Diabetic Retinopathy
Conditions
Keywords
Diabetic retinopathy. ., Diabetes, Artificial intelligence., Deep learning
Brief summary
Brief Summary: The main objective of EviRed project is to develop and validate a system assisting the ophthalmologist by improving prediction of evolution, and decision making during diabetic retinopathy (DR) follow-up for a patient. It will replace the current classification of diabetic retinopathy (DR) which provides an insufficient prediction precision. It will use modern available fundus imaging devices and artificial intelligence (AI) to properly integrate the amount of data they provide with other medical data of the patient. A cohort of 5000 diabetic patients will be recruited and followed for an average of 2 years in order to collect data to train and validate the new prediction system. An economic study will also be carried out.
Detailed description
A cohort of 5,000 diabetic patients with different stages of DR will be recruited and followed for an average of 2 years. Each year, general data as well as ophthalmological data will be collected. Retinal images and videos of both eyes will be acquired using different imaging modalities including ultrawidefield photography, OCT and OCT angiography. The EviRed cohort will be split in two groups: one group of 1,000 patients (validation cohort) will be randomly selected during the inclusion period by unbalanced draw to be representative of the general diabetic population. Their data will be used for the validation of the algorithms. The data of the remaining 4,000 patients (training cohort) will be used to train the algorithms. The main objective will be the validation of the prognostic tool and evaluate how accurately the algorithm can predict progression to severe retinopathy in the following year. Secondary objectives will be to evaluate how accurately the algorithm can assess DR severity and individual components of severe DR, predict progression to severe retinopathy, visual decrease, occurrence of fluid in the macula area in an eye, DR progression in an eye and for a patient, as well as to compare prediction by algorithm with the one made by ophthalmologists based on the current DR classification and their clinical experience. Secondary outcome measures will be sensitivity, specificity and AUC of the algorithm for detecting DR severity and individual components of severe DR, prediction of DR progression towards a severe form of DR, visual decrease, occurrence of fluid in the macular area in an eye, as well as to predict DR progression in an eye and for a patient. Economic study Healthcare reimbursement data will also be collected and analyzed from Health Insurance. The inclusion and follow-up period of the cohort runs from 12/21/2020 to 12/21/2024, with follow-up durations varying from 1 to 3 years.
Interventions
DIAGNOSTIC_TESTDiagnostic
Des images et des vidéos de la rétine des deux yeux seront acquises en utilisant différentes modalités d'imagerie, notamment la photographie grand champ, l'OCT et l'angiographie OCT. Toutes les images et les données seront collectées grâce à une plateforme commune et centralisées sur un serveur.
Sponsors
Assistance Publique - Hôpitaux de Paris
Study design
Allocation
NON_RANDOMIZED
Intervention model
SINGLE_GROUP
Primary purpose
PREVENTION
Masking
NONE
Intervention model description
Retinal images and videos of both eyes will be acquired using different imaging modalities including widefield photography, OCT and OCT angiography. All images and data will be collected thanks to a common platform and centralized on a server. The EVIRED cohort will be randomly split in two groups: one group of 4000 patients for building algorithms and one group of 1,000 patients to validate them (validation cohort)
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
No
Inclusion criteria
* patient with type 1 or type 2 diabetes or other, * aged 18 years or more, * diabetes duration greater than 10 years for type 1 diabetes, * patient affiliated to social security, * women of childbearing potential who are unwilling or unable to use a method to avoid pregnancy; women who are pregnant or breastfeeding can be included in the study.
Exclusion criteria
* ungradable fundus photography or OCT/OCTA imaging in both eyes, * previous treatment with vitrectomy in both eyes, * participant is unable or unwilling to comply with study procedures (including foreign language speakers who are not assisted by a native language speaker), * Gestational diabetes vulnerable participants (minors, legally detained), * patients under legal protection (guardianship), * prisoners or subjects who are involuntarily incarcerated, * patients with untreated or treated proliferative DR and macular edema involving the center of the macula in both eyes or with panretinal photocoagulation and untreated or treated macular edema involving the center of the macula in both eyes.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Progression to severe retinopathy | month 12 | Progression towards severe diabetic retinopathy form. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Algorithm performance | 3 years | Algorithms will be evaluated by comparing the algorithm performance to automatically assess diabetic retinopathy severity as well as individual components of severe retinopathy against the same grading made by human graders. |
| Comparison of algorithms prediction to human prediction. | 3 years | at each patient's visit of the validation cohort, ophthalmologists will evaluate the risk of DR progression based on the current DR classification and their clinical experience. The risk of DR will be expressed by the clinician as a continuous variable (its estimated probability of progression) or as a semi-quantitative variable. Performance of the human prediction will be compared to the algorithm using sensitivity, specificity and AUC. |
Countries
France
Outcome results
None listed