FP1526-Vitreo Retinal Diseases – IV -Determination and validation of retinal arteriolar and venular calibre measurements using cSLO-HRA.

Determination and validation of retinal arteriolar and venular calibre measurements using cSLO-HRA.

Purpose: To determine and validate retinal vascular caliber measurements (arteriolar and venular) using confocal Scanning Laser Ophthalmoscopy- Heidelberg Retina Angiograph (cSLO- HRA) system

Methods: 600 eyes of 300 individuals with no history of systemic or ocular illness were recruited. Non- mydriatic fundus images were captured using cSLO-HRA system. Infrared Reflectance image, Blue Reflectance image and Blue Peak Blue Autofluorescence images were captured for all individuals. The dimension of the retinal arterioles and venules using inbuilt calipers were measured at 1800µm from centre of optic disc (which corresponds to 2^nd circular zone of an inbuilt ETDRS grid). Internal (lumen) and external dimensions were measured using Blue Reflectance image and Blue Peak Blue Autofluorescence images respectively. Measurements were done by 3 independent observers. Inter and intra observer variations were noted. Appropriate statistical analysis was applied.

Results: Median age of individuals was 29 years with a range of 18 years to 50 years out of which 191 were males and 109 were females. Mean internal and external dimensions for arterioles were 85.1 ± 12.4µm and 105.04 ± 12.05µm respectively. Mean internal and external dimensions for venules were 133.84 ± 16.62µm and 145.42 ± 16.13µm respectively. Wall thickness for arterioles and venules were 19.71 ± 8.04µm and 11.02 ± 5.65µm respectively. Arteriole to venule ratio for internal and external dimensions was 0.66 ± 0.1 and 0.74 ± 0.09 respectively. Wall thickness to lumen ratio for arteriole and venule was 0.26 ± 0.12 and 0.1 ± 0.05 respectively. There was no significant difference between 3^rd, 4^th and 5^th decade for all dimensions. Both inter observer and intra observer variation between 3 observers was >0.95.

Conclusion: This is a new non-invasive methodology for capturing and measuring retinal vascular caliber measurements including the lumen and external diameters using cSLO-HRA system. Use of inbuilt calipers to assess retinal vessels showed a high inter observer and intra observer reliability. This is the first effort to determine the normative database using a simple clinically applicable method. Future studies using this normative database as baseline in varied systemic and ocular pathologies would help in better understanding of relationship between retinal vascular dimensions and diseases under study.

INTRODUCTION 

Ever since the development of the first method for fundoscopy by Hermann von Helmholtz in 1951, there has been a persistent interest in evolving a technique for the assessment of retinal vessel diameters. The retinal vasculature is often regarded as a representation of the systemic microvasculature, providing exemplary information about changes in vascular morphology and function, in a natural and non-invasive manner. Given that retinal, cerebral, and coronary blood vessels share similar anatomy and physiology, retinal blood vessels have routinely been evaluated as a part of clinical ocular investigation, especially in hypertensive and metabolic disorders that lead to systemic small vessel disease (1)(2)(3). The changes seen in the retinal vessels serve as ‘markers’ of systemic disease was described by the Scottish physician Robert Marcus Gunn (1)(4)

In fundoscopy, ratio of diameters of retinal arterioles and venules approximate 2:3. Assessment of this arteriovenous ratio (AVR), the ratio of arteriolar diameter to venule diameter, and the general appearance of retinal vessels are based on the normal light reflex of the retinal vasculature, which is formed by the reflection from the interface between the blood column and the vessel wall (5)(6).   In thin-walled veins, haemoglobin has strong absorbance of green light leading to the characteristic dark appearance (7). Retinal arteries have a more silvery look, indicating that light is reflected before it reaches the blood itself. Although these arteries have a strong muscular wall, they appear smaller than the veins, suggesting that the main site of reflection is close to the bloodstream and not near the outside of the vessel. AVR based on fundus evaluation closely reflects the relationship between the arterial and the venous inner diameter and does not include the outer diameter (i.e., the wall) of the vessel (8). Though clinical evaluation of retinal vessels by ophthalmoscopy involves no cost, it is a subjective test requiring large experience with good diagnostic skill. Hence, it has been shown to have poor interobserver reproducibility.

Recent studies using digital imaging and semi-automated software methods have allowed retinal vascular caliber to be measured more precisely from retinal photographs than from fundus examination.  Series of clinical and epidemiologic studies now show that retinal vascular changes are associated with cardiovascular and metabolic diseases such as hypertension (9) (10), diabetes (11) (12), obesity (13), stroke (14), coronary heart disease (15) and heart failure. Retinal vasculature changes are also seen in local pathologies such as primary open angle glaucoma (POAG) (16), central retina vein occlusion (CRVO) and branch retinal vein occlusion (BRVO) (17). Arterioles and venules are affected differently by different diseases. Arterioles are more affected in hypertension(9) (10) and coronary heart disease(18) whereas venules are more affected in diabetes(19), stroke(20) and inflammatory diseases(21). Sometimes, arterioles and venules may be affected concurrently. As recently as 2013, it has been highlighted that the lack of normative data was limitation towards understanding the vast amount of literature available with regard to retinal vasculature dimensions (22).

A first approach to measure retinal vessel diameters in vivo was introduced by Ruete and Landolt. The measurement principle utilized an optical image plane of the indirect ophthalmoscopy and then placing reference markers of well-defined size into it. The markers could be viewed in focus together with the vessels to gauge the vessel diameter by direct comparison. The first electrooptical system for measuring retinal vessel diameters was presented by Delori(23). In the image plane of an indirect ophthalmoscope, a light scanning system picks up brightness profiles perpendicular to the vessel course. The electronic signals representing these vessel cross sections were then analyzed in order to identify the vessel borders by the means of the half-height algorithm described in detail below. The specific purpose of this device was the measurement of oxygen saturation by use of different absorption profiles of oxy/deoxyhemoglobin including the diameter measurements. Stokoe and Turner (1966) in their study on normal retinal vascular pattern and arteriovenous ratio as a measure of arterial caliber, concluded that it was important to study the normal retinal vascular pattern as a preliminary to serial observations on the vascular changes in patients undergoing treatment for hypertension(24). Subsequently, Parr and Spears (1974) described a formula to calculate Central Retinal Arteriole Equivalent(CRAE), which represented the average arteriolar diameters of the eye(25) (26). They derived at the following formula to calculate CRAE,

W_c = (0.87W²_a + 1.01W²_b − 0.22W_aW_b − 10.73)^1/2

Where W _c was calculated as the trunk arteriole diameter, and included diameter from the smallest (W _a) to largest branches (W _b).

Hubbard (1992) then described a formula to calculate Central Retinal Venous Equivalent (CRVE)(9), which represented the average venular diameters of the eye:

W_c = (0.72W²_a + 0.91W²_b + 450.05)^1/2

Where W _c, W _a, W _b were corresponding venular diameter measurements.

These formulas were further modified by excluding vessels with less than < 25µ in the calculation(9). They also developed protocols for non-mydriatic fundus photography for evaluation of retinal vascular abnormalities. The most recent modification has been by Knudson et al.  who chose only the 6 most prominent arteriole and venule dimensions to assess the variability in retinal vessel measurements at different points in the pulse cycle(27). They concluded that measurements of large retinal venules are generally less variable than measurements of other retinal vessels.

For retinal microvascular imaging, images can be acquired by confocal microscopy in In-vitro conditions. For in vivo image capture several techniques are available including adaptive optics, speckle variance OCT (svOCT) and phase variance OCT or Doppler OCT (DOCT). For retinal macrovascular imaging, images can be acquired either by digitizing images captured using analog cameras or by capturing digital images directly using digital cameras (non-confocal or confocal fundus camera). cSLO camera is a digital confocal fundus camera. After acquiring images, they can be analyzed by various software or manually using caliper based computation. Digital image and semi-automated software based evaluation as an objective method of studying retinal vasculature has good inter-observer correlation. This method however is not cost effective and hence not used routinely in clinical practice.

MATERIALS AND METHODS 

All phakic individuals (n= 300 participants; 600 eyes) between 20 to 50 years of age having normal Intra-Ocular Pressure (IOP) with clear media were included in the study. Retinal imaging was done only after obtaining informed consent. Whereas patients with hypertension, diabetes, chronic smoking/ alcoholism, obesity, Coronary Artery Disease (CAD), Cerebrovascular Accident (CVA), myopia more than -1D, hypermetropia more than +1D and IOP >21 mm of hg were not included in the study

All Individuals included, had visual acuity of LogMAR ≤ 0.20, axial length between 22.00 to 24.50 mm with no significant systemic/ ocular or personal history. It was a cross sectional study conducted at Dr. Rajendra Prasad Centre for Ophthalmic Sciences, All India Institute of Medical Sciences (AIIMS) and was approved by the Institute Ethics Committee for Post Graduate Research, AIIMS (Ref. No.- IECPG- 44/27.11.2015). All general and potentially effect-modifying variables, such as age, gender, occupation, smoking, alcohol consumption, personal history and drug use were documented in individual proforma. In order to ensure data quality, observers had received specific training (initial 20 images) after technical discussion.

In the study, for all enrolled individuals non- mydriatic fundus images were captured using confocal Scanning Laser Ophthalmoscopy – Heidelberg Retina Angiograph. Image for the study was one that was centered on the optic nerve head using a 30 degree field. For all 600 eyes, Infrared Reflectance image, Blue Reflectance (equivalent of Red-free) image and Blue Peak Blue Autofluorescence images were captured (Figure 1). In suboptimal images, sharpness was improved using in built software of cSLO-HRA system. Details of image capture are summarized in Table 1.

Table 1: showing different parameters of images captured using cSLO-HRA system.

S. No.	Parameters
1.	Resolution Mode	High Resolution
2.	Scan Angle	30ﹾ
3.	Size X	1536 pixel (9.4mm)
4.	Size Y	1536 pixel (9.4mm)
5.	Scaling	6.10 µm/ pixel
6.	Camera Model	Spectralis HRA + OCT (S3610-CIFP)

For measurement, a circular zone between half a disc diameter and one disc diameter from the optic disc margin (this region is selected because its vessels are unequivocally arterioles instead of arteries. Furthermore, in this region there was less overlap between the vessels than near or on the optic disc making the measurements more reliable)(9) (28). For uniformity, an inbuilt ETDRS grid was used whose centre was placed at the root of the vessels in optic disc. The three circular grids of ETDRS were at a distance of 700 µm, 1800 µm and 3600 µm from inside out (Figure 2). The second circular grid was used to measure the dimension of the retinal arterioles and venules using inbuilt calipers.

The most prominent arteriole and venular pairs were identified and measured by inbuilt caliper using Infrared Reflectance images as reference (Figure 3). Blue Reflectance (Red-free) images were used to determine internal dimension of both arterioles and veins whereas Blue Peak Blue Autofluorescence images were used to determine their external dimensions.

Measurements were done by single observer for all 600 eyes. For inter-observer variation 3 observers measured 200 eyes (33.33% of 600 eyes) and for intra-observer variation same 3 observers did repeat measurements for 50 eyes (25% of 200 eyes). The values so obtained by each of the observers independently were entered into an excel sheet for later analysis and were represented as vertical simple graphs.

STATISTIC ANALYSIS

Vessel diameter parameters, including wall thickness, luminal diameter, vessel diameter, arteriole -to-venule ratio, and wall-to-lumen ratio, were determined on basis of age and gender. These parameters were compared with the right and left eye using Paired t-tests. Reliability of the measurements within-individuals was determined by calculating standard deviation using the square-root of the mean within- individual variance. ICC was used to determine inter-observer reproducibility and Intra-observer repeatability. We classified ICCs of ≥0.81 as almost perfect reliability, and ICCs of ≥0.61 and <0.8 as substantial reliability. Bland-Altman plot was used to compare measurements done by cSLO-HRA system with that done by Image J 1.51J8 [a Java- based image freeware processing program developed by the National Institute of Health (http://rsb.info.nih.gov/ij)] in a randomly selected sample of 40 eyes (6.67% of 600 eyes). For statistical difference between measurements two-sample t-test, two-sample Wilcoxon rank-sum (Mann-Whitney) test, Analysis of variance (ANOVA) and Kruskal-Wallis equality-of-populations rank test were used. All statistical tests were two- sided at 95% confidence interval (CI), and a P-value < 0.05 was considered statistically significant using the Stata/SE 13.1 software (StataCorp, College Station, TX, USA).

OBSERVATIONS AND RESULTS

Demographics

600 eyes of 300 individuals were studied. Median age of the individuals was 29 years with a range of 18 years to 50 years. 166 individuals (55.34%) were between 18 to 30 years, 102 individuals (34%) were between 31 to 40 years, and 32 individuals (10.66%) were between 41 and 50 years. 191 individuals (63.67%) were male and 109 individuals (36.33%) were female.

Table 2: Table showing normative database from the study

S.NO.	PARAMETERS	MEAN VALUES (µm)	95% CONFIDENCE INTERVAL (µm)
	Arteriolar
1.	Internal dimension- lumen	85.10 ± 12.40	59.58 to 109.18
2.	External dimension	105.04 ± 12.05	80.40 to 128.6
3.	Wall thickness	19.71 ± 8.04	2.92 to 35.08
4.	Wall thickness-to-lumen ratio	0.26 ± 0.12	0.01 to 0.53
	Venular
5.	Internal dimension- lumen	133.84 ± 16.62	99.26 to 165.74
6.	External dimension	145.42 ± 16.13	110.99 to 175.51
7.	Wall thickness	11.02 ± 5.65	0 to 22.03
8.	Wall thickness-to-lumen ratio	0.10 ± 0.05	0 to 0.2
	Arteriolar-to-Venular Ratio (AVR)
9.	AVR- Lumen	0.66 ± 0.1	0.46 to 0.86
10.	AVR- External dimension	0.74 ± 0.09	0.56 to 0.92

Table 2, is showing the data collected (normative) during study of 600 eyes from November 2015 to August 2017. Using appropriate statistic analysis we have evaluated mean values and 95% confidence interval for different parameters pertaining to retinal vasculature. Following were the formulas used in evaluating above parameters.

1. Wall thickness =

2. Arteriolar-to-Venular Ratio (AVR)=

3. Wall thickness-to-Lumen Ratio (WLR) =
4. Table 3: Table showing normative database from the study on the basis of age

S.NO.	PARAMETERS	MEAN VALUES (µm)				p- VALUE
		Category 1 18 to 30 years	Category 2 31 to 40 years	Category 3 41  to 50 years
	Arteriolar
1.	Internal dimension- lumen	86.17 ± 18.23	83.54 ± 17.92	82.64 ± 17.63		0.54
2.	External dimension	106.09 ± 17.45	103.72 ± 17.84	101.75 ± 17.90		0.783
3.	Wall thickness	19.72 ± 14.01	19.92 ± 13.58	18.98 ± 13.06		0.544
4.	Wall thickness-to-lumen ratio	0.26 ± 0.21	0.27 ± 0.20	0.25 ± 0.20		0.292
	Venular
5.	Internal dimension- lumen	133.96 ± 23.35	131.17 ± 23.07	126.1 ± 23.08		0.51
6.	External dimension	145.77 ± 22.11	142.42 ± 22.29	138.46 ± 22.80		0.81
7.	Wall thickness	11.19 ± 10.74	10.42 ± 10.89	12.05 ± 10.52		0.247
8.	Wall thickness-to-lumen ratio	0.10 ± 0.10	0.10 ± 0.10	0.11 ± 0.10		0.247
	Arteriolar-to-Venular Ratio (AVR)
9.	AVR- Lumen	0.66 ± 0.13	0.64 ± 0.12		0.68 ± 0.14	0.381
10.	AVR- External dimension	0.74 ± 0.11	0.74 ± 0.11		0.76 ± 0.11	0.493

 Table 3, is showing the data (normative) according to different age groups. On evaluating there was no significant statistical significance between different age groups.

Table 4, is showing the data (normative) according to gender basis and was evaluated for statistical differences among different parameters of retinal vasculature on the basis of gender.

Table 4: Table showing normative database from the study on the basis of gender

S.NO.	PARAMETERS	MEAN VALUES (µm)		p-VALUE	REMARK
		Male	Female
	Arteriolar
1.	Internal dimension- lumen	84.41 ± 17.93	85.76 ± 18.46	0.13	No significant difference
2.	External dimension	104.04 ± 17.83	106.19 ± 17.38	0.015	Male < Female
3.	Wall thickness	19.4 ± 13.88	20.26 ± 13.54	0.077	No significant difference
4.	Wall thickness-to-lumen ratio	0.27 ± 0.21	0.27 ± 0.20	0.256	No significant difference
	Venular
5.	Internal dimension- lumen	133.23 ± 24.04	130.27 ± 21.91	0.032	Male > Female
6.	External dimension	145.05 ± 23.20	141.70 ± 20.56	0.015	Male > Female
7.	Wall thickness	11.14 ± 11.15	10.80 ± 9.92	0.55	No significant difference
8.	Wall thickness-to-lumen ratio	0.10 ± 0.12	0.10 ± 0.10	0.867	No significant difference
	Arteriolar-to-Venular Ratio (AVR)
9.	AVR- Lumen	0.66 ± 0.17	0.68 ± 0.18	0.003	Male < Female
10.	AVR- External dimension	0.74 ± 0.15	0.77 ± 0.16	<0.0001	Male < Female

Inter Observer Reliability

3 observers calibrated randomly selected 200 eyes for measuring interobserver reliability; following values were calculated after averaging the measurements of internal and external dimensions by each observer. The reliability was evaluated by using ICC.

S. No.	Parameters	Internal Dimensions						External Dimensions
		Arteriole			Venule			Arteriole			Venule
		1st Observer	2nd Observer	3rd Observer	1st Observer	2nd Observer	3rd Observer	1st Observer	2nd Observer	3rd Observer	1st Observer	2nd Observer	3rd Observer
	Size	200	200	200	200	200	200	200	200	200	200	200	200
	Mean	87.28	88.42	88.34	134.53	135.54	134.61	105.43	107.25	106.5	145.76	146.84	146.82
	SD	11.28	10.62	11.13	15.75	15.76	15.97	11.03	10.69	10.57	15.69	15.21	14.94
	p50	86.75	89.06	87.69	132	133.35	132.19	105.06	107.06	106.06	143.62	144.27	145.53
	Minimum	63.62	63.5	62.88	104.88	107.25	102.75	82.25	84	85	118.88	123	121.62
	Maximum	119.38	121	121	201.67	201.5	200	138	137.38	136.62	212.67	213.33	209.5
	Intraclass correlation	0.9891			0.9956			0.9883			0.9950
	95% Confidence Interval	0.9878 to 0.9904			0.9950 to 0.9961			0.9868 to 0.9896			0.9944 to 0.9956

Intra Oberserver Reliabilty

Same 3 observers have done repeat calibration for randomly selected 50 eyes for measuring intraobserver reliability, following values were calculated after averaging the measurements of internal and external dimensions of each observer. The reliability was evaluated by using ICC.

S. No.	Parameters	Internal Dimensions
		Arteriole						Venule
		1st Observer		2nd Observer		3rd Observer		1st Observer		2nd Observer		3rd Observer
		1st reading	2nd reading	1st reading	2nd reading	1st reading	2nd reading	1st reading	2nd reading	1st reading	2nd reading	1st reading	2nd reading
	Size	50	50	50	50	50	50	50	50	50	50	50	50
	Mean	87.3	87.65	87.37	87.52	87.76	86.95	135.11	134..44	135.141	135.02	134.67	134.12
	SD	13.66	13.4	12.78	13.52	13.68	13.29	22.06	21.83	22.05	21.77	21.98	21.71
	p50	84	85.75	85.38	85.38	85.88	85.5	127.71	126.43	127.57	127.5	126	126.57
	Minimum	63.62	63.25	63.5	62.25	62.88	61.5	109.75	108.25	109.25	109.5	108.75	108.38
	Maximum	115.88	116.5	112.88	115.75	117.12	115.12	201.67	201.17	201.5	200.17	200	200.33
	Intraclass correlation	0.9961						0.9988
	95% Confidence Interval	0.9952 to 0.9969						0.9985 to 0.9991

 

S. No.	Parameters	External Dimensions
		Arteriole						Venule
		1st Observer		2nd Observer		3rd Observer		1st Observer		2nd Observer		3rd Observer
		1st reading	2nd reading	1st reading	2nd reading	1st reading	2nd reading	1st reading	2nd reading	1st reading	2nd reading	1st reading	2nd reading
	Size	50	50	50	50	50	50	50	50	50	50	50	50
	Mean	104.62	106.42	106.58	106.67	106.12	106.49	145.61	147.41	147.14	147.83	147.09	147.3
	SD	13.08	12.65	12.41	12.5	12.37	12.78	22.06	21.24	21.55	21.07	20.54	21.55
	p50	104	107	107.25	106.38	105.12	106.62	138.71	140.43	141.57	143.12	139.88	140.43
	Minimum	82.25	84	84	85.62	85	84.38	119.75	122.38	123.12	122.25	121.62	122
	Maximum	130	132.12	132.62	131.5	131.88	132	212.67	213.33	213.33	211.5	209.5	212.67
	Intraclass correlation	0.9956						0.9988
	95% Confidence Interval	0.9945 to 0.9965						0.9985 to 0.9990

Comparison with Image J

After determining inter- observer and intra-observer reproducibility, single observer measured vessel caliber for randomly selected 40 eyes using Image J software (freely made available by National Institute Health at the following link http://rsb.info.nih.gov/ij) and compared the values with that done using cSLO-HRA system after averaging the measurements of internal and external dimensions. The two methodologies were compared using Bland-Altman Plot.

S. No.	Parameters	Internal Dimensions				External Dimensions
		Arteriole		Venule		Arteriole		Venule
		Study Data	Image J	Study Data	Image J	Study Data	Image J	Study Data	Image J
	Size	40	40	40	40	40	40	40	40
	Mean	87.28	89.46	105.76	107.68	133.14	134.44	145.89	146.79
	SD	13.13	13.19	11.22	11.18	13.91	13.09	13.64	13.29
	p50	87.06	89.31	103.83	105.85	130.38	132.06	143.5	143.69
	Minimum	65.12	66.75	86.12	88	109.75	111.12	119.75	122
	Maximum	110.75	113.88	125.12	127.75	159.62	158.75	171.75	171.88

DISCUSSION 

Retinal vasculature provides a unique non-invasive window to assess vascular health and this has been emphasized since early days of ophthalmoscopy. Both clinicians and researchers have been interested in developing methods not just to visualize fundus, but also determine the exact retinal vessel size. Recent advances in retinal image analysis have provided exciting possibilities, but the applicability of the reported methods in the clinical setting is yet to be established. Use of AVR as a biomarker for ocular and systemic disease has always been debatable largely because the measurements have been largely subjective and qualitative in nature. If retinal vascular measurements are to truly become markers for systemic and ocular diseases it is important to have quantitative methods that are not only precise but cost effective are user friendly.

Initially, methods based on empirical formulas were derived for individual retinal vascular calibers. Parr-Hubbard and his team examined a large number of retinal images using a root mean square deviation model to measure CRAE and CRVE(9) (25) (26). More recently, Knudtson et al. (29) modified previously used Parr-Hubbard formulas for retinal vascular caliber. As these are the summary measures, further refinements are required to reflect true values of the retinal vascular calibers.

Though there is a vast amount of data on retinal vascular measurement from population based studies, availability of normative data is lacking (29). To develop a clinical marker, it is important to have a normal baseline so that we could clearly define normal, abnormal and variant of normal. One of the challenges in deriving normative data using studies in the adult populations has been the difficulty to completely control for the confounding effect of systemic (e.g., hypertension, diabetes, smoking, or medication) processes on retinal vascular caliber measurements. It has been alluded to earlier that studies in healthy children and young adults, who are generally free of these influences, may provide better understanding of important retinal vascular variables.

In the present study, we propose a new non- invasive method for measuring the caliber of the retinal blood vessels in an undilated pupil using inbuilt calipers after capturing a fundus picture using cSLO-HRA system. We were able to measure lumen diameter and external diameter of 4 prominent vessel pairs in the zone between half a disc diameter and one disc diameter from the optic disc margin (2^nd circle on ETDRS grid). We have also calculated wall thickness and AVR in relation to both lumen and external diameter. All these parameters were calculated separately for OD and OS, gender-wise and age-wise for 3^rd, 4^th and 5^th decade and their normal values were analyzed in detail. All measurements were done in a large healthy cohort. Based on the ICCs and the Bland-Altman plot, we confirmed the reliability, reproducibility and validity of cSLO-HRA system measurements of retinal vessels.

Previously, Doppler OCT and SD-OCT have been used to measure retinal blood flow and vessel diameter in retinal and optic nerve head diseases(30)(31). Wang et al. (31) reported on the diameters of all veins around the optic disc that were scanned with two fixed diameter circular scans. We have measured the retinal vessel diameter at a known constant distance from the optic disc centre regardless of the disc diameter (which is not possible with a fixed diameter circular scans) unlike the technique used by Wang et al.

Previous in vivo studies reported measurements on the caliber of retinal vessels mainly based on fundus color photographs analyzed by a computerized algorithm (29). Hogan and Feeney conducted a postmortem study and reported the mean diameter of the principal Retinal arterial branches (peripapillary location) were 130 µm (a luminal diameter of 100 µm and wall thickness of 15 µm on each side). Our results of a mean arterial luminal and outer vessel diameter of 85.1 ± 12.4 µm and 105.04 ± 18.2 µm measured at 1800 µm from the optic disc centre are significantly different from their findings.  As the measurements of the arterioles and venules in our study were performed more peripherally compared with those of Hogan and Feeney the diameter is expected to be smaller.

Limited success was seen with previously studied AVR as a marker for systemic and ocular disease(32). In the present study, the mean AVR of lumen and external diameter were 0.66 ± 0.1 and 0.74 ± 0.09, respectively. Mean AVR was reported as 0.65 (130/200 mm) in Hogan and Feeney study (33), which are comparable with our findings of luminal AVR. The compatibility of the results in both of these studies emphasizes the potential for accuracy in retinal vessel measurement using a noninvasive technique, such as cSLO-HRA system. Liew et al. (34), demonstrated that AVR was not a particularly informative summary of retinal vessel measurement, as it conveyed less information than that provided by individual arteriolar and venular calibers. A major reason for the initial use of AVR, to correct for magnification differences, may not be as important as previously thought. Refractive error, when available, can be used to control for magnification (35), whereas, in its absence, bias from magnification differences is not profound in most eyes within the refractive power range of  ±3 D (35).

The arteriolar walls are thicker than venular walls with mean as 19.71 ± 8.04 µm and 11.02 ± 5.65 µm respectively, which was consistent with previous other studies. Chui et al.  (2013)(36) demonstrated that venular wall was relatively thinner compared with arterioles with similar lumen diameters due to their differing structure. Hogan et al. (33) described that retinal arteries differ from muscular arteries of other organs by their unusually well-developed media for vessel of this size, by the absence of an internal elastic lamina, and absence of elastin in the basement membrane system of the intima and media, where as the only pertinent observation that can be made about the larger veins is that they lack a muscularis, except those on the disc or possibly in its immediate vicinity.

The mean arteriolar and venular WLR (wall to lumen ratio) in this study was 0.26 ± 0.04 and 0.1 ± 0.05 respectively, which was calculated using the formula, (arteriole outer diameter −lumen diameter) / lumen diameter (Michelson et al.  2007(37)). In 2014, Zhu et al. (38) using SD-OCT, had reported mean arteriolar WLR as 0.36 ± 0.04. This difference may be due to the different devices used for measuring retinal vascular caliber. A study directly comparing the method described herein versus SD-OCT based approach may be necessary to understand reasons for this discrepancy. In 2009, Ritt and Schmieder(39) reported that arteriolar WLR changes may reflect vascular structure remodeling and in same year a study by Baleanu et al. (40) has shown that WLR was a more sensitive indicator than arteriole/venule ratio in the assessment of hypertensive cerebrovascular damage. In addition, Cuspidi and Sala(41) in 2011 suggested that arteriolar WLR could be a potential marker of endothelial dysfunction in both the retinal and systemic vasculatures.

Direct measurements of vessel diameters using image J software in previous studies have reported high repeatability and reproducibility(42). In our study, the intra and inter observer ICCs for cSLO-HRA system vessel measurements were also highly repeatable and reproducible, the ICCs were >0.95. whereas The Atherosclerosis Risk in Communities (ARIC) study(9) reported intraobserver and interobserver correlation coefficients of 0.69 and 0.74 for arteriole caliber, 0.89 and 0.77 for venule caliber. Garcı´a-Ortiz et al. (43) reported an interobserver ICC ranging from 0.96 to 0.98 for veins and arteries. Intraobserver ICC was also very high, ranging from 0.97 to 0.99. Sherry et al. (44) used a method based on the ARIC study, which improved both interobserver correlations (ranging from 0.78 to 0.90) and intraobserver correlations (ranging from 0.79 to 0.92), although the values achieved in our study were not reached. Muraoka et al. (2013) (45), using SD-OCT (Spectral-HRA + OCT), reported intervisit, interexaminer, and interevaluator ICCs ranging from 0.944 to 0.982 using an optic disc centered circle scan method, similar to that observed in our study.

The Bland-Altman plot shows that the difference between measurements using two methods, cSLO-HRA system and Image J software lie within the limits of agreement approximately 95% of the time. Both of these methods can be used to measure vessel caliber interchangeably.

Interobserver reproducibility was calculated by observations made by two retinal specialists, and one ophthalmology resident [who were naïve in terms of cSLO-HRA system but had received specific training (Initial 20 images) after technical discussion]. Thus, with adequate training it is possible that a non- ophthalmologist may also be capable of interpreting measurements of vessel diameters.

In theory, the retinal vessel wall cannot be visualized by fundus picture because it is transparent (Patton et al.  2006) (46). Pakter et al. (2011) (47) utilized fundus angiography as a gold standard for vessel lumen measurement and compared the results of vessel diameter measurement in fundus picture, they found that the vessel diameters obtained from the two methods were very close and suggested that vessel diameter measured in fundus picture actually was blood column diameter, as it was described in a previous study (Rassam et al. 1994 (48)). In our study, the lumen of vessels were measured by using Blue Reflectance (red free) images in which the laser is reflected from the blood column whereas Blue peak blue Autofluorescence images were used to measure external diameter as the laser is reflected from RPE making it as hyperreflective and leaving vessels (including lumen and wall) as hyporeflective, due to blocked reflectance by vessels.

Table 5: showing p value to determine significance of differences on comparing different parameters

S.NO.	PARAMETERS	p VALUE	COMMENT
OVERALL
1.	OD vs OS	0.09	No difference between OD and OS values
2	ST-A vs rest of arterioles	< 0.0001	ST-A is thickest for both lumen and external dimensions
3.	IT-V vs rest of venules	< 0.0001	IT-V is thickest for both lumen and external dimensions
4.	Inferior vs superior arteriolar wall thickness	0.20	No significant difference
5.	IT-A arteriolar wall vs rest of arteriolar wall	< 0.0001	inferotemporal arteriolar wall being thickest
6.	Inferior vs superior venular wall thickness	0.26	No significant difference
7.	AVR Nasal vs Temporal	<0.0001	AVR of lumen and external dimension both have significant difference with nasal half more than temporal half (AVR of lumen and external dimension correspond to 2:3 and 3:4, respectively)
8.	WLR Nasal vs Temporal	<0.0001	WLR of lumen and external dimension both have significant difference with nasal half more than temporal half.

Table 6: showing comparison between different parameters values evaluated in different previous studies and its comparison with the present study. (All values are in µm)

Studies		Hogan et al(33)	SB Lee et al (49)	Garcı´a-Ortiz et al (43)	Goldenberg et al (30)	TP Zhu et al (38)	TH Rim et al (50)	Present study
Year of study		1963	1998	2012	2013	2014	2016	2017
methodology		Histological dissection	optic disc photographs	Retinal photographs	SD-OCT with Image J	volume scan in SD-OCT	SD-OCT with intensity graph	cSLO
Distance of measurement		Peri-papillary border	Peri-papillary border		At 960 µm	between half and one disc distance		1800 µm from centre of disc
S.No.	Parameters
	Arteriolar
1.	Internal dimension- lumen	100				80.94 ± 11.78	95.1±16.1	85.10 ± 12.40
2.	External dimension	130				109.33 ± 13.18		105.04 ± 12.05
3.	Mean arteriolar caliber		102 ± 16 (ST) 93 ± 12 (IT)	106.00	127.81 ± 13.42
4.	Wall thickness	15				14.19 ± 1.10	23.9±4.9 (inner sides) 21.2±3.5 (outer sides)	19.71 ± 8.04