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

U.S. Food and Drug Administration

A 3D Quantitative Optical Coherence Tomography Angiography Method for Measuring Retinal Blood Flow in Humans

Catalog of Regulatory Science Tools to Help Assess New Medical Devices 

This regulatory science tool (RST) is a quantitative optical coherence tomography angiography (qOCTA) method to measure the absolute velocity of individual blood cells within retinal microvasculature and map 3D flow rates.

Technical Description

This quantitative OCTA method measures absolute blood velocity and vessel lumen diameter to calculate retinal blood flow (RBF) rates within retinal microvasculature. Blood velocity is measured by tracing streaks from individual blood cells within a set of B-scans used to create an OCTA volume. The streaks are processed with a modified 3D Radon transform to extract depth-resolved velocity for individual blood cells. We use a high-speed adaptive optics (AO) system with 3.4 MHz Fourier domain mode-locked swept-source OCT to achieve the micron-level spatial and fractional millisecond temporal resolution to track individual retinal blood cells coursing through vessels. OCT systems with appropriately high spatio-temporal resolution may be compatible with this method, although performance for other systems has net been investigated. Further, the tool requires no modification to a typical OCTA image acquisition protocol. The tool includes a processing pipeline for measurement of single cell velocity, vessel caliber and a 3D flow rate (in absolute µL/min units) using the same OCT intensity volume. 

A typical OCTA imaging protocol (1.5° × 1.5° field of view with 8 repeat B-scans) is adopted to simultaneously generate the OCT intensity volume and the corresponding OCTA volume [1, 2]. The OCTA volume, processed using a typical amplitude decorrelation method, is used to quantify morphological parameters such as vessel diameter and 3D vessel orientation that are needed for blood flow calculation. Individual streaks caused by blood cell motion in the intensity volume are analyzed with a modified 3D Radon approach to extract depth-resolved streak orientation. Velocity and flow rates are computed voxel-wise by integrating vessel morphology and streak orientation measurements and visualized in MATLAB [2]. Velocity profiles within vessels and a wide dynamic range (0.5–54 mm/s), including inner retinal capillaries, are demonstrated. This methodology enables single-volume, quantitative mapping of flow rates in a research setting. The approach is agnostic to the type of particle producing streaks in the spatio-temporal OCT volume, making it adaptable for reporting velocity and flow rates of blood cells with or without contrast agents or any moving particles in 3D (e.g., flow cells or chambers used in non-biologic applications). Applicability to non-biologic applications has not been evaluated in this study and is noted as a potential future research direction only.

The capabilities demonstrated of this approach include:

  • The ability to measure absolute RBF rates on an individual cell and individual vessel segment basis.
  • The ability to measure RBF rates from any vessels within a 3D tissue slab, assuming it is within the penetration depth of the source light so that an OCTA image can be acquired.
  • A wide dynamic range including: the ability to measure parabolic flow profiles across larger vessels with laminar flow conditions, and the ability to track single-cell flow in capillaries.
  • The ability to distinguish between different types of vessels, such as arteries and veins, based on streak orientation.
  • Temporal resolution to measure pulsatile flow in vessels during the cardiac cycle (systole-diastole).
  • High contrast OCT images to extract vessel wall metrics as input into the blood flow calculation. This capability may support future research into autoregulation and structural vascular changes, though disease applications were not investigated in this study. 

Intended Purpose 

This tool is intended to help measure absolute blood velocity and volumetric flow rate within human retinal vasculature using OCT. The method is applicable to high-resolution OCT devices. This RST may be relevant to, but is not limited to, submissions under product codes SAX and OBO. Applicability should be assessed on a case-by-case basis.

Testing

All demonstrated capabilities are based on a study of six healthy volunteers [1, 2]. In this study, 20 retinal vessels with caliber between 5 and 120 µm were quantified with the qOCTA method. A wide range of blood velocities was captured from 0.5 to 54 mm/s, including low flow rates in retinal capillaries that surround the foveal avascular zone. Accuracy was demonstrated in a branched arteriole where conservation of flow rates was measured to have <4% difference between the parent and sum of the branches. The results showed a moderate linear correlation between mean velocity and vessel diameter with a slope of 0.18 ± 0.04 (r2 = 0.53). The regression analysis between the mean blood flow rate and the vessel diameter (log-log) showed a slope of 2.55 ± 0.15 (r2 = 0.94). The velocity profile in depth in retinal vessels showed a blunt parabolic shape.

Limitations

The method provides greater precision and accuracy at the expense of computation speed. The accuracy of determining streak orientation using the Radon transform can be limited when prioritizing rapid analysis. The resonant scanner frequency, which determines the B-scan rate, dictates the temporal resolution and, consequently, the measurable velocity range. At the lower end, the ability to measure single-file flow at a few mm/s in retinal capillaries is constrained by the sensitivity required to detect very few short streak angles. While all vessels within the field of view can produce streaks, discrete streak quantification requires a shallow intersection angle relative to the fast-scanning axis. This limitation, due to the fixed orientation of the fast axis scanner in the imaging system, reduces the number of vessels that can be accurately quantified, limiting flow maps to specific regions.

Supporting Documentation

  1. Raghavendra, A., Saeedi, O., Hammer, D. X., & Liu, Z. (2024). Absolute Retinal Blood Velocity and Flow Rate Measurement in Humans with Adaptive Optics-Optical Coherence Tomography Angiography. Investigative Ophthalmology & Visual Science, 65(7), 4944-4944.
  2. Raghavendra, A. J., Saeedi, O. J., Hammer, D. X., & Liu, Z. (2026). Single-scan adaptive optics-enabled quantitative optical coherence tomography angiography for absolute three-dimensional retinal blood flow mapping. Optica, 13(1), 125-134.

Contact

Tool Reference