An Approach to Measure Blood Flow in Single Choroidal Vessel Using Doppler Optical Coherence Tomography

Abstract

Purpose.: To evaluate the absolute blood flow rate in a single choroidal vessel using Doppler optical coherence tomography (OCT).

Methods.: Three choroidal vessels were selected in the right eye of three normal subjects, and were measured with Doppler OCT at a 1020-nm probe wavelength. The pulsatile change of the blood flow was obtained from synchronized measurement of Doppler OCT and plethysmography. Absolute blood flow rates in choroidal vessels were calculated from Doppler OCT volume data.

Results.: The cyclic change of the blood flow was quantitatively obtained. Absolute blood flow velocities and blood flow rates at peak systole [mean (SD)] were 46.9 (12.5) mm/s and 5.9 (3.6) μL/min, respectively. The coefficient of variation of three sets of measurements [mean (SD)] was 9.3 (4.9%).

Conclusions.: Doppler OCT and plethysmography provided an accurate quantitative assessment of the blood flow in choroidal vessels. This measurement technique could prove valuable to the study of choroidal blood flow in normal and pathologic conditions.

Introduction

The choroidal vasculature is a major supply for the outer retina, including the photoreceptors. ¹ The evaluation of age-related macular degeneration and various other retinochoroidal diseases requires precise measurement of choroidal blood flow. ^1,2 Previously, techniques such as pulsatile choroidal blood flow, ³ indocyanine green angiography (ICGA), ⁴ laser speckle flowgraphy ⁵ and laser Doppler flowmetry ⁶ have been used to evaluate human choroidal blood flow in vivo. These techniques have yielded significant findings about the mechanisms of choroidal blood flow, but none have proven capable of measuring the absolute blood flow rate. 

Doppler optical coherence tomography (OCT) is an extension of OCT capable of measuring Doppler shifts arising from blood flow. ^7,8 This technique provides depth-resolved, cross-sectional images of the retina and choroid while preserving the information derived from blood flow. Two different applications of Doppler OCT have been reported for the evaluation of retinal circulation: the first involves three-dimensional (3-D) volumetric ocular vascular imaging, in which Doppler signals are used as a contrast source to noninvasively pinpoint the location of blood flow in structural OCT, ^9–16 and the second involves a quantitative evaluation of retinal blood flow that permits the calculation of absolute blood flow rate in the human retina in vivo. ^17,18 In these latter studies, retinal blood flow measurements with Doppler OCT were performed with light source of 840 nm. ^17,18 It is known that light with a wavelength of approximately 1.0 μm is absorbed less in the retinal pigment epithelium (RPE) ^19,20 ; therefore, Doppler OCT at this wavelength is expected to be a suitable tool for measuring choroidal blood flow. ^14,20,21 In this study, we evaluated the absolute blood flow rates in choroidal vessels by Doppler OCT at 1020 nm, and evaluated the accuracy and reproducibility of this measurement. 

Methods

Subjects

All subjects were treated in accordance with the Declaration of Helsinki and approved by the Institutional Review Boards of the University of Tsukuba and Tokyo Medical University. The right eyes of three healthy male volunteers of varying ages (subject A: 52 years, subject B: 35 years, and subject C: 25 years) were examined; volunteers did not display any detectable ocular or systemic diseases. 

Doppler OCT

Blood flow velocity and the diameter of choroidal vessels were measured by prototype Doppler OCT, and a pulsatile flow profile was reconstructed by classifying each Doppler OCT scan according to each profile's pulse phase. ²² The prototype Michelson interferometer-based Doppler spectral-domain OCT system used in this study was built by Computational Optics Group at the University of Tsukuba. The system's light source was a superluminescent diode with a central wavelength of 1020 nm and a full-width-of-half-maximum bandwidth of 100 nm (Superlum, Carrigtwohill, Ireland). A high-speed indium gallium arsenide line-scanning camera with 1024 pixels (Goodrich, Charlotte, NC) was used as the detection system. The measurement speed was 47,000 depth-scans/s, and the depth resolution was measured to be 4.3 μm; the depth measurement range was 1.3 mm. 

The Doppler shift of the OCT signals was calculated using the phase difference between the adjacent depth-scans where the phase difference was obtained by complex division of adjacent depth-scans. The time interval between the two sequential depth-scans was 21.3 μs, and the integration time of the line camera was 13.9 μs. Based on these parameters, the maximum detectable Doppler shift was defined as 23.5 kHz and the corresponding axial velocity component in the eye as 8.68 mm/s. The dominant factor, which limits the minimum detectable axial velocity, is the ratio of the spatial sampling spacing to the beam-spot diameter on tissues. ²³ In our system, the ratio of the spatial sampling spacing to the beam-spot radius was 0.33; thus, the minimum detectable axial velocity was expected to be at least 0.51 mm/s with 5 × 3 (lateral × axial) averaging filter. 

Blood Flow Measurement

A raster-scanning protocol of each OCT volume with 750 depth-scans × 64 B-scans covering a 2.0 × 2.0-mm region on the retina was used for volumetric scans (Fig. 1). Three-dimensional vascular imaging, so-called Doppler optical coherence angiography, ⁹ was used to detect the choroidal vessels with bulk motion correction. ⁹ Squared Doppler shift (i.e., the power of Doppler shift) underneath the RPE was projected (Fig. 2). ⁹ Because the detectable velocity range of Doppler OCT imaging is limited at relatively high velocity, it was assumed that the visualized vessels were arterial. Hence, the obtained choroidal vascular images resembled an arterial-phase ICGA image (Fig. 2). 

Figure 1. 

 

Fundus photograph and Doppler OCT images obtained from right eye of subject A. In fundus photograph (A), the white box indicates the measurement area of the Doppler OCT volume, and the black line indicates the scanning line of the B-scan OCT image (B, C). (B) B-scan intensity OCT image. (C) Bidirectional Doppler OCT image. White arrow indicates the Doppler signal from the measured choroidal vessel.

Figure 2. 

 

En-face Doppler OCT vascular images obtained from subject A (A), subject B (B) and subject C (C). White arrows indicate the direction of choroidal blood flow. White circles show the measurement points of the choroidal blood flow rate. (D) Arterial phase of indocyanine green angiography obtained from subject A. The white box indicates the measurement area of Doppler OCT. (E) High magnification image of indocyanine angiography the area of the white box in (D)

Measurement points were arbitrarily selected from where clear choroidal vascular images could be obtained in en-face Doppler choroidal vascular imaging, and three measurement points were selected in each subject (Fig. 2). The axial location of the choroidal vessel in the Doppler OCT B-scan image was manually detected at each measurement point (Fig. 1). The Doppler phase shift at the center of the choroidal vessel was measured, and the average of the Doppler phase shift of two adjacent Doppler OCT B-scan images was used for further calculation. The Doppler angle (θ) was computed from the relative axial position of the measurement points in the adjacent Doppler OCT B-scan image. Blood flow velocity (V) was then calculated as 

 

where λ_c is the central wavelength of the light source, f_A is the A-scan rate of OCT, Δφ is the Doppler phase shift, and n is the refractive index of the tissue. Based on the assumption that there is a parabolic distribution of the flow in a lumen, blood flow rate (F) was calculated as 

 

where V_max is equal to the blood flow velocities at the center of the blood vessel and D is the vessel diameter as measured using an en-face Doppler OCT choroidal vascular image. Blood flow direction was determined with Doppler frequency shift and 3-D vascular images (Fig. 2). Because the choroidal vessel in each measurement point was nearly perpendicular to the incident beam, the observed Doppler frequency shifts were relatively small. Thus, excessive phase-wrapping did not occur in each measurement. 

Using previously published methodology, we obtained a pulsatile change of blood flow in a single choroidal vessel for each subject from the synchronized measurement of Doppler OCT volume and plethysmograph. ¹⁸ The acquisition speed of each volume measurement was 1.02 s/volume, covering an average of 1.6 heartbeats. For pulse synchronization, 10 sets of four Doppler OCT volumes were measured continuously. As a consequence, 40 sets of Doppler OCT volume data were obtained, and we chose the best 20 volumes based on motion artifact and sensitivity. 

Plethysmography was recorded by a pulse oximeter with a finger probe as the Doppler OCT measurement was performed. Doppler flow signals were classified as belonging to one of seven heartbeat phases based on the plethysmography data (Fig. 3). The pulse curves of the Doppler signals during a single-heartbeat period were synthesized based on the classified heartbeat phase (Fig. 3). For each subject, three sets of choroidal blood flow measurement with pulse synchronization were performed, and an average of three measurements was used for evaluation. 

Figure 3. 

 

X-marks and the solid line indicate the pulsatile change of the choroidal blood flow rate at V1 in subject A. The dotted line indicates parallel recorded plethysmographs obtained with a pulse oximeter.

Results

The pulse curve of the Doppler signals during the single-heartbeat measurement period was successfully synthesized based on the classified heartbeat phase (Fig. 3), ²² and mean blood flow rates at peak systole were calculated. Absolute blood flow velocities at peak systole [mean (SD)] were 46.9 (12.5) mm/s, and blood flow rates at peak systole were 5.9 (3.6) μL/min (Table). Vessel diameter [mean (SD)] was 70 (14) μm. Peak blood flow velocities showed a significant positive correlation with vessel diameter (R ² = 0.46, P = 0.045; Fig. 4). The coefficient of variation of the three measurements [mean (SD)] was 9.3 (4.9%) (Table). 

Figure 4. 

 

Scatterplot of vessel diameter and peak blood flow velocity.

Table. 

 

Blood Flow Measurement in Choroidal Vessels in Three Normal Human Subjects

Subject	Vessel	Peak Velocity (mm/s)	Flow Rate (μL/min)	Diameter (μm)	Angle (deg C)	CV (%)
A	V1	45.2	6.0	75	80	6.4
	V2	27.5	3.4	73	81	5.6
	V3	65.6	9.9	80	76	4.9
B	V4	60.2	12.8	95	80	13.4
	V5	52.4	4.9	63	83	14.9
	V6	55.7	7.6	76	85	8.5
C	V7	43.0	3.6	60	80	11.4
	V8	35.8	2.1	50	80	16.4
	V9	37.1	2.6	55	79	2.2
	mean (SD)	46.9 (12.5)	5.9 (3.6)	70 (14)	85.0 (1.4)	9.3 (4.9)

 CV, coefficient of variation.

For subjects A and B, choroidal vessels before and after a bifurcation were used to evaluate measurement accuracy. Subject C was not used in this accuracy evaluation because no bifurcation with clear Doppler signal was observed. In subject A, V3 represented incoming blood flow, and V1 and V2 represented outgoing blood flow (Fig. 2). The mean total outgoing blood flow rate and incoming blood flow rate were 9.9 μL/min and 9.4 μL/min, respectively (Table). In subject B, V4 represented incoming blood, and V5 and V6 represented outgoing blood flow (Fig. 2). The mean total outgoing blood flow rate and incoming blood flow rate were 12.8 μL/min and 12.3 μL/min, respectively (Table). 

Discussion

In this study, pulse synchronization of Doppler OCT at 1020 nm proved capable of successfully measuring the in vivo absolute blood flow rate in choroidal vessels in human beings. Measurement accuracy was confirmed by comparing outgoing and incoming blood flow rates at bifurcation in two of the three subjects. 

Laser Doppler flowmetry has been widely used to evaluate human choroidal blood flow in various pathophysiological conditions. ^1,6,24,25 For measuring choroidal blood flow, Doppler OCT varies greatly from laser Doppler flowmetry in several respects. First, laser Doppler flowmetry provides only a relative blood flow rate; hence, it is not relevant to compare two or more values obtained at different locations of a single subject or multiple subjects. On the other hand, unlike laser Doppler flowmetry, Doppler OCT is capable of providing an absolute blood flow rate. It potentially enables the comparison of multiple values obtained at different locations or from different subjects. Second, laser Doppler flowmetry measures relative net flow in the capillaries in a certain region, ^6,24 while Doppler OCT measures the absolute flow in a specific vessel; thus, these two modalities measured different quantities associated with choroidal circulation. In addition, the depth resolution of Doppler OCT enables spatially 3-D selection of the specific vessel to be measured. We expect that the combination of laser Doppler flowmetry and Doppler OCT might provide comprehensive evaluation of choroidal blood flow. 

There were several limitations to this study. First, the axial position of each measurement point in the Doppler OCT B-scan image was manually detected. The measurement error of absolute blood flow velocity depending only on the Doppler angle detection ΔV can be expressed as 

 

where Δθ is the Doppler angle measurement error. The error of absolute blood flow velocity when the Doppler angle θ = 80° and its discrepancy is 1° (Δθ = π/180 [rad]), ΔV/V = 0.099 (∼10%). An automatic analysis of the Doppler angle is generally preferred, as it provides increased objectivity and generality of the analysis, as well as a considerably lower error rate. The use of a bidirectional Doppler technique ²⁶ might be an effective alternative approach for automatic analysis of the Doppler angle. 

Second, 40 sets of OCT volume measurement were used for pulsatile flow acquisition, requiring nearly 10 minutes for each measurement series; this extended scanning time impedes our technique's applicability in a clinical setting. The use of a faster Doppler OCT system might decrease the measurement time. 

Third, the selected vessels in this study were almost perpendicular to the incident light; hence, the phase shift was within the measurement range. If the incident angle is less perpendicular than that examined here, the axial component of the flow velocity, which is the quantity measured by the Doppler OCT, becomes large, and phase wrapping and fringe washout occur; these changes disturb the flow evaluation. Although increasing the A-scan rate is a potential solution to this problem, this would also increase the minimum detectable velocity. Because the minimum detectable velocity is improved with a high signal-to-noise ratio and advanced Doppler detection technology, such as dual-beam Doppler OCT, ^15,16 further development of the Doppler OCT system will widen the Doppler measurement range and achieve effective measurement of multiple choroidal vessels. 

Fourth, in the current condition, there are no alternative techniques to measure the absolute blood flow rate in human choroidal vessels in vivo. Because no current method can be a reference standard to classify the choroidal vessels into arteries and veins, we cannot fully validate that the vessels examined in this study were arteries. However, according to the comparison between the Doppler OCT and a corresponding ICGA, the vessel can be believed to be an artery. 

Fifth, the diameter of the specific vessel imaged for each subject was calculated from Doppler OCT images; hence, the diameter could be slightly underestimated by our analysis. Further development of an advanced algorithm for the estimation of vessel diameter would increase measurement accuracy. 

As observed in retinal blood flow, ²⁷ the blood flow velocities in choroidal vessels increased with diameter. Some similarity might be expected between choroidal and retinal blood flow. However, vascular architecture of the choroid is completely different from the retina, ²⁸ and a difference in auto-regulation between these blood flows has been reported. ¹ The blood flow rate in the choroidal vessels measured in this study was higher compared with retinal blood flow with an equivalent vessel diameter. ²⁷ Comparison of choroidal and retinal blood flow, including responses to various stimulations, could provide important insight into the mechanisms and supply of ocular blood flow. Further analysis is required to evaluate the specificity of choroidal blood flow. 

In conclusion, absolute blood flow velocities and absolute blood flow rate in a single choroidal vessel can be quantitatively determined by Doppler OCT with plethysmography. Further studies with a greater number of subjects are required to explore fundamental details about the regulation of choroidal blood flow in normal and pathological conditions. 

Footnotes

 Supported by KAKENHI (24592682) and the Japan Science and Technology Agency through a program of the Development of Systems and Technology for Advanced Measurement and Analysis.

Footnotes

 Disclosure: M. Miura, None; S. Makita, None; T. Iwasaki, None; Y. Yasuno, None