Noninvasive and Direct Monitoring of Erythrocyte Aggregates in Human Retinal Microvasculature Using Adaptive Optics Scanning Laser Ophthalmoscopy

Abstract

Purpose.: To investigate erythrocyte aggregates in parafoveal capillaries by adaptive optics scanning laser ophthalmoscopy (AO-SLO).

Methods.: AO-SLO videos were acquired from the parafoveal areas of one eye in 10 healthy subjects. Erythrocyte aggregates were detected as “dark tails” that were darker regions than vessel shadows. The lengths of the dark tails were measured in target capillaries, and their time-dependent changes in length were analyzed using spatiotemporal images. The dark tail elongation rate was calculated as the change of dark tail length per unit length of the target capillary.

Results.: The overall average dark tail length was 112.1 ± 36.9 μm. The dark tail became longer in a time-dependent manner in every monitored capillary (P < 0.0001). The dark tail elongation rate and average velocity were 0.51 ± 0.37 and 1.49 ± 0.36 mm/s, respectively.

Conclusions.: AO-SLO can be used for noninvasive and direct monitoring of blood dynamics in the retinal microvasculature without dying agents. Erythrocyte aggregates were detected as dark tails and were elongated in a time-dependent manner in the parafoveal capillaries of normal subjects. Monitoring the characteristics of dark tails has promising potential for evaluating retinal hemodynamics.

Introduction

Retinal vessels can be observed directly through the window of the pupils. Information regarding abnormalities and diseases related to circulation are often first revealed by the observation of retinal vessels, which leads to speedy and appropriate treatment for patients. ¹ The appearance of vessels contributes to the grading of noteworthy systemic disorders such as hypertension, ^2,3 arteriosclerosis, and diabetic retinopathy. ^1,4 Furthermore, investigating hemodynamics in the human retina provides potentially helpful information and clues for understanding the pathogenesis of retinal vascular disorders such as diabetic retinopathy, ⁵ glaucoma, ⁶ age-related macular degeneration, ⁷ retinal vein occlusion, ⁸ and collagen diseases. ⁹  

Many approaches have been developed for evaluating blood vessels or dynamic blood flow, including the dye dilution technique, ^10–12 blue field entoptic phenomenon, ^13–15 laser Doppler velocimetry, ^16–19 laser speckle phenomenon, ²⁰ optical coherence tomography (OCT), ²¹ and adaptive optics scanning laser ophthalmoscopy (AO-SLO). ^22–25 These techniques, except for blue field entoptic phenomenon and AO-SLO, evaluate the blood flow in the retina but do not monitor the blood corpuscles in the retinal capillary. 

Many reports have described erythrocyte dynamics in animal studies and ex vivo capillary models and simulation. ^26–29 Erythrocyte behaviors such as aggregations, deformations, or abnormalities change in response to vascular status or the presence of diseases like sickle-cell anemia, ³⁰ diabetes mellitus, ³¹ systemic lupus erythematosus (SLE), ⁹ or Behçet's disease. ³² Meanwhile, few reports have described erythrocyte aggregations in the human eye with the exception of the conjunctiva bulbi. ³³ An in vivo approach for monitoring erythrocyte aggregates directly in the retinal microcirculation has not been reported. 

Recently, confocal AO-SLO has enabled imaging of retinal cells such as photoreceptors and leukocytes. ^5,24,34,35 AO-SLO is a SLO-equipped adaptive optics technology that provides high-resolution and high-contrast retinal images by correcting ocular aberrations. ^36,37 AO-SLO images allow the noninvasive monitoring of leukocyte movements as bright particles flowing in dark parafoveal capillaries and the measurement of their velocities without the use of contrast dyes. ^10–12 We previously reported that the bright particles moving in the dark vessel shadows may be reflections of the photoreceptor aggregates that pass through circulating transparent objects such as leukocytes or plasma gaps. Furthermore, we described the so-called dark tail, which could be seen as a region darker than the vessel shadow that occurred closely behind moving particles and might correspond to aggregated erythrocytes upstream of the leukocytes that block the AO-SLO laser (Fig. 1A). ²⁴  

Figure 1

 

Schema of dark tail elongation in the parafoveal capillary. (A) Diagrammatic representation of the optical properties of blood plasma and blood cells in the parafoveal capillaries. The adaptive optics scanning laser ophthalmoscopy (AO-SLO) laser passes through leukocytes, and the reflective light from photoreceptors (hexagons) enables leukocytes to be detected as bright moving objects when the scanning layer is focused on the photoreceptor layer. Erythrocyte aggregates prevent the AO-SLO laser from reaching the photoreceptor and are detected as dark shadows on the photoreceptor layer (dark tail). Blood cells flow in single file in the capillary lumen, and the dark tail follows closely behind the slow-moving leukocytes. (B) Leukocytes interfere with the flow of erythrocyte aggregates and cause the erythrocytes to become packed together sequentially, resulting in the detection of an elongated dark tail closely behind leukocytes in AO-SLO videos.

In this study, we performed direct, noninvasive, and objective quantification of the “dark tail” in the parafoveal capillaries of healthy human eyes by using our AO-SLO system. 

Methods

This study was approved by the Institutional Review Board and the Ethics Committee at Kyoto University Graduate School of Medicine and was performed in accordance with the tenets of the Declaration of Helsinki. Written informed consent was obtained from each participant after a detailed explanation of the nature and possible consequences of the study procedures. 

Subjects

AO-SLO videos were acquired from the parafoveal areas of 10 healthy Japanese subjects (mean age ± standard deviation [SD], 32.6 ± 6.7 years; range, 22–42 years) without a history of ocular or systemic diseases. All subjects were dilated before imaging with one set of tropicamide (0.5%) and phenylephrine hydrochloride (0.5%) drops and were examined for approximately 10 minutes per eye in a seated posture. 

Adaptive Optics Scanning Laser Ophthalmoscopy Imaging

We developed a novel AO-SLO system (Canon, Inc., Tokyo, Japan) with a high wavefront correction efficiency using a dual liquid-crystal phase modulator (LCOS-SLM; X10468-02; Hamamatsu Photonics, Hamamatsu, Japan), as described previously (Fig. 2). ³⁸ Briefly, the imaging wavelength was 840 ± 25 nm, and the wavelength of beacon light for the measurement of wavefront aberrations was 760 ± 5 nm. The imaging light and the beacon light were set at 330 and 40 μW, respectively, by calculating the incident power of both light sources in accordance with the safety limits set by the American National Standards Institute. ³⁹ The videos were acquired at 64 frames/s. The scan area was 1.4 × 2.8° at the retina and was sampled at 200 × 400 pixels. 

Figure 2

 

Schematic of the optical structure of AO-SLO. Two spatial light modulators (SLM) are used for wavefront correction, and the cylindrical lens (CyL) is used to compensate for the large astigmatism of eyes. The aspherical mirror is used to compensate for the aberration from the off-axis layout mirrors. The two red arrows represent the infrared light used to observe eye fixation.

Video Processing

Desinusoiding.

We calculated each pixel value of the AO-SLO image from the intensity of the reflective light measured at regular time intervals. The resonant scanner changes its scan speed sinusoidally, and the AO-SLO obtains the intensity of the reflective light at regular time intervals. Therefore, it obtains a few intensity values near the center and many intensity values near edge of the image. 

In this study, we presumed that the AO-SLO image was composed of the intensity of the reflective light from the imaging light that is scanned on the retina with the scanner. 

Each pixel value of the AO-SLO image was calculated using the weighted mean method based on the time for which the imaging light was scanned on the retina, under the assumption that its scan speed changed sinusoidally. 

Video Stabilization.

Raw videos were corrected for scanning distortions and were stabilized to correct for eye motion by using ARIA (AO-SLO Retinal Image Analyzer; Canon, Inc.). The registration technique of ARIA is based on scan line warping, and a fixed reference frame is used for image warping. ^24,40  

AO-SLO Movie Acquisition

AO-SLO videos were recorded for 4 seconds per scan area, and 10 to 25 scan areas were collected per subject to cover the parafoveal areas. AO-SLO imaging was performed by focusing on the photoreceptor layer to enable detection of the cone mosaic pattern. 

Quantitative Image Analysis of the Dark Tail

The dark tail was defined as a region darker than the vessel shadow that occurred closely behind the moving bright particles (Fig. 1B). In this study, we investigated the dark tail as a shadow of aggregated erythrocytes that blocked the AO-SLO laser. All digital images, except for video stabilization and capillary visualization, were processed by a single operator (SA) using the public-domain software ImageJ (available in the public domain at http://rsb.info.nih.gov/ij/index.html; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD). The grayscale ranged from 0 (black) to 255 (white). 

Capillary Visualization.

The capillary images were constructed as projections of the moving objects in sequential frames by using the motion contrast-enhancement technique, which was first introduced by Tam et al. as an image processing technique for capillary visualization using AO-SLO videos. ^41,42 In brief, motion contrast enhancement was conducted by dividing the pixels between sequential frames followed by calculating the variance of pixels among all division images in each X–Y position, resulting in visualization of the contrast-enhanced capillary images. Visualized vessels were used for the following analyses to set the regions of interest (ROI) along the target vessels precisely. All digital image processing for capillary visualization was manipulated by ARIA automatically. 

Velocity Measurement of the Dark Tail.

The target vessels with dark tail flow were selected from one branch to another to ensure that they were free of another bifurcation. The dark tail velocity was calculated by using a spatiotemporal image generated by reslicing the sequential frames, with the length of the line on the horizontal axis and the frame number on the vertical axis. ^24,43 After reslicing of the frames along the line set on the target vessel (Fig. 3A), the velocity was obtained by calculating the reciprocal of the slope of the borderline between the white band and black band depicted in the spatiotemporal image, which correspond to the loci of moving bright particles and the dark tail, respectively (Fig. 3B). A steeper slope indicates a lower velocity. We measured three individual capillaries per subject, which were selected randomly. The velocities of three individual successive dark tails were measured for each capillary, yielding 90 measurements. 

Figure 3

 

Measurement of dark tail velocity and length using a spatiotemporal image. (A) Schema of stacked sequential AO-SLO images. Red lines are focused on the target vessel in preparation for reslicing of the sequential images in order to generate the spatiotemporal image. (B) The dark tail velocity and length were calculated by using a spatiotemporal image that was plotted with the length of the target vessel on the horizontal axis and the frame number on the vertical axis. The white narrow band and the wider black band represent the trajectory of bright moving objects and dark tail, respectively. The thickness of the black band (t) represents the time required for the dark tail to pass through a point located on the target vessel. The reciprocal of the slope of the borderline between the white and black bands represents the velocity of the head of the dark tail. The spatiotemporal image was vertically separated into three zones from upstream to downstream of the blood flow (zones a, b, and c), and the dark tail length was calculated for each zone to evaluate the time-dependent changes in the dark tails. (C) Plot profile of the red line set on the spatiotemporal image in (B). Points x, y, and z and the length of time (t) correspond to x, y, z, and t in (B), respectively. The frame number of the gray value range that was lower than the averaged gray value was measured as the time required for the dark tail to pass the point on the target vessel. The dark tail length was computed by multiplying the velocity by the time. Dark tail velocities were calculated for each zone to evaluate the time-dependent changes in dark tails.

Measurement of Dark Tail Length.

Dark tail length was calculated by using a spatiotemporal image in which the dark tail velocities and the time required for the dark tail to pass the points on the target vessel could be seen simultaneously. First, a straight vertical line was set on the spatiotemporal image and a plot profile was prepared. The frame number of the gray value range that was lower than the background gray value was measured as the time required for the dark tail to pass the point on the target vessel (Figs. 3B, 3C). The background gray value was defined as the average gray value of a region free of dark tails and bright moving objects such as leukocytes or plasma gaps. Together with the velocity of the dark tail mentioned above, the dark tail length was calculated as follows:  

 

where l and v are the length and velocity of the dark tail, respectively, and t is the time required for the dark tail to pass through a point located on the target vessel. In addition, the spatiotemporal image was separated vertically into three zones from upstream to downstream of the blood flow (zones a, b, and c), and the dark tail length was calculated for each zone to evaluate time-dependent dark tail changes. The dark tail elongation rate was defined as follows:  

 

where r, L, l_c, and l_a are the elongation rate, the distance between zone a and zone c, the dark tail length at zone c, and the dark tail length at zone a, respectively. We measured three individual and randomly selected capillaries in each subject. The length of three individual dark tails was measured for each capillary, yielding a total of 270 measurements.  

Capillary Diameter Measurement

Capillary diameters were measured to evaluate the association with the elongation rate of the dark tail. The parafoveal capillary diameters were measured on the constructed capillary images and were measured for each of the three zones manually. 

Statistical Analysis

All values are presented as the mean ± SD. The statistically significant correlations between the elongation rate and velocity, between the elongation rate and diameter, between the velocity and diameter, and between the length of the dark tail and diameter were evaluated with the Pearson's correlation coefficient. The differences in dark tail length and capillary diameter among the three zones were evaluated with analysis of variance (ANOVA) followed by repeated measures ANOVA. All calculations were performed using StatView (version 5.0; SAS, Inc., Cary, NC) except for the intraclass correlation coefficient (ICC), which was calculated using SPSS (IBM SPSS statistics 19; IBM, Inc., Armonk, NY). P values less than 0.05 were considered significant. 

Results

Appearance of the Dark Tail

By using our prototype AO-SLO, we successfully captured bright moving objects and dark tails flowing in parafoveal capillaries in all subjects, as previously reported. ²⁴ All dark tails were observed as black “tadpole tail”-like regions that were darker than the vessel shadow, following closely after the bright moving objects that were previously reported as leukocytes or plasma gaps (Fig. 4 and Supplementary Movie S1). ²⁴ Although dark tails were detected in parafoveal capillaries, they were not detected in larger vessels such as the terminal artery or collecting venules. Moreover, dark tails were not detected in all of the capillaries, and they appeared to flow in a fixed path. They suddenly appeared at a branch of the parafoveal capillary, flowed in the capillary network, and disappeared when they reached the larger vessels. 

Figure 4

 

Dark tail elongation. The images show five consecutive frames with a bright particle (dotted red circles) and dark tail (yellow, two-headed arrows) flowing closely behind the bright particle in the parafoveal capillaries, which might correspond to leukocytes and erythrocyte aggregates, respectively. Scale bar: 100 μm. As the bright particles moved forward, the dark tails followed closely. The lengths of the dark tails became longer in a time-dependent manner, as shown by the yellow arrows.

Dark Tail Velocity

The average velocity of the dark tails of normal subjects was 1.49 ± 0.36 mm/s, with a range of 0.79 to 2.19 mm/s (Table). 

Table

 

Characteristics of Subjects and Dark Tails

Subject	Sex	Age	Blood Pressure, mm Hg	Blood Data		Eye	Axial Length, mm	Average Velocity, mm/s	Average Dark Tail Length, μm			Average Dark Tail Elongation Rate	Average Vessel Diameter, μm
				RBC, ×104/μL	Hb, g/dL				Zone a	Zone b	Zone c
A	M	35	100/61	498	15.2	OD	26.2	1.14 ± 0.25	93.7 ± 53.5	119.7 ± 51.3	144.6 ± 42.4	0.54 ± 0.31	9.7 ± 0.3
B	F	42	98/61	415	12.7	OD	24.5	1.62 ± 0.30	61.9 ± 22.1	77.5 ± 20.0	95.3 ± 15.8	0.34 ± 0.27	8.7 ± 1.2
C	F	31	104/70	454	14.3	OD	25.1	1.48 ± 0.25	90.5 ± 20.9	144.0 ± 40.8	174.4 ± 56.5	0.54 ± 0.26	9.1 ± 0.9
D	F	23	103/64	428	13.5	OS	23.6	1.48 ± 0.28	75.7 ± 41.0	108.1 ± 42.0	135.3 ± 45.9	0.54 ± 0.22	13.0 ± 4.5
E	F	34	101/70	436	13.5	OD	24.7	1.22 ± 0.33	52.9 ± 35.2	98.7 ± 40.1	149.1 ± 47.5	0.55 ± 0.25	10.2 ± 0.4
F	M	38	107/71	500	15.5	OS	23.8	1.77 ± 0.54	86.7 ± 18.2	113.2 ± 34.2	142.1 ± 33.4	0.39 ± 0.23	10.4 ± 1.0
G	M	33	135/78	500	13.8	OD	26.3	1.58 ± 0.32	109.7 ± 35.3	150.1 ± 51.3	180.5 ± 63.5	0.74 ± 0.72	12.0 ± 0.9
H	M	40	110/64	479	14.5	OD	24.9	1.60 ± 0.34	84.1 ± 24.0	107.4 ± 23.4	135.4 ± 45.2	0.59 ± 0.56	12.2 ± 2.5
I	M	28	120/74	508	14.3	OD	27.1	1.53 ± 0.35	83.3 ± 36.8	121.3 ± 26.6	151.5 ± 32.9	0.48 ± 0.30	12.4 ± 2.0
J	F	22	102/46	522	15.7	OD	25.9	1.51 ± 0.21	73.1 ± 28.9	88.7 ± 29.1	114.2 ± 39.8	0.37 ± 0.17	11.8 ± 1.1
Mean ± SD		32.6 ± 6.7					25.2 ± 1.1	1.49 ± 0.36	81.2 ± 35.1	112.9 ± 41.4	142.2 ± 48.2	0.51 ± 0.37	11.0 ± 2.3

 RBC, red blood cells; Hb, hemoglobin; OD, right eye; OS, left eye.

Dark Tail Length

The lengths of the dark tails were calculated successfully in all subjects by using spatiotemporal images acquired from target vessels (Fig. 5). Examples of spatiotemporal images are shown in Figure 6. All examples show several dark bands that corresponded to the trajectories of the dark tails and were narrowest upstream and broadest downstream, indicating gradual elongation of the dark tails. The overall average length of the dark tails was 112.1 ± 36.9 μm, with a range of 21.1 to 302.2 μm. The average lengths of the dark tails at zones a, b, and c were 81.2 ± 35.1, 112.9 ± 41.4, and 142.2 ± 48.2 μm, respectively (Table). The length of the dark tails became longer in a time-dependent manner (P < 0.0001) (Figs. 4, 7A). The lengths of all dark tails were longer at zone c than at zone a. The overall average elongation rate was 0.51 ± 0.37, ranging from 0.02 to 1.87. Although the velocity of the dark tail was considerably slower when the length of the dark tail was longer, the elongation rate was not correlated with the velocity (P = 0.56, r = −0.062). The reproducibility of the dark tail elongation measurements was calculated by using ICC, and the obtained ICC value was 0.954. 

Figure 5

 

Dark tail analysis. (A) Color fundus photograph. (B) The AO-SLO image corresponds to the outlined area in (A). Scale bar: 200 μm. (C) The first frame of the AO-SLO video recorded in the outlined area in (B). Scale bar: 50 μm. The numerous white dots represent cone photoreceptors, and the dark lines represent the shadows of the parafoveal capillaries on the photoreceptors. (D) Constructed image of the capillaries. (E) The target vessel with dark tail flow was selected from one branch to another to ensure that they were free of another bifurcation (blue line). (F) A spatiotemporal image generated from the AO-SLO video by reslicing the frames along the line set on the target vessel (blue line in [E]) showing a white band paired with a dark band, which correspond to the trajectories of bright moving objects and dark tails, respectively. (G) Plot profile of the yellow line set on (F). The frame number of the gray value range that was lower than the averaged gray value (t) was measured as the time required for the dark tail to pass the point on the target vessel. The grayscale ranged from 0 (black) to 255 (white).

Figure 6

 

Examples of spatiotemporal images with dark tails. The letters for each image represent the subjects described in the Table. All examples show several dark bands that correspond to the trajectories of dark tails and are narrowest upstream and broadest downstream, indicating gradual elongation of dark tails.

Figure 7

 

Differences in dark tail length and capillary diameter among three zones. Target vessels were separated into three zones from upstream to downstream of blood flow (zones a, b, and c) in order to evaluate the time-dependent change in dark tail length. (A) Significant differences in the dark tails were found among zones a, b, and c. The dark tails became longer in a time-dependent manner (P < 0.0001). (B) The diameters of the target vessels were not significantly different among the three zones (P = 0.43), suggesting that dark tail elongation is not caused by the gradual narrowing of the capillary lumen from upstream to downstream of blood flow.

Capillary Diameter

The mean capillary diameter in the parafovea was 11.0 ± 2.3 μm, with a range of 7.5 to 19.1. The diameters of the zones were not significantly different (P = 0.43) (Fig. 7B). A significant correlation was not found between diameter and velocity (P = 0.31, r = 0.109), between diameter and elongation rate (P = 0.80, r = 0.027), or between diameter and the length of the dark tail (P = 0.15, r = −0.153), suggesting that dark tail elongation was not caused by the gradual narrowing of the capillary lumen from upstream to downstream of blood flow. 

Discussion

In this study, we applied the AO-SLO system to monitor erythrocyte aggregates in healthy human retinas and demonstrated the phenomenon of erythrocyte aggregate elongation in the parafoveal capillaries for the first time. Careful observation of AO-SLO images revealed the pairing of a bright particle and a dark region flowing in the parafoveal capillaries (Fig. 4 and Supplementary Movie S1), which correspond to leukocytes and erythrocyte aggregates, respectively, as previously reported. ²⁴ Their velocities and changes in length were clearly depicted in the spatiotemporal images; the dark bands, which were represented geometrically in the spatiotemporal images and corresponded to the dark tail trajectories, had a triangular or trapezoidal shape. Dark tail elongation was demonstrated by analyzing these shapes in all dark bands. 

All of the dark tails showed elongation, but not shortening. Because all samples were recorded in healthy young subjects, elongation may be a physiological phenomenon that is observed normally in the human retina. Unfortunately, the mechanism of this phenomenon remains unknown because we could not finely visualize the elongating dark tail at a cellular level in the current study. However, we believe that dark tail elongation was caused by the packed erythrocytes that could not overtake the slow-moving leukocytes in the capillary lumen, thereby strongly blocking the imaging light of AO-SLO. ²⁹ To eliminate the influence of vessel diameter difference on dark tail elongation, that is, influence on the moving object in the tapering lumen, which would be stretched and elongated, we evaluated the diameter uniformity of target vessels by dividing the vessels into thirds and measuring the diameter of each zone. The average diameter of the target vessels was 11.0 ± 2.3 μm, and significant differences were not found among the diameters of each zone (P = 0.43). Moreover, a significant association was not observed between the average diameter of the target vessel and the elongation rate (P = 0.80). These findings support the notion that dark tail elongation was not caused by the tapering lumen of vessels during the flow and that dark tail elongation would be induced by blood cell kinetics. Leukocytes might interfere with the flow of erythrocytes and induce packing of sequential erythrocytes, resulting in the detection of elongating dark tails closely behind the bright moving objects known as leukocytes in AO-SLO videos. ²⁴  

It is widely accepted that erythrocytes play an extremely important role in determining the flow properties of blood. ⁴⁴ Changes in the condition of erythrocytes cause impairment in microcirculation. To clarify microcirculation in diseases, erythrocyte aggregation has been used as a surrogate marker through analysis of the collected blood samples of patients. For example, increased erythrocyte aggregations were reported to be cardiovascular risk factors in patients with diabetes mellitus. ⁴⁵ Erythrocyte aggregation was also increased in patients with SLE, causing a decreased flow that might contribute to the thromboembolic process. ⁵ In patients with Behçet's disease, increased erythrocyte aggregates are related to increased fibrinogen, but not to thrombosis and uveitis. ³² These findings motivated us to measure the dark tail velocity and the elongation of dark tail length to analyze the dynamics of erythrocyte aggregates in parafovea capillaries. Although a few studies have reported direct and noninvasive observation of erythrocyte aggregation using conjunctival blood vessels, ³³ to the best of our knowledge, this is the first study to report noninvasive and direct monitoring of erythrocyte aggregates in human retinal capillaries and the phenomenon of erythrocyte aggregate elongation. In the future, the measurement of dark tail length or elongation rate will be tested in various diseases using methods that are similar to those described in this study, and their clinical importance will be investigated. 

The peculiarity of AO-SLO is the correction of ocular lower- and higher-order aberrations, enabling noninvasive observation of each retinal layer, including the photoreceptor layer,²⁴ nerve fiber layer,⁴⁶ and capillary layer,³⁴ under high resolution. Although the aim of this study was to analyze blood cells, the scanning layer of AO-SLO was focused on the photoreceptor layer rather than the capillary layer, and blood cells were observed as shadings on the shadows of the bright cone mosaic patterns of photoreceptors. As previously reported, the reflected light characteristics of the AO-SLO laser from photoreceptors is affected by blood cells because of their differences in scattering coefficients.^24,47 When the scanning layer is focused on the photoreceptor layer, leukocytes are candidates for the bright particles moving in the dark vessel shadows due to the low absorptivity of the AO-SLO laser, resulting in detection of leukocytes as background illumination of photoreceptors. Blood plasma is another bright particle candidate, but its signal is slightly weaker than that of leukocytes.^24,41 Conversely, erythrocytes are strong candidates for the region that is darker than the vessel shadow because they block the AO-SLO lasers, and their aggregation is thought to be a dark tail. Because these differences in the shades on the photoreceptor layer help to distinguish blood components in the retinal microvasculature, we believe that the best focus for monitoring erythrocyte aggregation is the photoreceptor layer (Supplementary Movie S2). Meanwhile, only the numerous high-intensity particles that may correspond to a mixture of reflected light from erythrocytes and from the retinal layers near the capillary layer can be observed when the scanning layer is focused on the capillary layer, and a distinction between blood components is impossible with the accuracy of the current confocal optic system (Supplementary Movie S3). 

Detection of a pair of bright particles and a dark tail flowing in the retinal circulation using AO-SLO bears a remarkable resemblance to the perception of a bright moving object and its dark tail under blue field entoptic phenomenon; this type of physiological phenomenon is perceived as numerous bright particles that move in a flowing manner with a synchronous rhythmic acceleration that corresponds to the cardiac cycle against bright, diffuse illumination. ¹⁵ Sinclair et al. suggested that leukocytes are the source of the bright particles perceived under the blue field entoptic phenomenon and that erythrocytes are detected as dark particles using animal preparations; that is, the entoptic images are the perception of the gaps between erythrocytes created by leukocytes, which have large cell bodies. ¹⁵ Erythrocyte aggregates have also been observed in retinal circulation in studies that used scanning laser ophthalmoscopy with fluorescein angiography. These studies suggested that the dark (hypofluorescent) spots detected in the stained capillaries were rouleaux formations of erythrocytes and represented erythrocyte aggregates. The fluorescein characteristics showed that the hemoglobin and oxyhemoglobin absorption bands were clearly visible in the fluorescence spectra and accounted for hypofluorescent erythrocytes. ^14,48,49 Unlike the above-mentioned techniques for observing erythrocyte aggregates, an approach by AO-SLO enables the analysis of erythrocyte aggregates quantitatively and without dying agents. Herein, the dark tail appearance was regarded as erythrocyte aggregates without a direct comparison of the blood components between animal preparations and AO-SLO imaging, which is a potential limitation of the current study. However, the articles described above let us speculate that the dark tail is correlated strongly with erythrocyte aggregates and that monitoring real-time changes in dark tails by AO-SLO would help us to understand the circulation physiology of the retina. 

Dark tails could not be seen in vessels larger than capillaries in the current study. These large vessels were observed as dark shadows, and moving objects could not be detected in them when the scanning layer was focused on the photoreceptor layer. This can be explained by the occupation of numerous blood cells, particularly erythrocytes, in the larger vascular lumen where blood cells can easily pass through each other, preventing the imaging light of the AO-SLO from reaching the photoreceptor layer. Meanwhile, blood cells flow in single file in the capillary lumen, resulting in the detection of transparent leukocytes as bright particles and hyperreflective erythrocyte aggregates as dark shadows. ²⁴ Moreover, dark tails could not be detected in all capillaries, and they appeared to flow in a fixed path. A possible reason for this preference is that the leukocytes, which block the current of erythrocytes through the capillary, preferred this flow path, as previously reported. ⁴² However, our recording time of AO-SLO was 4 seconds per one video, which is presumably too short to evaluate the flow path preference. Long-duration recording in addition to analysis of day-to-day or circadian variation of dark tail flow is required in a future study. 

Given that erythrocytes cannot overtake leukocytes in the capillary lumen, it is convenient to think that the velocity of the head of the dark tail is an approximation of leukocyte velocity. In fact, the average velocity of dark tails in this study was 1.49 ± 0.36 mm/s, which is similar to the previously reported leukocyte velocities of 1.37 ²² and 1.30 mm/s ³⁵ obtained using AO-SLO. Tam et al., who first introduced spatiotemporal image analysis as a leukocyte velocity measurement, reported a faster velocity of 1.80 mm/s. ⁴² Although we also used spatiotemporal images to calculate the velocity of moving objects in capillaries, the velocity reported by Tam et al. was adjusted by the correction of error due to raster scan, and pulsatility was evaluated in order to calculate an accurate mean velocity, which is different from the method used herein. To improve the accuracy of velocity measurements in this study, the relationship between the direction of moving objects and the direction of the raster scan should be considered. The spatiotemporal images in this study lack slope modification to reduce the error associated with the raster scan. Although the blood dynamics in the retinal circulation require evaluations of pulsatility and raster scan error, we believe that excluding this parameter had a minimal effect on our results because our monitoring time was very short. Furthermore, our AO-SLO system can perform raster scans at 64 frames/s, which contributed to a smaller influence of raster scan error in this study. The mean error rate was calculated at 6.02% (range, 3.27%–9.11%) with reference to the formula of percent error reported by Tam et al. ⁵⁰  

In conclusion, the use of AO-SLO provided a direct, noninvasive, and objective method of monitoring erythrocyte aggregates in the parafoveal capillaries of normal subjects. Erythrocyte aggregates were observed as moving dark shadows on the photoreceptor layer and became elongated in a time-dependent manner. 

Supplementary Materials

pdf S4 

mov S1 

mov S2 

mov S3 

Acknowledgments

Supported in part by the Innovative Techno-Hub for Integrated Medical Bio-imaging Project of the Special Coordination Funds for Promoting Science and Technology from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. 

Disclosure: S. Arichika, None; A. Uji, None; M. Hangai, Canon (F, C, R); S. Ooto, None; N. Yoshimura, Canon (F, C, R)