Evaluation of NeoRetina Artificial Intelligence Algorithm for the Screening of Diabetic Retinopathy at the CHUM

Overview

This prospective study aims to validate if NeoRetina, an artificial intelligence algorithm developped by DIAGNOS Inc. and trained to automatically detect the presence of diabetic retinopathy (DR) by the analysis of macula centered eye fundus photographies, can detect this disease and grade its severity.

Full Title of Study: “The Use of Artificial Intelligence in the Early Detection and the Follow-Up of Diabetic Retinopathy of Diabetic Patients Followed at the CHUM: Evaluation of NeoRetina Automated Algorithm (DIAGNOS Inc.)”

Study Type

  • Study Type: Interventional
  • Study Design
    • Allocation: N/A
    • Intervention Model: Single Group Assignment
    • Primary Purpose: Diagnostic
    • Masking: None (Open Label)
  • Study Primary Completion Date: January 2024

Detailed Description

More than 880 000 Quebecers (more than 10% of the population) suffer from diabetes, which is the main cause of blindness in diabetic adults under 65 years of age, and around 40% of people with diabetes suffer from diabetic retinopathy (DR). The early detection of DR and a regular follow-up is thus crucial to prevent the progression of this disease. However, the public health care system in Quebec does not actually have the capacity to allow all people with diabetes to see an ophthalmologist within a short delay. Artificial intelligence might help in screening DR and in refering to eye doctors only patients who suffer from this eye disease. The investigators of this study hypothesize that artificial intelligence (AI) is a useful technology for the screening of diabetic retinopathy (DR) that can detect the absence or the presence of DR with an efficiency and an accuracy similar to that of an ophthalmological evaluation. The goal of this study is to compare the screening results of DR obtained with NeoRetina pure artificial intelligence algorithm (automated analysis of color photos of the retina) with the results of a routine ophthalmological evaluation done in a clinical context at the Centre hospitalier de l'Université de Montréal (CHUM). The main objective of this study is to determine if artificial intelligence (AI) could be a useful technology for the early detection and the follow-up of diabetic retinopathy (DR). The first specific objective is to determine the efficiency and the accuracy of NeoRetina (DIAGNOS Inc.) automated algorithm for the screening and the grading of the severity of diabetic retinopathy (DR) by the analysis of eye fundus images from diabetic patients compared to that of an eye examination done by an ophthalmologist in a clinical context. The second specific objective is to evaluate if NeoRetina can determine, with efficiency and accuracy, the absence of diabetic retinopathy (DR), the presence of diabetic retinopathy (DR) and the severity of the disease. Recruited diabetic participants will be screened for DR by AI with NeoRetina. Participants will also have a full eye examination (blind assessment) with an ophthalmologist of the CHUM in order to determine if they suffer from this eye complication of diabetes. The results of the screening done by AI with NeoRetina will be compared to those of the ocular evaluation done by an ophthalmologist. Ophthalmologists from the CHUM will also revise the retinal images acquired by DIAGNOS (blind assessment) in order to determine if DR is present and will manually grade the severity of the disease.

Interventions

  • Diagnostic Test: Screening of DR and DME with artificial intelligence using NeoRetina
    • Macula-centered eye color fundus photos will be acquired by DIAGNOS team using a non-mydriatic digital camera (without pupil dilation). After a numerical treatment, retinal images will be analyzed by NeoRetina artificial intelligence (AI) algorithm in order to find eye lesions characteristics of diabetic retinopathy (DR) and diabetic macular edema (DME). The severity of DR and DME will be graded by NeoRetina according to the ”Early Treatment Diabetic Retinopathy Study” (ETDRS) international classification standards.
  • Diagnostic Test: Routine ophthalmological evaluation of DR and DME
    • Standard of care eye examination (blind assessment) will be performed by an ophthalmologist of the CHUM in order to find lesions characteristics of diabetic retinopathy (DR) and diabetic macular edema (DME). The severity of DR and DME will be graded by the doctor according to the ”Early Treatment Diabetic Retinopathy Study” (ETDRS) international classification standards.
  • Diagnostic Test: Manual grading of DR and DME by CHUM ophthalmologists based on retinal photographies acquired by Diagnos
    • Ophthalmologists of the CHUM will revise the macula-centered eye color photos acquired by DIAGNOS in order to find lesions characteristics of diabetic retinopathy (DR) and diabetic macular edema (DME). The severity of DR and DME will be graded (blind assessment) according to the ”Early Treatment Diabetic Retinopathy Study” (ETDRS) international classification standards.

Arms, Groups and Cohorts

  • Experimental: Diabetic Retinopathy (DR)
    • Screening of DR with artificial intelligence (NeoRetina algorithm) and diagnostic evaluation with a standard of care ophthalmological examination.

Clinical Trial Outcome Measures

Primary Measures

  • Artificial Intelligence – Absence or Presence of Diabetic Retinopathy (DR)
    • Time Frame: Baseline
    • Analysis of retinal images by artificial intelligence (NeoRetina) to determine the absence or the presence of diabetic retinopathy (DR) R0 : No DR R+ : Presence of DR
  • Eye Examination – Absence or Presence of Diabetic Retinopathy (DR)
    • Time Frame: Baseline
    • Eye examination done by an ophthalmologist to determine the absence or the presence of diabetic retinopathy (DR) (blind assessment) R0 : No DR R+ : Presence of DR
  • Manual Analysis of Retinal Images – Absence or Presence of Diabetic Retinopathy (DR)
    • Time Frame: Baseline
    • Manual analysis of retinal images acquired by Diagnos by an ophthalmologist of the CHUM to determine the absence or the presence of diabetic retinopathy (DR) (blind assessment) R0 : No DR R+ : Presence of DR
  • Artificial Intelligence – Severity of Diabetic Retinopathy (DR)
    • Time Frame: Baseline
    • Analysis of retinal images by artificial intelligence (NeoRetina) to grade the severity of diabetic retinopathy (DR) R1 – Mild NPDR: Mild Nonproliferative Diabetic Retinopathy R2 – Moderate NPDR: Moderate Nonproliferative Diabetic Retinopathy R3 – Severe NPDR : Severe Nonproliferative Diabetic Retinopathy R4 – PDR : Proliferative Diabetic Retinopathy
  • Eye Examination – Severity of Diabetic Retinopathy (DR)
    • Time Frame: Baseline
    • Eye examination done by an ophthalmologist to grade the severity of diabetic retinopathy (DR) (blind assessment) R1 – Mild NPDR: Mild Nonproliferative Diabetic Retinopathy R2 – Moderate NPDR: Moderate Nonproliferative Diabetic Retinopathy R3 – Severe NPDR : Severe Nonproliferative Diabetic Retinopathy R4 – PDR : Proliferative Diabetic Retinopathy
  • Manual Analysis of Retinal Images – Severity of Diabetic Retinopathy (DR)
    • Time Frame: Baseline
    • Manual revision of retinal images acquired by Diagnos by an ophthalmologist of the CHUM to grade the severity of diabetic retinopathy (DR) (blind assessment) R1 – Mild NPDR: Mild Nonproliferative Diabetic Retinopathy R2 – Moderate NPDR: Moderate Nonproliferative Diabetic Retinopathy R3 – Severe NPDR : Severe Nonproliferative Diabetic Retinopathy R4 – PDR : Proliferative Diabetic Retinopathy
  • Artificial Intelligence – Absence or Presence of Diabetic Macular Edema (DME)
    • Time Frame: Baseline
    • Analysis of retinal images by artificial intelligence (NeoRetina) to determine the absence or the presence of diabetic macular edema (DME) M0 : No DME M+ : Presence of DME
  • Eye Examination – Absence or Presence of Diabetic Macular Edema (DME)
    • Time Frame: Baseline
    • Eye examination done by an ophthalmologist to determine the absence or the presence of diabetic macular edema (DME) (blind assessment) M0 : No DME M+ : Presence of DME
  • Manual Analysis of Retinal Images – Absence or Presence of Diabetic Macular Edema (DME)
    • Time Frame: Baseline
    • Manual analysis of retinal images acquired by Diagnos by an ophthalmologist of the CHUM to determine the absence or the presence of diabetic macular edema (DME) (blind assessment) M0 : No DME M+ : Presence of DME
  • Artificial Intelligence – Severity of Diabetic Macular Edema (DME)
    • Time Frame: Baseline
    • Analysis of retinal images by artificial intelligence (NeoRetina) to grade the severity of diabetic macular edema (DME) M1 : Non Central DME M2 : Central DME
  • Eye Examination – Severity of Diabetic Macular Edema (DME)
    • Time Frame: Baseline
    • Eye examination done by an ophthalmologist to grade the severity of diabetic macular edema (DME) (blind assessment) M1 : Non Central DME M2 : Central DME
  • Manual Analysis of Retinal Images – Severity of Diabetic Macular Edema (DME)
    • Time Frame: Baseline
    • Manual analysis of retinal images acquired by Diagnos by an ophthalmologist of the CHUM to grade the severity of diabetic macular edema (DME) (blind assessment) M1 : Non Central DME M2 : Central DME

Secondary Measures

  • Performance of NeoRetina Algorithm – Diabetic Retinopathy (DR)
    • Time Frame: 3 years
    • The performance of NeoRetina algorithm for the detection and the grading of diabetic retinopathy (DR) will be evaluated. The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) and area under the receiver operating characteristic curve (AUC, 95% CI) will be calculated. The levels of agreement will be determined by kappa analyses.
  • Performance of NeoRetina Algorithm – Diabetic Macular Edema (DME)
    • Time Frame: 3 years
    • The performance of NeoRetina algorithm for the detection and the grading of diabetic macular edema (DME) will be evaluated. The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) and area under the receiver operating characteristic curve (AUC, 95% CI) will be calculated. The levels of agreement will be determined by kappa analyses.

Participating in This Clinical Trial

Inclusion Criteria

1. Patients of 18 years old and older; 2. Ability to provide informed consent; 3. Diagnostic for diabetes : 3a) Type 1 diabetes of a lest 5 years of evolution; or 3b) Type 2 diabetes; 4. Diabetic patient followed and refered by a physician of the Centre hospitalier de l'Université de Montréal (CHUM) : 4a) followed by an endocrinologist of the CHUM; or 4b) hospitalized at the CHUM; or 4c) on the waiting list of the Ophthalmology Clinic of the CHUM for the evaluation of DR. Exclusion Criteria:

1. Patients less than 18 years old; 2. Inability to provide informed consent; 3. Patient who already had a treatment (surgery, laser, injection, etc.) for any retinal condition : Age-related macular degeneration (AMD), retinal vascular occlusion (RVO); etc.

Gender Eligibility: All

Minimum Age: 18 Years

Maximum Age: N/A

Are Healthy Volunteers Accepted: No

Investigator Details

  • Lead Sponsor
    • Centre hospitalier de l’Université de Montréal (CHUM)
  • Collaborator
    • DIAGNOS Inc.
  • Provider of Information About this Clinical Study
    • Sponsor
  • Overall Official(s)
    • Salim Lahoud, MD, Principal Investigator, Centre hospitalier de l’Université de Montréal (CHUM)
  • Overall Contact(s)
    • Marie-Catherine Tessier, MSc, 514-890-8000, marie-catherine.tessier.chum@ssss.gouv.qc.ca

References

Unité d'évaluation des technologies et des modes d'intervention en santé (UETMIS). Centre hospitalier de l'Université de Montréal. Projet pilote : application de l'intelligence artificielle en ophtalmologie. Revue de la littérature et étude de terrain, phase I. Préparée par Imane Hammana et Alfons Pomp. Février 2020.

Shaban M, Ogur Z, Mahmoud A, Switala A, Shalaby A, Abu Khalifeh H, Ghazal M, Fraiwan L, Giridharan G, Sandhu H, El-Baz AS. A convolutional neural network for the screening and staging of diabetic retinopathy. PLoS One. 2020 Jun 22;15(6):e0233514. doi: 10.1371/journal.pone.0233514. eCollection 2020.

Clinical trials entries are delivered from the US National Institutes of Health and are not reviewed separately by this site. Please see the identifier information above for retrieving further details from the government database.

At TrialBulletin.com, we keep tabs on over 200,000 clinical trials in the US and abroad, using medical data supplied directly by the US National Institutes of Health. Please see the About and Contact page for details.