An Overview of Public Retinal Optical Coherence Tomography Datasets: Access, Annotations, and Beyond

Area of Science:

• Ophthalmology
• Medical Imaging
• Data Science

Background:

• Optical Coherence Tomography (OCT) is essential for diagnosing and planning treatments in ophthalmology.
• Publicly available OCT datasets accelerate research by enabling algorithm development with diverse imaging data.

Purpose of the Study:

• To provide a comprehensive analysis of publicly available OCT datasets.
• To create a catalog of OCT datasets, focusing on properties, disease representation, and accessibility.
• To identify challenges in dataset accessibility, format, annotations, and metadata.

Main Methods:

• Systematic review and analysis of publicly available Optical Coherence Tomography (OCT) datasets.
• Evaluation of dataset properties including disease representation and accessibility features.
• Cataloging of existing OCT imaging resources.

Main Results:

• Inconsistent accessibility, data formats, annotations, and metadata across OCT datasets hinder efficient research.
• Significant variation exists in the representation of different diseases within available datasets.
• There is a clear need for improved data-sharing practices and standardized documentation for OCT resources.

Conclusions:

• Standardization and improved accessibility of OCT datasets are crucial for advancing ophthalmology research.
• A comprehensive catalog of OCT datasets can enhance data discovery and utilization.
• Addressing data inconsistencies will foster more robust algorithm development and clinical applications.