Age-Related Macular Degeneration Benchmark Imaging Dataset (ABID)

Overview

1000 participants from up to 25 international locations who are at least 50 years old with either healthy eyes or a diagnosis of Age-Related Macular Degeneration (AMD) will be consented to provide images of their eyes for a new dataset. This dataset is an important step in developing an Artificial Intelligence (AI) based screening tool for AMD.

Age-related macular degeneration (AMD) is a leading cause of blindness in the world. The early and intermediate stages of AMD progress to late stage resulting in vision loss due to either geographic atrophy or neovascular AMD. Preventive measures to reduce treatment burden and prevent blindness are important. While there are no approved therapies for early disease, active research is underway. However, advancing therapeutic trials for AMD prevention remains challenging due to the lack of primary care involvement in early-stage AMD diagnosis. A clear need exists for a reliable screening tool, deployable in optometry or primary care facilities to identify those with an early stage of disease. While an Artificial Intelligence (AI) based screening tool would be an ideal solution, there are obstacles to its prospective validation due to difficulty in enrolling sufficient cases in a screening environment. A key step towards promoting this field lies in collecting a diverse dataset of multimodal imaging and providing a reference standard level 1 classification from a wide array of patients with AMD. Having such a benchmark dataset available for research purposes will empower the development and validation of AI models for AMD. This data can serve as the much-sought pathway to the rapid development of screening models, facilitating the referral of patients with the appropriate disease spectrum, and ultimately leading to the prevention of late AMD. This is a prospective, cross-sectional, multi-center, observational study to collect and develop a meticulously curated and diverse AMD benchmark dataset, featuring reference standard level 1 classification and comprehensive annotation of images. Study procedures include: * Patient history: demographics (age, sex, ethnicity, race), smoking history, family history of AMD * Physical exam: height and weight * Snellen best-corrected visual acuity (BCVA) * AMD classification (Beckman scale) Eligible participants will undergo one retinal imaging session of both eyes for the following: * Single field stereo color fundus photography (cFP) - pre and post dilation * Macular spectral domain-optical coherence tomography (SD-OCT) Due to the need for diversity in the dataset, sites need to represent a wide array of geographical locations. In addition, balancing will be done on enrolled participants with or without AMD including age, sex, and AMD level (worse eye selected as study eye).

Conditions

Interventions

Eligibility

Locations (14)

Outcomes

Central Contacts