U.S. flag An official website of the United States government

U.S. Food and Drug Administration

AO-OCTA Method and Dataset for Human Choriocapillaris Imaging

Catalog of Regulatory Science Tools to Help Assess New Medical Devices 

This regulatory science tool (RST) is an adaptive optics – optical coherence tomography angiography (AO-OCTA) method to resolve and quantify retinal vessels including the choriocapillaris (CC) in the living human eye, along with a dataset of CC montages and their morphologic metrics across the macula from healthy individuals. This RST has multiple purposes: 1. It offers one possible approach for performing OCTA for vessel imaging; 2. It describes an approach for measuring CC image quality using a new flow signal to noise ratio (SNR) metric; 3. The dataset can serve as a normative baseline for future studies of retinal pathology and as labelled ground truth data for the assessment of novel algorithms for CC segmentation.

Technical Description

Twenty healthy volunteers were imaged with a custom, high-speed Fourier domain mode-locked laser (FDML)-based AO imaging platform [1]. The following describes key hardware and acquisition parameters of the AO-OCT platform. The AO-OCT subsystem operates at an A-line acquisition rate of 3.348 MHz set by the buffered FDML swept-source laser (λc=1060 nm, ∆λ=76 nm) providing a nominal axial resolution of 8.4 µm in retinal tissue (refractive index n=1.38). The system transverse resolution is estimated to be 2.9 µm with a 7.5 mm pupil diameter. To operate in the angiography mode, similar to conventional OCTA scanning, the resonant scanner (3.27 kHz resonant frequency) of the system uses a MB-scan mode, in which a series of M subsequent B-scans (512 A-lines/B-scan, bi-directional scans) is acquired at the same retinal location before the slow scanner moves to the next position to continue to form an AO-OCTA volume. A total of 512 separate slow scan positions are captured in each volume, which equates to 512×n total B-scans, where n is number of repeated fast scans per slow scan position. 

For the dataset, one eye of each participant was imaged, and that eye was dilated and cycloplegia induced with Tropicamide 1%. After alignment in the system, the fixation target was set to direct the gaze of the participant to eleven regions of interest (ROIs) from the fovea to 12.5° temporal retina, where each imaging location was separated by 2.5°. At each location, 3° × 3° field-of-view (FOV) (approximately 0.9 x 0.9 mm at the retina) AO-OCT volumes were recorded, and the total imaging region occupies approximately 27°×3° at the retina. Factors that impact CC image quality, including number of repeated B-scans, volume averaging, imaging field of view (FOV), and angiography method, were quantified. 

The final AO-OCTA volumes used for CC quantification were acquired with eight repeated B-scans and a speckle variance (SV) method was used for OCTA calculation. En-face CC images were created from the AO-OCTA volumes by summing depth pixels in a slab ~7-11 µm in thickness (5-7 pixels) using average intensity projection, where the thickness varied depending on location and OCTA signal strength. The en-face CC images were then stitched together to create the final montage across the macula, where the best quality in overlapped regions was chosen by visual inspection. From the en-face CC macular montage, 1°×1° regions of interest (ROI) were extracted, spaced at approximately 1° increments for further extraction of CC capillary metrics. Key CC morphological metrics include vessel density, vessel diameter, void diameter, void area, and tortuosity. 

Intended Purpose 

This tool is a high-resolution imaging method to resolve the CC as well as a normative dataset of CC morphological measures from human subjects. It is designed to be used by retinal imaging device developers, manufacturers, clinicians, and researchers who perform OCTA for retinal vessel imaging, analyze CC image quality, or develop artificial intelligence/machine learning (AI/ML) algorithms for vessel morphology quantification and vessel segmentation.

Testing

The method was demonstrated in twenty healthy volunteers, and the measurement results were compared with other published human in vivo and ex vivo reports, details of which are provided in [2, 3]. In the initial study, a minimum 9 dB flow SNR was needed for the method to reliably resolve individual CC segments. All CC capillary metrics except lobule count were independently calculated by two graders, and Lin’s concordance coefficient (LCC) was used to assess inter-rater variability between the graders. We observed a continuous CC capillary network for all HVs. LCC values ranged from 0.65 to 0.91 across various CC metrics. 

Limitations

The limitations of this dataset include:

  • The FOV of single AO-OCTA dataset (3°×3°) is still relatively small as compared with a clinical OCTA image (10°×10°).
  • The study cohort age range (mean±SD: 43.1±13.0 years) may limit generalizability to older populations.
  • The shadow artifact – a known imaging artifact in OCTA caused by flow dynamics in overlying retinal vessels – is present in this dataset and has not been corrected in the current implementation. 

Supporting Documentation

Additional data description and details on data access can be found here:  https://www.synapse.org/Synapse:syn64741794/wiki/635416

Peer-reviewed journal article describing the dataset [2-3], the imaging system [1] and the AO-OCTA method [3]:

  1. Liu Z., Zhang, F., Zucca K., Agrawal A., and Hammer, D.X., Ultrahigh-speed multimodal adaptive optics system for microscopic structural and functional imaging of the human retina. Biomed Opt Express, 2022;13(11):5860-5878. https://doi.org/10.1364/BOE.462594.
  2. Hammer D.X., Kovalick K., Saeedi O.J., Cukras C., and Liu Z., Quantification of human choriocapillaris measures across the macula. Invest. Ophthalmol. Vis. Sci. 2025;66(12):26. https://doi.org/10.1167/iovs.66.12.26.
  3.   Liu Z., Hammer D.X., Wide-Field Choriocapillaris Mapping with 3.4 MHz Adaptive Optics – Optical Coherence Tomography Angiography. Biomed. Opt. Express 16, 3255-3269 (2025). https://doi.org/10.1364/BOE.550936.

Contact

Tool Reference