Morphometry of the normal retinal periarteral capillary-free zone and changes during severe nonproliferative diabetic retinopathy

Abstract

PURPOSE:

To explore the normal morphological features of the retinal periarteral capillary free zone (paCFZ) and the changes associated with severe non-proliferative diabetic retinopathy (NPDR) by optical coherence tomography angiography (OCTA).

METHODS:

A total of 148 normal subjects and 115 patients with severe NPDR were enrolled. Spectral-domain OCTA was used to acquire the image with a Cirrus prototype. Diameter and area of each paCFZ quadrant and adjacent large artery diameter was estimated.

RESULTS:

In healthy subjects, the maximum width of paCFZ in the temporal quadrant (169.38 ± 19.26 μm) was smaller than the other three quadrants (all P <0.001). The ratio of maximum paCFZ width to artery caliber was much larger in the nasal quadrants than the rest quadrants (Ps <0.05). In patients with severe NPDR, both maximum width and area of paCFZ were significant larger, and the retinal artery inner diameters significant smaller in each quadrant compared to normal subjects (Ps <0.05). The ratio of paCFZ maximum width to artery caliber maximum width was significant greater, and the ratio of paCFZ area to artery caliber was significantly higher in all quadrants compared to normal eyes (Ps <0.05).

CONCLUSIONS:

OCTA provides noninvasive and quantitative measurement of paCFZ dimensions. The maximum width and area of paCFZ quadrants and the ratios of these parameters to adjacent inner artery width and area are elevated in severe NPDR, suggesting that changes in paCFZ dimension can be used as clinical indices for diseases associated with changes in retinal microcirculation and oxygen pressure.

PRÉCIS:

Optical coherence tomography angiography revealed differences in retinal periarteral capillary free zone (paCFZ) morphometry between health retina and severe non-proliferative diabetic retinopathy. Morphometric indices of paCFZ may be useful for monitoring disease occurrence and progression.

Keywords

Non-proliferative diabetic retinopathy periarteral capillary free zone optical coherence tomography angiography

1 Introduction

The periarterial capillary-free zone (paCFZ) was first demonstrated by His [1] and is thought to be the structural expression of transmural oxygenation from the major retinal arteries down to the level of the smaller arterioles [2-6]. During retinal development, the oxygen pressure within arteries and capillaries establishes the paCFZ. The paCFZ is visible along retinal arteries but not along retinal veins, and an additional capillary free zone is located around the optic nerve head [7]. The presence of the paCFZ is compelling anatomical evidence that oxygen diffuses directly into the mammalian inner retina from pre-capillary vessels [8]. Previous studies [9, 10] have shown that the dimensions of the paCFZ are associated with retinal oxygen level, suggesting that various morphometric parameters of the paCFZ may provide an objective means for monitoring retinal oxygen saturation and oxygen delivery. Moreover, early and subtle changes in the paCFZ may indicate the first signs of retinopathies associated with aberrant retinal perfusion [11]. Thus, an effective and noninvasive method for monitoring paCFZ would provide a means for early detection of changes related to various ischemic or hyper-perfusion pathologies as well as a more thorough understanding of the underlying pathophysiology.

Examination of paCFZ by fundus fluorescein angiography (FFA) has several limitations, including low resolution and invasiveness [12, 13]. This low resolution may be particularly problematic in cases of retinal vascular leakage and vascular inflammation. In addition, FFA cannot separately visualize the intraretinal structures of the major capillary networks, as images of the superficial capillaries and deep capillaries overlap in two dimensions. Optical coherence tomography angiography (OCT-A) is a newer noninvasive functional extension of OCT for visualizing retinal and choroidal microvasculature. Unlike the two-dimensional views provided by FFA, OCT-A permits topographic isolation of vascular flow across the retinal layer. Therefore, OCT-A may be useful for morphometric analysis of the paCFZ. To explore clinical application of OCT-A for paCFZ analysis, we established normal dimensional values of the paCFZ by OCT-A for a large cohort of normal eyes, then compared these values to those measured in severe NPDR, one of the most common retinal pathologies associated with low oxygen partial pressure.

2 Methods

2.1 Subjects

Healthy participants and patients with severe NPDR were enrolled from Zhongshan Ophthalmic Center, Sun Yat-Sen University, between September 2016 to July 2017. Severe NPDR was diagnosed and graded according to the Early Treatment of Diabetic Retinopathy Study (ETDRS) classification [14]. In accordance with the tenets of the Declaration of Helsinki, the study was reviewed and approved by the institutional ethics committee of Zhongshan Ophthalmic Center, Sun Yat-Sen University. Written informed consent was collected from all study subjects prior to investigation-related procedures. No patient had a history of ocular surgery or eye trauma. For healthy participants, only one eye was arbitrary selected. The inclusion criteria were a best corrected visual acuity (BCVA) of 20/20 or better, and refraction less than or equal to -3D. The exclusion criteria were a family history of glaucoma in a first-degree relative, signs of myopic degeneration or a pathological form of myopia, and other ophthalmic diseases or the presence of any systemic disease that may affect retinal blood flow, such as diabetic retinopathy, hyperlipidemia nephropathy, or hypertension. For severe NPDR patients, only first diagnosed was eligible, and either eye was arbitrary selected.Patients with previous treatments such as laser photocoagulation, intravitreal steroids, or systemic or intravitreal anti-VEGF agents were excluded. Poor OCT-A images due to eye movements or media opacities were also excluded according to previous criteria [15].

2.2 OCT angiography data acquisition and processing

All subjects underwent OCT-A using an AngioPlex OCT-A system (Zeiss, Inc., Dublin, CA, USA) run by a single operator. The device has a high acquisition speed of 70,000 A-scans per second. The scan area was 6 × 6 mm for all subjects. Analyses were performed on OCT-A images generated from the superficial and deep retinal vascular plexuses.

Two independent graders reviewed the images. To quantify paCFZ parameters and retinal artery inner caliber, the optic OCT angiograms were exported and processed by Image Pro Plus 6.0 (Wayne Rasband, National Institutes of Health, USA). The optic disc boundary for each image was manually delineated along the choroid capillary layer image. The boundaries were then extrapolated to the superficial vascular layer OCT angiogram map (Fig. 1A) for disc region segmentation according to a previous study [13].

Fig.1

The schema for selection of the region of interest. A: Optic disc boundary for each image was manually delineated along the choroid capillary layer image. B: The region of interest around half to one disc diameter from the optic disc margin. C: The width of artery caliber and maximum width of paCFZ. D: The area of paCFZ.

For each peripapillary quadrant, the largest artery diameter was marked. Two independent graders measured the width of the marked arteries in a concentric circle half to one disc diameter from the optic disc margin (Fig. 1B) [16]. The artery caliber and paCFZ widths in the concentric circle were outlined manually (Fig. 1C) and calculated the area (Fig. 1D) by Image Pro Plus 6.0. The grader then assessed the validity of each measurement by evaluating the correlation between the average and central widths.

2.3 Repeatability and reproducibility

Intra-observer repeatability and inter-observer reproducibility for artery inner caliber, maximum width, and area of the paCFZ were evaluated on all images by two graders who measured the same scan from each eye on two different days. Intra-class correlation (ICC) and Bland-Altman plots were used to assess repeatability and reproducibility.

2.4 Statistical analysis

All analyses were performed using SPSS for Windows, version 22.0 (SPSS, Inc., Chicago, IL, USA). Student’s t-tests were used to compare artery inner caliber, maximum width, and area of the paCFZ for each quadrant between normal subjects and severe NPDR patients. Intraclass correlation was used to assess repeatability and reproducibility (ICC values of 0.81–1.00 indicate good to excellent agreement between repeated measurements; values less than 0.40 indicate poor to fair agreement). Bland–Altman analysis tested for proportional biases between repeated measurements. Pearson’s correlation test was used to determine the strength of the relationship between paCFZ maximum width and retinal artery inner caliber and between paCFZ area and retinal artery area in each quadrant. For all tests, P <0.05 was considered statistically significant.

3 Results

3.1 Demographics

A total of 148 normal eyes and 115 eyes with severe NPDR were included in this study. Among them, 26 cases were bilateral severe NPDR. Pertinent demographic data are shown in Table 1. The healthy and severe NPDR groups did not differ significantly in age or sex distribution.

Table 1
Demographics of healthy and diabetic subjects

Subject characteristics Healthy Severe NPDR P value

No. of subjects, n 148 115 -

No. of eyes, n 148 115 -

Male: female 43 : 42 35 : 32 0.84

Age range, y 54.00 ± 9.04 53.95 ± 11.85 0.75

Refractive error (D) 2.25 ± 0.50 1.75 ± 0.75 0.64

Blood glucose (mg/dl) 85 ± 4 132 ± 7 0.01^*

Systolic blood pressure (mmHg) 124 ± 9 132 ± 8 0.18

Diastolic blood pressure (mmHg) 75 ± 8 83 ± 5 0.09

TG (mmol/L) 1.56 ± 0.26 1.63 ± 0.12 0.55

TC (mmol/L) 5.35 ± 0.26 5.56 ± 0.21 0.31

BMI (kg/m²) 23.38 ± 1.58 23.82 ± 1.67 0.35

Last recorded HbA_1c (%) - 5.8 -

Cr (μmol/L) 84.55 ± 6.46 98.25 ± 5.40 0.10

BUN (mmol/L) 5.41 ± 1.34 7.17 ± 1.96 0.08

IOP (mmHg) 13.51 ± 2.14 15.10 ± 1.58 0.36

Subject characteristics	Healthy	Severe NPDR	P value
No. of subjects, n	148	115	-
No. of eyes, n	148	115	-
Male: female	43 : 42	35 : 32	0.84
Age range, y	54.00 ± 9.04	53.95 ± 11.85	0.75
Refractive error (D)	2.25 ± 0.50	1.75 ± 0.75	0.64
Blood glucose (mg/dl)	85 ± 4	132 ± 7	0.01^*
Systolic blood pressure (mmHg)	124 ± 9	132 ± 8	0.18
Diastolic blood pressure (mmHg)	75 ± 8	83 ± 5	0.09
TG (mmol/L)	1.56 ± 0.26	1.63 ± 0.12	0.55
TC (mmol/L)	5.35 ± 0.26	5.56 ± 0.21	0.31
BMI (kg/m²)	23.38 ± 1.58	23.82 ± 1.67	0.35
Last recorded HbA_1c (%)	-	5.8	-
Cr (μmol/L)	84.55 ± 6.46	98.25 ± 5.40	0.10
BUN (mmol/L)	5.41 ± 1.34	7.17 ± 1.96	0.08
IOP (mmHg)	13.51 ± 2.14	15.10 ± 1.58	0.36

NPDR, non proliferative diabetic retinopathy; y, year; TG, triglyceride; TC, total cholesterol; BMI, body mass index; Cr, creatinine; BUN, blood urea nitrogen; IOP, intraocular pressure. *: P <0.05

3.2 Repeatability and reproducibility

In normal eyes, the superficial retinal arteriolar and venous structures could be clearly distinguished on OCT-A images (Fig. 1B). Artery and paCFZ measures from normal eyes are summarized in Table 2. The ICCs for intra-observer repeatability were 0.88 for artery inner caliber, 0.92 for paCFZ maximum width, and 0.89 for paCFZ area, while the ICCs for inter-observer reproducibility were 081 for paCFZ width, 0.85 for artery inner diameter, and 0.87 for paCFZ area. Bland-Altman plots show the good repeatability and reproducibility for both artery and paCFZ parameters. Inter-observer reproducibility of all measurements was slightly lower than intra-observer repeatability. Bland–Altman plots were showed in Fig. 2 and Fig. 3.

Table 2
Retinal paCFZ and artery calibers measurements in normal subjects

Quadrants Age (years) Width of paCFZ (μm) Area of paCFZ (mm²) Artery caliber (μm) Width of paCFZ/artery caliber ratio Area of paCFZ/artery caliber ratio

Superior temporal 40-49 120.6 ± 24.1 0.087 ± 0.029 105.2 ± 17.6 1.17 ± 0.19 0.79 ± 0.20

50-59 122.5 ± 29.3 0.079 ± 0.026 102.5 ± 14.3 1.19 ± 0.20 0.71 ± 0.15

60-69 116.5 ± 25.4 0.088 ± 0.022 98.3 ± 12.3 1.18 ± 0.15 0.81 ± 0.18

Total 118.13 ± 31.68 0.083 ± 0.020 101.77 ± 12.26 1.19 ± 0.31 0.83 ± 0.23

Superior nasal 40-49 127.6 ± 27.4 0.086 ± 0.020 87.5 ± 9.6 1.50 ± 0.20 1.02 ± 0.31

50-59 129.5 ± 21.1 0.079 ± 0.030 88.3 ± 10.3 1.55 ± 0.21 1.10 ± 0.30

60-69 130.6 ± 29.2 0.085 ± 0.025 86.7 ± 13.4 1.64. ± 0.28 1.07 ± 0.23

Total 129.661 ± 33.65 0.083 ± 0.021 82.78 ± 13.12 1.52 ± 0.39 1.07 ± 0.31

Inferior temporal 40-49 90.8 ± 26.9 0.063 ± 0.018 101.2 ± 17.8 0.97 ± 0.22 0.58 ± 0.10

50-59 95.2 ± 25.1 0.046.3 ± 0.027 101.7 ± 15.3 0.96 ± 0.21 0.70 ± 0.23

60-69 99.8 ± 29.8 0.054 ± 0.039 94.5 ± 16.9 1.05 ± 0.25 0.72 ± 0.25

Total 94.53 ± 29.32 0.056 ± 0.015 100.63 ± 13.94 0.98 ± 0.28 0.64 ± 0.21

Inferior nasal 40-49 136.2 ± 28.2 0.075 ± 0.023 87.4 ± 12.7 1.66 ± 0.34 1.13 ± 0.20

50-59 138.1 ± 20.4 0.074 ± 0.024 77.9 ± 10.9 1.73 ± 0.35 1.10 ± 0.29

60-69 143.2 ± 22.6 0.073 ± 0.022 88.5 ± 15.3 1.82 ± 0.46 1.05 ± 0.28

Total 139.37 ± 32.69 0.074 ± 0.018 81.10 ± 13.37 1.75 ± 0.35 1.09 ± 0.32

Fig.2

The Bland-Altman plot of intra-observer reproducibility.

Fig.3

The Bland-Altman plot of inter-observer repeatability.

3.3 paCFZ measurements

3.3.1 In normal eyes

The mean maximum width of paCFZ did not differ between corresponding quadrants among age groups (Ps >0.05, Table 2 & Table 4). The mean width of the paCFZ 0.5 to 1 disc diameter from the optic disc margin was significantly smaller in the inferior temporal quadrant (94.53 ± 29.32 μm) compared to the other three quadrants (superior temporal: 118.13 ± 31.68 μm, superior nasal: 129.661 ± 33.65 μm, inferior nasal: 139.37 ± 32.69 μm; P <0.05) and combined nasal width was significant larger than the combined temporal width (P <0.05).

Similarly, mean paCFZ area did not differ between corresponding quadrants among age groups (Ps >0.05, Table 2 & Table 4).

The paCFZ area of concentric circles in healthy subjects was significantly larger in the superior hemisphere than the inferior hemisphere (superior temporal: 0.083 ± 0.020 mm² and superior nasal: 0.083 ± 0.021 mm² vs. inferior temporal: 0.056 ± 0.015 mm² and inferior nasal: 0.074 ± 0.018 mm²; P <0.05), and paCFZ area in the inferior nasal quadrant was significant smaller than the other three quadrants (Ps <0.05).

3.3.2 In eyes with severe NPDR

Mean paCFZ width did not differ among quadrant in eyes with severe NPDR (superior temporal: 142.87 ± 35.08 μm, superior nasal: 153.22 ± 36.14 μm, inferior temporal: 143.60 ± 15.40 μm, inferior nasal: 171.88 ± 25.16 μm; Ps >0.05). Similarly, paCFZ area in regions of interest (ROIs) did not differ among quadrants (superior temporal: 0.123 ± 0.026 mm², superior nasal: 0.120 ± 0.024 mm², inferior temporal: 0.113 ± 0.042 mm², inferior nasal: 0.122 ± 0.027 mm²; Ps >0.05). However, all quadrant widths and ROI areas were significantly larger than the corresponding quadrants of normal eyes (Ps <0.05, Table 3 & Table 4).

Table 3
Retinal paCFZ and artery calibers measurements in patients with severe non-proliferative diabetic retinopathy

Quadrants Age (years) Width of paCFZ (μm) Area of paCFZ (mm²) Artery caliber (μm) Width of paCFZ/artery caliber ratio Area of paCFZ/artery caliber ratio

Superior temporal 40-49 137.6 ± 19.61 0.127 ± 0.021 85.2 ± 12.7 1.34 ± 0.19 1.27 ± 0.22

50-59 141.42 ± 21.94 0.119 ± 0.022 93.5 ± 16.7 1.42 ± 0.23 1.40 ± 0.25

60-69 144.25 ± 24.56 0.118 ± 0.025 95.3 ± 13.2 1.51 ± 0.31 143 ± 0.18

Sub-total 142.87 ± 35.08 0.123 ± 0.026 90.11 ± 10.23 1.45 ± 0.39 1.38 ± 0.26

Superior nasal 40-49 148.89 ± 24.67 0.136 ± 0.019 67.9 ± 6.9 1.94 ± 0.20 1.72 ± 0.22

50-59 155.54 ± 22.28 0.119 ± 0.028 73.3 ± 11.3 2.09 ± 0.21 1.81 ± 0.27

60-69 158.46 ± 22.59 0.185 ± 0.027 68.7 ± 14.2 2.21 ± 0.26 1.96 ± 0.24

Sub-total 153.22 ± 36.14 0.120 ± 0.024 70.83 ± 11.58 2.08 ± 0.50 1.84 ± 0.36

Inferior temporal 40-49 138.58 ± 19.62 0.113 ± 0.011 89.2 ± 11.8 1.17 ± 0.22 1.48 ± 0.28

50-59 144.70 ± 21.54 0.124 ± 0.023 85.7 ± 13.2 1.19 ± 0.16 1.60 ± 0.38

60-69 146.48 ± 18.92 0.119 ± 0.029 82.5 ± 19.6 2.26 ± 0.28 1.62 ± 0.43

Sub-total 143.60 ± 15.40 0.113 ± 0.042 85.47 ± 13.44 1.22 ± 0.38 1.55 ± 0.65

Inferior nasal 40-49 166.92 ± 18.27 0.125 ± 0.021 77.4 ± 11.4 2.05 ± 0.34 1.93 ± 0.32

50-59 172.15 ± 18.40 0.124 ± 0.019 73.9 ± 9.4 2.25 ± 0.35 2.01 ± 0.41

60-69 173.7 ± 12.58 0.133 ± 0.031 71.5 ± 13.6 2.32 ± 0.46 2.06 ± 0.38

Sub-total 171.88 ± 25.16 0.122 ± 0.027 70.08 ± 11.17 2.25 ± 0.55 2.00 ± 0.54

Quadrants	Age (years)	Width of paCFZ (μm)	Area of paCFZ (mm²)	Artery caliber (μm)	Width of paCFZ/artery caliber ratio	Area of paCFZ/artery caliber ratio
Superior temporal	40-49	137.6 ± 19.61	0.127 ± 0.021	85.2 ± 12.7	1.34 ± 0.19	1.27 ± 0.22
	50-59	141.42 ± 21.94	0.119 ± 0.022	93.5 ± 16.7	1.42 ± 0.23	1.40 ± 0.25
	60-69	144.25 ± 24.56	0.118 ± 0.025	95.3 ± 13.2	1.51 ± 0.31	143 ± 0.18
	Sub-total	142.87 ± 35.08	0.123 ± 0.026	90.11 ± 10.23	1.45 ± 0.39	1.38 ± 0.26
Superior nasal	40-49	148.89 ± 24.67	0.136 ± 0.019	67.9 ± 6.9	1.94 ± 0.20	1.72 ± 0.22
	50-59	155.54 ± 22.28	0.119 ± 0.028	73.3 ± 11.3	2.09 ± 0.21	1.81 ± 0.27
	60-69	158.46 ± 22.59	0.185 ± 0.027	68.7 ± 14.2	2.21 ± 0.26	1.96 ± 0.24
	Sub-total	153.22 ± 36.14	0.120 ± 0.024	70.83 ± 11.58	2.08 ± 0.50	1.84 ± 0.36
Inferior temporal	40-49	138.58 ± 19.62	0.113 ± 0.011	89.2 ± 11.8	1.17 ± 0.22	1.48 ± 0.28
	50-59	144.70 ± 21.54	0.124 ± 0.023	85.7 ± 13.2	1.19 ± 0.16	1.60 ± 0.38
	60-69	146.48 ± 18.92	0.119 ± 0.029	82.5 ± 19.6	2.26 ± 0.28	1.62 ± 0.43
	Sub-total	143.60 ± 15.40	0.113 ± 0.042	85.47 ± 13.44	1.22 ± 0.38	1.55 ± 0.65
Inferior nasal	40-49	166.92 ± 18.27	0.125 ± 0.021	77.4 ± 11.4	2.05 ± 0.34	1.93 ± 0.32
	50-59	172.15 ± 18.40	0.124 ± 0.019	73.9 ± 9.4	2.25 ± 0.35	2.01 ± 0.41
	60-69	173.7 ± 12.58	0.133 ± 0.031	71.5 ± 13.6	2.32 ± 0.46	2.06 ± 0.38
	Sub-total	171.88 ± 25.16	0.122 ± 0.027	70.08 ± 11.17	2.25 ± 0.55	2.00 ± 0.54

Table 4

Average retinal artery calibers, paCFZ around artery diameters, artery area and paCFZ area in patients with severe non-proliferative diabetic retinopathy

Quadrants	Group	Width of paCFZ (μm)	Area of paCFZ (mm²)	Artery caliber (μm)	Width of paCFZ/artery caliber	Area of paCFZ/artery caliber
Superior temporal	Normal	182.11 ± 21.46	0.082 ± 0.017	102.35 ± 14.71	1.19 ± 0.31	0.81 ± 0.19
	NPDR	209.69 ± 21.41	0.123 ± 0.026	90.11 ± 10.23	1.45 ± 0.39	1.38 ± 0.26
	P value	<0.001	0.001	0.030	<0.001	<0.001
Superior nasal	Normal	173.71 ± 19.14	0.083 ± 0.020	86.90 ± 10.52	1.52 ± 0.39	1.04 ± 0.26
	NPDR	206.08 ± 24.37	0.120 ± 0.024	70.83 ± 11.58	2.08 ± 0.50	1.84 ± 0.36
	P value	<0.001	0.004	<0.001	<0.001	<0.001
Inferior temporal	Normal	169.38 ± 19.26	0.054 ± 0.016	98.96 ± 16.74	0.98 ± 0.28	0.61 ± 0.23
	NPDR	207.37 ± 28.69	0.113 ± 0.042	85.47 ± 13.44	1.22 ± 0.38	1.55 ± 0.65
	P value	<0.001	<0.001	0.002	<0.001	<0.001
Inferior nasal	Normal	181.46 ± 18.37	0.084 ± 0.021	85.59 ± 13.47	1.75 ± 0.35	1.08 ± 0.32
	NPDR	215.08 ± 25.55	0.122 ± 0.027	70.08 ± 11.17	2.25 ± 0.55	2.00 ± 0.54
	P value	<0.001	<0.001	<0.001	<0.001	<0.001

3.4 Artery inner caliber measurement

3.4.1 In normal eyes

There were no significant differences in mean retinal artery inner diameter between corresponding quadrants among age groups (Ps >0.05, Table 2 & Table 4). Across all age groups, mean retinal artery inner diameter was significantly larger in the temporal hemisphere than the nasal hemisphere (101.77±12.26 μm in superior temporal and 100.63±13.94 μm in inferior temporal vs. 82.78±13.12 μm in superior nasal and 81.10±13.37 μm in inferior nasal quadrants; P = 0.00, Table 2).

3.4.2 In eyes with severe NPDR

In patients with severe NPDR, mean artery inner diameters were significantly smaller than in normal eyes for all quadrants (superior temporal: 90.11±10.23 μm, superior nasal: 70.83±11.58 μm, inferior temporal: 85.47±13.44 μm, inferior nasal: 70.08±11.17 μm; Ps <0.05, Table 3 & Table 4).

3.5 Ratio of paCFZ to the adjacent artery

3.5.1 In normal eyes

The ratio of CFZ width to artery inner diameter did not differ between corresponding quadrants among age groups (Ps >0.05, Table 2 & Table 4). The overall paCFZ width to artery inner diameter ratio of the temporal hemisphere was significantly smaller than that of the nasal hemisphere (superior temporal: 1.19±0.31 and inferior temporal: 0.98±0.28 vs. superior nasal: 1.52±0.39 and inferior nasal: 1.75±0.35; Ps <0.05, Table 2 & Table 4).

The ratio of paCFZ area to artery area was significantly smaller in the inferior temporal quadrant compared to all others (inferior temporal: 0.64±0.21 vs. superior temporal: 0.83±0.23, superior nasal: 1.07±0.31, inferior nasal: 1.09±0.32; Ps<0.05).

3.5.2 In eyes with severe NPDR

In severe NPDR eyes, the ratios of paCFZ width to the counterpart artery inner diameter were significantly larger than in the corresponding quadrants of normal eyes (superior temporal: 1.45±0.39, superior nasal: 2.08±0.50, inferior temporal: 1.22±0.38, inferior nasal: 2.25±0.55; Ps <0.05, Table 3 & Table 4). In addition, the ratios in nasal quadrants were significantly larger than in temporal quadrants of NPDR eyes (Ps <0.05).

The ratios of paCFZ area to counterpart artery area were also significantly larger than in the corresponding quadrants of normal eyes (superior temporal: 1.38±0.26, superior nasal: 1.84±0.36, inferior temporal: 1.55±0.65, inferior nasal: 2.00±0.55; Ps <0.05).

3.6 Correlation between paCFZ and the retinal artery

In healthy subjects, paCFZ width was positively correlated with the counterpart artery caliber (r = 0.378, P<0.001). In contrast, paCFZ area was not significantly correlated with retinal artery caliber and artery area (r = -0.149, P = 0.128).

4 Discussion

Although FFA is a classic tool for revealing retinal microcirculation perfusion, the ability to examine vessel invasion into the paCFZ may be limited during the later disease stages by vascular fluorescence leakage. In contrast, OCT-A provides higher resolution and more stable images, and resolution will not change with disease progression. OCT-A can reveal pathogenic changes and their localization in much greater detail than FFA, including microaneurysms, enlarged foveal avascular zone, area of retinal nonperfusion, reduced capillary density, capillary tortuosity, and dilatation of both superficial or deep choroid capillaries. In addition, OCT-A can also be used for the quantitative analysis of retinal caliber. Finally, as a non-invasive modality, OCT-A is safer than FFA. Despite these advantages, no previous study has examined human paCFZ morphology in vivo. To the best of our knowledge, this is the first study characterizing normal paCFZ morphology in healthy retina and changes associated with NPDR using OCT-A. Results showed that OCTA can be a reliable high-resolution modality for quantitative and qualitative analysis of paCFZ morphometry and retinal vascular caliber for clinical assessment of vascular changes and studies on the underlying pathogenesis.

In this study, the mean maximum width of the paCFZ in normal retina was 139 μm, greater than 120 μm by fluorescein angiography in McLeod study [17]. The reason for this discrepancy is that OCT-A measurements are based on changes in reflection and backscattering of light, and not all capillaries are open during OCT scanning, which requires 1 to 2 s [18, 19]. Thus, some of the capillaries around the CFZ might be temporarily excluded, so vascular circulation halted or reduced during OCT scanning, while fluorescence in the capillary persists for approximately 10 minutes during fluorescein angiography.

No significant differences in paCFZ quadrant parameters were found among age groups. Rose et al [20]. also reported that the oxygen partial pressure distribution in large retinal vessels is not related to age. Similar results in the current study confirm paCFZ morphology as an important indirect index reflecting intravascular oxygen partial pressure. In contrast to this homogeneity among age groups, there were multiple significant differences in maximum width and area among the four retinal quadrants, suggesting that the parameters from each quadrant are important for understanding the characteristics of paCFZ in health and disease.

The paCFZ adjacent to the primary branches of retinal arterioles was narrower than that adjacent to secondary branches, possibly due to structural differences between primary and secondary branches and concomitant effects on oxygen diffusion. Hence, the relationship between paCFZ width and vessel caliber was analyzed. In addition, the ratios of maximum paCFZ width and area to adjacent artery caliber were also analyzed to provide normal physiological reference values as indicators of vessel oxygen partial pressure. The paCFZ was clearly asymmetrical, as there were significant difference in area, maximum width, and ratios to artery caliber among quadrants. The area and ratio of width to artery caliber were significantly smaller in the temporal inferior quadrant than the other three quadrants, consistent with the report of Mase et al [21]. It thought that this asymmetric morphology may reflect the high capillary density and thicker nerve fiber layer around the optic disc as the inferior temporal quadrant is located near the central fovea. The temporal peripapillary nerve fiber layer was significantly thicker and the temporal peripapillary retinal capillary density greater than on the nasal side, resulting in significantly greater CFZ width on the nasal side. This asymmetry in turn suggests that the PO2 distribution in retinal vessels may differ among quadrants. Indeed, Jiang et al [22, 23]. found that the temporal (superior and inferior) arterial diameters were significantly larger than the nasal (superior inferior) arterial diameters, and consistent with our study concluded that the distribution of PO2 in large retinal vessels differs significantly among quadrants.

Given that paCFZ is an important sign of PO2 changes in retinal vessels, while PO2 changes are associated with DR [24], differences in paCFZ parameters among patients with severe NPDR may provide an indirect way to monitor DR progression. Studies have shown that the PO2 in retinal vessels is higher in patients with diabetic retinopathy than in healthy controls due to shunting of blood through preferential channels and the bypass of nonperfused capillaries in the capillary network [25]. Venous PO2 is relatively high in patients with diabetic retinopathy, and the saturation increases with the severity of retinopathy. Further, changes in PO2 can lead to corresponding alterations in paCFZ morphology. Indeed, all paCFZ quadrant widths and areas as well as the ratios of maximum width and area to artery caliber were significantly larger in patients with severe NPDR than in normal subjects. Thus, these changes may be useful indicators of DR occurrence and development.

In this study, the width and area of the paCFZ and the ratio of these measures to adjacent artery caliber were quantitatively analyzed in both healthy retinas to define references values and in NPDR patients to examine the utility of these metrics for monitoring DR progression. The major limitation of this study is the sample size drawn from a single institution. Thus, further studies are required to explore how these paCFZ metrics change during the different stages of DR. In addition, it remains unclear if these changes can be detected in the early stages of DR as a sign of abnormal retinal microcirculation. Furthermore, characteristics of those paCFZ can be used as index in the auto-detection of early diabetic retinopathy by artificial intelligence technique.

Footnotes

Acknowledgments

None.

Conflict of interest

The authors have no financial or other conflicts of interest concerning this study.

Financial support

This work was supported by the National Natural Science Foundation of Guangdong, China (grant number 2016A030313364) and the Science and Technology Program of Guangzhou, China (grant number 2016070010070).

References

His

. Abbildungen uber des Gefaszsystem dermenschlichen Netzhaut und derjenigen des Kaninchens. Arch f Anat u Entwickelungsg (1880).

Wolf

, Arend

, Schulte

, et al., Quantification of retinal capillary density and flow velocity in patients with essential hypertension, Hypertension 23(4) (1994), 464–467.

Friedland

A.B.

, A mathematical model of transmural transport of oxygen to the retina, Bull Math Biol 40(6) (1978), 823–837.

Toussaint

, Kuwabara

, Cogan

D.G.

, Retinal vascular patterns, II. Human retinal vessels studied in three dimensions. Arch Ophthalmol 65 (1961), 575–581.

Pournaras

C.J.

, Rungger-Brandle

, Riva

CE.

, et al. Regulation of retinal blood flow in health and disease. Prog Retin Eye Res 2008;27(3):284–330. doi: 10.1016/j.preteyeres.2008.02.002

Michaelson

I.C.

, Herz

, Lewkowitz

, Kertesz

, Effect of increased oxygen on the development of the retinal vessels; an experimental study, Br J Ophthalmol 38(10) (1954), 577–587.

Claxton

, Fruttiger

, Role of arteries in oxygen induced vaso-obliteration, Exp Eye Res 77(3) (2003), 305–311.

Harrison

W.W.

, Chang

, Cardenas

M.G.

, et al. Blood pressure, vessel caliber, and retinal thickness in diabetes. Optom Vis Sci 2012;89(12):1715–20. doi: 10.1097/OPX.0b013e3182775c62

Fruttiger

, Development of the retinal vasculature, Angiogenesis 2007;10(2):77–88. doi: 10.1007/s10456-007-9065-1

10.

Arend

, Wolf

, Jung

, et al., Retinal microcirculation in patients with diabetes mellitus: dynamic and morphological analysis of perifoveal capillary network, Br J Ophthalmol 75(9) (1991), 514–518.

11.

Mendis

K.R.

, Balaratnasingam

, Yu

, et al. Correlation of histologic and clinical images to determine the diagnostic value of fluorescein angiography for studying retinal capillary detail. Invest Ophthalmol Vis Sci 2010;51(11):5864–9. doi: 10.1167/iovs.10-5333

12.

Patel

R.C.

, Gao

S.S.

, Zhang

, et al. Optical coherence tomography angiography of choroidalL neovascularization in four inherited retinal dystrophies. Retina 2016;36(12):2339–47. doi: 10.1097/iae.0000000000001159

13.

Hodge

J.V.

, Parr

J.C.

, Spears

G.F.

, Comparison of methods of measuring vessel widths on retinal photographs and the effect of fluorescein injection on apparent retinal vessel calibers, Am J Ophthalmol 68(6) (1969), 1060–1068.

14.

Wilkinson

C.P.

, Ferris

F.L.

, Klein

RE.

, et al. Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales. Ophthalmology 2016;110(9):1677–82. doi: 10.1016/s0161-6420(03)00475-5

15.

Matsunaga

, Yi

, Puliafito

C.A.

, Kashani

AH.

. OCT angiography in healthy human subjects. Ophthalmic Surg Lasers Imaging Retina 2014;45(6):510–5. doi: 10.3928/23258160-20141118-04

16.

Wong

T.Y.

, Islam

F.M.

, Klein

, et al. Retinal vascular caliber, cardiovascular risk factors, and inflammation: the multi-ethnic study of atherosclerosis (MESA). Invest Ophthalmol Vis Sci 2006;47(6):2341–50. doi:10.1167/iovs.05-1539

17.

McLeod

, Krogh cylinders in retinal development, panretinal hypoperfusion and diabetic retinopathy, Acta Ophthalmol 2010;88(8):817–35. doi:10.1111/j.1755-3768.2009.01796.x

18.

Krogh

, The number and distribution of capillaries in muscles with calculations of the oxygen pressure head necessary for supplying the tissue, J Physiol 52(6) (1919), 409–415.

19.

Krogh

, The supply of oxygen to the tissues and the regulation of the capillary circulation, J Physiol 52(6) (1919), 457–474.

20.

Rose

, Flanagan

J.G.

, Patel

SR.

, et al. Retinal blood flow and vascular reactivity in chronic smokers. Invest Ophthalmol Vis Sci 2014;55(7):4266–76. doi: 10.1167/iovs.14-14022

21.

Mase

, Ishibazawa

, Nagaoka

, et al. Radial Peripapillary Capillary Network Visualized Using Wide-Field Montage Optical Coherence Tomography Angiography. Invest Ophthalmol Vis Sci 2016;57(9):OCT504–10. doi: 10.1167/iovs.15-18877

22.

da Silva

A.V.

, Gouvea

S.A.

, da Silva

A.P.

, et al. Changes in retinal microvascular diameter in patients with diabetes. Int J Gen Med 2015;8: 267–73. doi: 10.2147/ijgm.s83749

23.

Jiang

, Wei

, Shi

, et al. Altered Macular Microvasculature in Mild Cognitive Impairment and Alzheimer Disease. J Neuroophthalmol 2017. doi: 10.1097/wno.0000000000000580

24.

Luan

, Roberts

, Sniegowski

, et al. Retinal thickness and subnormal retinal oxygenation response in experimental diabetic retinopathy. Invest Ophthalmol Vis Sci 2006;47(1):320–8. doi: 10.1167/iovs.05-0272

25.

Hardarson

S.H.

, Stefansson

. Oxygen saturation in branch retinal vein occlusion. Acta Ophthalmol 2012;90(5):466–70. doi:10.1111/j.1755-3768.2011.02109.x