. This can lead to better health outcomes for diabetic patients, potentially reducing the incidence of severe diabetic retinopathy and preventing blindness, which in turn can lower healthcare costs and improve quality of life. By demonstrating the cost-efficiency of automated diabetic retinopathy grading, the project highlights the potential for reducing healthcare expenditure associated with manual screening processes. The PerDiRe project conducted public awareness campaigns through various media channels to educate the broader community about diabetic retinopathy, its risks, and the benefits of the novel screening methods developed by the project.
Indicators achieved in the project: 3 joint scientific publications have been prepared, 3 joint project applications for further funding were submitted, 18 researchers were supported.
Summary of bilateral results
The project was a positive test of cooperation and innovation. Cooperation was strengthened, knowledge exchange took place intensively, and all project partners gained new skills and experience. Quality clinical care was provided to diabetic patients and the results of the project will later contribute to further research in the field of diabetic retinopathy. We are glad that during the project rallies, we got to know the screening approaches of diabetic retinopathy in the institutions of the project participants and there was an exchange of experiences.In the future, the scientific direction of the project will be further developed. It is planned to apply to Eurostars and ICPerMed project calls to continue the work on the analysis of eye images with artificial intelligence and the study of new biomarkers of diabetic retinopathy. As part of these new project applications, it is planned to expand the PerDiRe consortium and continue the research and analysis of the data obtained. This experience is not only medically significant, but also provides an opportunity to develop new collaboration and research opportunities.