Artificial Intelligence-assisted Diagnosis in Ophthalmology

NCT07497815 | Not Yet Recruiting

• 1 other identifier: CEC-HCB-E023-2026
• Study Type: observational
• Target: 15,000
• Locations: 1 country
• Sites: 2

Brief Summary

This is a retrospective, multicenter, observational study designed to develop and validate an artificial intelligence (AI) system capable of detecting and classifying major ophthalmic diseases (glaucoma, cataract, diabetic retinopathy, and other retinal pathologies) in the Costa Rican population. The study will use approximately 15,000 existing medical images from digital archives of two ophthalmic centers in Costa Rica, without active participant recruitment or capture of new images. The primary motivation is that AI systems developed in other countries (primarily Asian, European, or North American populations) do not necessarily perform with the same accuracy when applied to Latin American populations. This study seeks to establish a precedent for the importance of locally validating any medical AI technology before clinical implementation.

Conditions

Macular Degeneration; Diabetic Retinopathy (DR); Glaucoma; Keratoconus; Cataract

Keywords

artificial inteligence; ophthalmology; imaging; bias

Outcome Measures

Primary Outcomes (2)

• Area Under ROC Curve (AUC)

  Area under the receiver operating characteristic curve (AUC-ROC) for each of the pathologies detection by the AI system, evaluated on the independent validation set of 3,000 images. AUC-ROC is a comprehensive measure of diagnostic performance across all possible decision thresholds. Values range from 0.5 (random guessing) to 1.0 (perfect classification). Success criterion: AUC ≥ 0.90.

  At study completion (Month 24)

• Specificity

  Specificity (true negative rate) of the AI system for glaucoma detection, defined as the proportion of non-glaucoma cases correctly identified as negative. Success criterion: Specificity ≥ 85%.

  At study completion (Month 24)

Secondary Outcomes (1)

• Sensitivity

  At study completion (Month 24)

Study Arms (1)

Images of patients over 18 years old

This is a diagnostic validation study without intervention. All images are analyzed using the same methodology. There are no comparison groups, treatment arms, or cohorts. The study evaluates AI system performance against expert ophthalmologist diagnoses (ground truth).

Other: No interventions

Interventions

No interventions OTHER

This retrospective observational study involves no therapeutic interventions, no treatment modifications, no patient contact, and no comparison groups. It is purely diagnostic technology development and validation using existing historical data.

Images of patients over 18 years old

Eligibility Criteria

• Age: 18 Years+
• Sex: all
• Healthy Volunteers: Yes
• Age Groups: Adult (18-64), Older Adult (65+)
• Sampling Method: Non-Probability Sample

Study Population

Ophthalmic medical images from adult patients (≥18 years) who received eye care at two ophthalmology centers in Costa Rica (Asociados de Mácula y Vítreo de Costa Rica in San José, and Centro Ocular in Heredia). Images were captured during routine clinical care for various clinical indications, including routine screening, follow-up visits, and diagnostic evaluations. The study population represents the real-world spectrum of patients seeking ophthalmology care in these centers, including healthy individuals, patients with various stages of eye diseases, and patients with multiple ocular pathologies.

You may qualify if:

• Image corresponds to patient ≥18 years of age at time of capture
• Image modality is one of: fundus photography, posterior segment OCT, anterior segment photography, automated perimetry, or video-OCT
• Image quality sufficient for diagnostic interpretation (adequate resolution, focus, illumination, complete visualization of anatomical area of interest, no major artifacts)
• Minimum clinical data available (age or age group, sex, and diagnosis or clinical indication)
• Image captured during routine clinical care (not specifically for research)
• No patient objection to use of medical data for research (when applicable per center policy)

You may not qualify if:

• CLINICAL:
• Images from eyes with recent intraocular surgery (\<3 months)
• Images from eyes with severe ocular trauma distorting anatomy
• Images from patients with rare or unique ocular pathologies not allowing generalization
• Images post-recent laser treatment where acute changes may confuse analysis
• TECHNICAL:
• Severely degraded image quality (extreme blur, severe under/overexposure, major artifacts preventing interpretation)
• Duplicate images of same eye on same date
• Images with missing or clearly erroneous metadata
• Images in non-standard or corrupted formats that cannot be processed
• Sex/Gender: All Minimum Age: 18 Years Maximum Age: No limit Accepts Healthy Volunteers: Yes (images of healthy eyes without pathology are included as controls)

Sponsors & Collaborators

• Marisse Masis-Solano: lead
• Iriscience Inc: collaborator

Study Sites (2)

Centro Ocular

Heredia, Centro, Costa Rica

Location

Asociados de Mácula y Vítreo de Costa Rica

San José, Provincia de San José, Costa Rica

Location

Related Publications (3)

• Li Z, He Y, Keel S, Meng W, Chang RT, He M. Efficacy of a Deep Learning System for Detecting Glaucomatous Optic Neuropathy Based on Color Fundus Photographs. Ophthalmology. 2018 Aug;125(8):1199-1206. doi: 10.1016/j.ophtha.2018.01.023. Epub 2018 Mar 2.

  PMID: 29506863 BACKGROUND

• De Fauw J, Ledsam JR, Romera-Paredes B, Nikolov S, Tomasev N, Blackwell S, Askham H, Glorot X, O'Donoghue B, Visentin D, van den Driessche G, Lakshminarayanan B, Meyer C, Mackinder F, Bouton S, Ayoub K, Chopra R, King D, Karthikesalingam A, Hughes CO, Raine R, Hughes J, Sim DA, Egan C, Tufail A, Montgomery H, Hassabis D, Rees G, Back T, Khaw PT, Suleyman M, Cornebise J, Keane PA, Ronneberger O. Clinically applicable deep learning for diagnosis and referral in retinal disease. Nat Med. 2018 Sep;24(9):1342-1350. doi: 10.1038/s41591-018-0107-6. Epub 2018 Aug 13.

  PMID: 30104768 BACKGROUND

• Gulshan V, Peng L, Coram M, Stumpe MC, Wu D, Narayanaswamy A, Venugopalan S, Widner K, Madams T, Cuadros J, Kim R, Raman R, Nelson PC, Mega JL, Webster DR. Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs. JAMA. 2016 Dec 13;316(22):2402-2410. doi: 10.1001/jama.2016.17216.

  PMID: 27898976 BACKGROUND

Related Links

• Information regarding software as medical device

MeSH Terms

Conditions

Macular Degeneration; Diabetic Retinopathy; Glaucoma; Keratoconus; Cataract

Condition Hierarchy (Ancestors)

Retinal Degeneration; Retinal Diseases; Eye Diseases; Diabetic Angiopathies; Vascular Diseases; Cardiovascular Diseases; Diabetes Complications; Diabetes Mellitus; Endocrine System Diseases; Ocular Hypertension; Corneal Diseases; Lens Diseases

Study Officials

• Marisse Masis-Solano

  Iriscience Inc

  STUDY DIRECTOR

Central Study Contacts

Marissé Masís Solano, MD PhD

CONTACT

19293278463; marisse@irisciencetech.com

Lihteh Wu, MD

CONTACT

marisse@irisciencetech.com

Study Design

• Study Type: observational
• Observational Model: OTHER
• Time Perspective: RETROSPECTIVE
• Sponsor Type: INDUSTRY
• Responsible Party: SPONSOR INVESTIGATOR
• PI Title: Collaborator

Study Record Dates

• First Submitted: March 23, 2026
• First Posted: March 27, 2026
• Study Start: May 1, 2026
• Primary Completion (Estimated): May 1, 2028
• Study Completion (Estimated): May 1, 2029
• Last Updated: April 1, 2026

Record last verified: 2026-03

Data Sharing

• IPD Sharing: Will not share

Locations

CO (2)