You are on page 1of 10

Special Issue

Optical Coherence Tomography Angiography Analysis of
Perfused Peripapillary Capillaries in Primary Open-Angle
Glaucoma and Normal-Tension Glaucoma
Nicole K. Scripsema,1 Patricia M. Garcia,1 Richard D. Bavier,1,2 Toco Y. P. Chui,1 Brian D.
Krawitz,1 Shelley Mo,1 Steven A. Agemy,1,3 Luna Xu,1 Yijie B. Lin,1 Joseph F. Panarelli,1 Paul A.
Sidoti,1 James C. Tsai,1 and Richard B. Rosen1
1Department of Ophthalmology, The New York Eye and Ear Infirmary of Mount Sinai and Icahn School of Medicine at Mount Sinai,
New York, New York, United States
2
Rutgers New Jersey Medical School, Newark, New Jersey, United States
3
Department of Ophthalmology, SUNY Downstate Medical Center and SUNY Downstate College of Medicine, Brooklyn, New York,
United States

Correspondence: Richard B. Rosen, PURPOSE. To compare perfused peripapillary capillary density in primary open-angle glaucoma
310 East 14th Street, 5th Floor South (POAG), normal-tension glaucoma (NTG), and normal patients using optical coherence
Building, New York, NY 10003, USA; tomography angiography (OCT-A).
RRosen@nyee.edu.
METHODS. A retrospective review of POAG, NTG, and normal patients imaged with OCT-A was
Submitted: December 14, 2015
Accepted: August 19, 2016
performed. En face OCT angiograms identifying peripapillary vessels were obtained using a
spectral-domain OCT system (Avanti RTVue-XR). A custom image analysis approach identified
Citation: Scripsema NK, Garcia PM, perfused peripapillary capillaries, quantified perfused capillary density (PCD), and generated
Bavier RD, et al. Optical coherence color-coded PCD maps for 3.5- and 4.5-mm-diameter scans. We compared PCD values, PCD
tomography angiography analysis of
maps, standard automated perimetry (Humphrey visual field [HVF]) parameters, and OCT
perfused peripapillary capillaries in
primary open-angle glaucoma and retinal nerve fiber layer (RNFL) thickness analyses across all groups.
normal-tension glaucoma. Invest RESULTS. Forty POAG, 26 NTG, and 26 normal patients were included. Annular PCD in POAG
Ophthalmol Vis Sci. (34.24 6 6.76%) and NTG (37.75 6 3.52%) patients was significantly decreased compared to
2016;57:OCT611–OCT620. DOI: normal patients (42.99 6 1.81%) in 4.5-mm scans (P < 0.01 and P < 0.01, respectively).
10.1167/iovs.15-18945
Similar trends and statistical significances were seen in 3.5-mm scans. Linear regression
analysis resulted in moderate correlations between annular PCD values and other
glaucomatous parameters. Pearson coefficients comparing annular PCD from 4.5-mm scans
in POAG and NTG groups to HVF mean deviation, HVF pattern standard deviation, and
average RNFL thickness all showed statistical significance (P < 0.05). Color maps showed that
POAG and NTG patients had a reduction of perfused capillaries that progressed in size when
comparing early, moderate, and severe glaucoma groups.
CONCLUSIONS. Optical coherence tomography angiography can uniquely identify changes in
peripapillary PCD in glaucoma patients. Optical coherence tomography angiography may
offer insights into the pathophysiology of glaucomatous damage and risk factors for disease
progression.
Keywords: optical coherence tomography, glaucoma posterior segment, angiography, low-
tension glaucoma, optic nerve head

laucoma is one of the leading causes of blindness difference between systemic blood pressure and IOP. Several
G worldwide, currently afflicting more than 60 million
people and predicted to affect 79.6 million people by
studies including the Baltimore Eye Survey and the Barbados
Eye Study have identified both high IOP and low ocular
2020.1–3 Given the insidious course and irreversible damage perfusion pressure as risk factors for glaucomatous nerve
caused by glaucoma, early diagnosis, prompt treatment, and damage.10–14 The role of systemic hypertension in glaucoma-
close monitoring are critical in disease management. Even tous damage is less well established, with varied reports in the
when optimal intraocular pressure lowering is consistently literature.6,10,11,15–18 Hypotension, particularly nocturnal hypo-
achieved, a subset of patients continue to progress. The tension, has shown increasing evidence as a strong risk factor
pathophysiological theories correlating high intraocular pres- for disease progression.19–26 Despite this evidence, our
sure (IOP) with mechanical optic nerve damage cannot always understanding of the role of optic nerve perfusion in the
account for this progression. Studies suggest that vascular pathophysiology of glaucoma is limited.
dysfunction causing optic nerve hypoperfusion may contribute Standard automated perimetry remains the gold standard for
to glaucoma progression in patients with both high and normal visual field testing and glaucoma assessment. While perimetry
IOPs.4–9 Alterations in systemic blood pressure or vasospasm provides information essential for the proper management of
may result in low ocular perfusion pressure, defined as the the disease and interobserver reproducibility among glaucoma

iovs.arvojournals.org j ISSN: 1552-5783 OCT611

This work is licensed under a Creative Commons Attribution 4.0 International License.

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/935468/ on 03/27/2018

defined as the network of capillaries from the categorized into early. Dublin. Patients whose images had inadequate signal cup-to-disc ratio (CDR) asymmetry of >0. obtained with a commercial spectral-domain OCT system Primary open-angle glaucoma. used to compare to OCT-A data. The specifics of similar algorithms have reproducible visual field tests with glaucomatous defects. glaucoma treatment. hypertension.40 Other groups have reported performed within 1 month of imaging.5- rim thinning. The research study adhered to the tenets of the Declaration of Helsinki and was approved by the OCT-A Image Acquisition Institutional Review Board of the New York Eye and Ear Optical coherence tomography angiography images were Infirmary of Mount Sinai. Average RNFL thickness and RNFL quadrant analysis were used for comparison with OCT-A data. Glaucomatous optic neuropathy was defined based consisting of both 3. optic disc structure are increasingly used in the assessment of with other points elsewhere at least 10 dB higher than the glaucoma. OCT RNFL provide images of perfused capillaries in both the perifoveal thickness analysis. USA) as previously imaged with OCT-A from April 2015 to August 2015 were described.2 not explained by strength were excluded. Patients with prior cataract extraction or progression. gonioscopy. reliable.5-mm-diameter peripapillary scans (4. Images with these OCT-A artifacts history of elevated IOP (>24 mm Hg) either with or without were excluded. for each scan. Inc.5 3 4.41–45 A study analyzing false positives < 15%. Inclusion criteria for NTG patients were a history of A SSADA algorithm was used to identify perfused vessels. central corneal thickness (CCT).40 Each patient underwent a single imaging session included.5 3 3. When the OCT-A image is poor due to significant specialist confirming these clinical findings on a minimum of motion artifact or low signal strength. insights into the status of the retinal microvasculature could be an excellent indicator of the Visual Field Testing presence or potential for disease. Inclusion criteria for POAG patients required a angiographic reconstruction. Exclusion criteria Optical coherence tomography angiography (OCT-A) is a included diabetes mellitus. Inc. A been previously published. our group demonstrated the use of perifoveal OCT-A imaging in the Patients were included only if they had a reliable visual field staging of diabetic retinopathy. All patients had undergone a complete ophthalmic retinal vasculature.43. NTG. the minimal visual field defect consisted of a cluster of three algorithm distinguishes the movement of red blood cells within adjacent points depressed by at least 5 dB from normal age the lumen of retinal and choroidal vessels between cross- values. Optovue.47–50 To summarize briefly.39 Currently. OCT Angiography Peripapillary Analysis in Glaucoma IOVS j Special Issue j Vol. or an intereye vertical mm scan). moderate. At least three points of such a cluster variability. metabolic demand.arvojournals.) was performed prior Study Population to OCT-A. and false negatives < 33%.org/pdfaccess.. or severe glaucoma sub- inner limiting membrane (ILM) to the posterior boundary of groups based on the Hodapp-Anderson-Parrish (HAP) Visual the RNFL. signal strengths of less than 7/10 were excluded. can detect variations between primary open-angle Field Severity Score. with one of these points depressed by at least 10 dB sectional scans. 9 j OCT612 experts is high. and performed on all patients within 6 months of imaging. Most patients included have been followed in our or gaps indicating insufficient information for complete clinic for years. including glaucoma. In a previous study.41. In addition. Carl Zeiss Meditech. Peak IOP. current one of several modalities currently used as an objective. the SSADA algorithm can patients with Humphrey visual field (HVF). Lowenstein Retina Center of the New York Eye and Ear Infirmary of Mount Sinai. Only noninvasive manner. the image contains lines two visits.28 Consequently.5-mm-diameter peripapillary scans on clinical exam findings including focal or diffuse neuroretinal (3. several conditions. patients were lary capillaries. and if our custom image analysis approach that provides deviation (MD) and pattern standard deviation (PSD) were qualitative and quantitative analysis of the perfused peripapil. including best-corrected visual acuity.42 with size III white stimulus and a standard SITA (Swedish The purpose of the present study was to determine if OCT-A interactive thresholding algorithm). OCT Circumpapillary Retinal Nerve Fiber Layer Analysis METHODS Optical coherence tomography RNFL thickness analysis (Cirrus HDT OCT 500. Visual fields peripapillary glaucomatous changes with OCT-A reported a were performed with the Humphrey Field Analyzer II (Carl decrease in perfused peripapillary vasculature in glaucoma Zeiss Meditec. or other ocular novel technique that employs en face reconstruction of OCT conditions with the exception of cataract or prior cataract combined with motion contrast processing to reveal perfused extraction. Split-spectrum amplitude-decorrelation examination.27. age. 57 j No. objective tests that analyze the were required to be on one side of the horizontal meridian. visual fields often have substantial intertest from normal age values. A dilated fundus exam was angiographic images of perfused vessels in a fast.46 glaucoma (POAG). IOP on the day of of peripapillary retinal nerve fiber layer (RNFL) thickness is imaging. normal-tension glaucoma (NTG). and normal patients. The decorrelation algorithm identifies per- Downloaded From: http://iovs. Fremont. and surgical history were also recorded quantifiable measure for detecting glaucoma and assessing for each patient. Patients with visual field is capable of differentiating normal eyes versus glaucomatous abnormalities had at least one confirmatory examination. All glaucoma patients including larger retinal vessels and the surrounding capillary meeting these criteria were also required to have at least two network. and disc photos performed within 1 month and peripapillary regions. All patients required documentation by a glaucoma image.33–38 glaucoma surgery were not excluded. As the neuroretina has a high of imaging were included.29–32 Optical coherence tomography (OCT) analysis densest point in the scotoma. Assessment of adequate signal differences in disc size.5-mm) and 4. and normal patients (Avanti RTVue-XR. CA) using a 24-2 threshold test patients compared to normal subjects. slit-lamp angiography (SSADA) algorithm detects the movement of red biomicroscopy. Reliable performance the utility of optic disc and peripapillary OCT-A imaging for on visual field testing was defined as fixation errors < 20%. untreated peak IOP of 21 mm Hg or less. Mean eyes. and Goldmann applanation to- blood cells within retinal and choroidal vessels to generate nometry on the day of OCT-A. focal or diffuse RNFL loss. Analyses with This retrospective review was conducted at the Bendheim. CA. An open angle on gonioscopy was also strength was achieved by analysis of the resulting OCT-A required. treatment.ashx?url=/data/journals/iovs/935468/ on 03/27/2018 .

which must be compensated for in order to yield NTG 26 34 67.org/pdfaccess.5-mm OCT-A image of a healthy control. (A) Contrast-stretched grayscale 4. 57 j No. (F) Perfused capillary area within the annular ROI (in cyan) superimposed with (A).07 Correction Technology MCT) compensates for motion artifacts by combining matching horizontal-priority and vertical-priority * No statistically significant difference was observed between study scan volumes using an orthogonal registration algorithm to groups with respect to age (1-way ANOVA). Downloaded From: http://iovs.ashx?url=/data/journals/iovs/935468/ on 03/27/2018 .34 decorrelation. fused retinal vessel from surrounding static tissue based upon signal amplitude variation differences in nonstatic tissue compared to static tissue. Study Group N.92 6 9. It cannot quantitate flow differences TABLE 1. (B) Major blood vessel mask created using global thresholding.72 6 9. Demographic Data of Study Groups below a certain threshold of motion. (C) Binary image after local adaptive thresholding with major blood vessels removed.35 acceptable vascular maps. (E) Annular ROI centered at the optic nerve head. The Optovue software (Motion Normal 26 38 63. Subjects % Male Age P Value* Subject movements during image acquisition also generate POAG 40 55 66.30 6 11. 9 j OCT613 FIGURE 1. OCT Angiography Peripapillary Analysis in Glaucoma IOVS j Special Issue j Vol. meaning that vessels with slow or no flow are not visualized with this technique.arvojournals. OCT-A image processing steps. (D) Color-coded perfused capillary density map.49 0.

appear blue on the color maps.55 on the grayscale OCT-A image with the value Repeatability and Reproducibility 1 (white) and the remaining pixels with the value 0 (black).27 96. PCD within a fixed annular ROI centered In this study we analyzed the perfused peripapillary at the optic nerve head was extracted for quantitative analysis.82 69.10 6 14. Each resized number of pixels in the entire image after removal of the inner image was contrast stretched by using the lowest and highest 1.08 543.26 Normal 1.84 6 9. In both 3.ashx?url=/data/journals/iovs/935468/ on 03/27/2018 . CDR.and 4.08 6 33. The Optovue built-in software automatically enced examiner. majority of commercially available OCT systems. NTG. These analyses were performed for all controls and glaucoma patients using both 3.58 0. and Pearson’s correlation coefficient (r) were A color-coded perfused capillary density (PCD) map was calculated. global thresholding was performed glaucoma using the 3.45- peripapillary capillaries located between ILM and the posterior mm (outer) diameters were placed manually by an experi- boundary of RNFL. lm.44 Comparisons were made between POAG and NTG patients to demonstrate a similar severity of glaucoma across the groups. allowing for the creation of en face OCT angiograms.5. relationship between PCD and other variables. mm Hg. * P < 0.58 6 37. 1B) for the removal were previously published and are consistent with the of the pixels associated with the major blood vessels on the literature. The means.91 26. 1A).43–45. No.04 6 1.94 <0. Annular PCD Although both eyes were imaged. remove bulk motion and produce a merged OCT image volume OCT-A Quantitative Analysis: Annular Perfused with essentially no residual motion artifacts.25 6 15. (%) regional peripapillary capillary density variations across the Early 22 (55) 13 (50) entire OCT-A image (Fig.5. NTG.arvojournals. bright red Moderate 6 (15) 7 (27) indicates a high density of perfused capillaries. No. 1E. MA.41.01* 538. in all OCT-A images. 1F). Larger retinal vessels excluded from the analysis also Based on Hodapp-Anderson-Parrish Visual Field Severity Score.25 0.95-mm (inner) and 3. 1C). OCT-A Image Analysis All OCT-A images were analyzed using a custom-developed Global Perfused Capillary Density (%) MATLAB program (The Mathworks.76 6 0.06 0.5-mm scans.15 20.25 73.org/pdfaccess.54 6 44.05 was considered statistically significant (unpaired t-test). Comparisons were made with both annular and global Capillary Density Maps PCD for 3. 9 j OCT614 TABLE 2. In the color maps. USA). Glaucoma Severity in POAG and NTG Groups 3 16-pixel sampling window with 8-pixel overlaps over the final binary image (Fig.5.45-mm outer circle diameter z-projection of the maximum decorrelation value to generate an represented the standard circumpapillary scan dimension for en face OCT-A image (Fig.82 530. which includes the larger blood vessels as well as In brief.09 18.5-mm OCT-A images.52 Optical coherence tomography angiog- OCT-A image. to convert the contrast-stretched grayscale OCT-A image into a binary image by replacing all pixel intensity greater than the level value of 0. 1C).25 0. Global PCD was calculated by generating a percentage of the First. Natick. setting the decorrelation value for bulk tissue to approximately To ensure that the same region of interest (ROI) was included zero. HVF PSD. a second image thresh- olding was performed using a local adaptive thresholding algorithm with a sampling window size of 15 3 15 pixels.and 4. and Normal Patients Average Tmax.15 6 5. This color-coded density map Glaucoma allowed for quick interpretation and better visualization of Severity POAG Patients. These 1% pixel intensity values of the image as the lower and upper analyses were also performed for all controls and patients with limits.40. dark blue Severe 12 (30) 6 (23) indicates area of low or no perfused capillaries. Glaucoma Parameters for POAG.72 0. 1D). annular and global generated by computing capillary area density within each 16.95-mm circular area and the major blood vessels.and 4. including average OCT RNFL thickness.32 6 0.5-mm OCT-A images. and normal TABLE 3.75 mm (Figs. CCT.5-mm scans. standard deviations. two concentric circles with 1. contrast-stretched grayscale OCT-A. Group Mean 6 SD P Value Mean 6 SD P Value Mean 6 SD P Value Mean 6 SD P Value Mean 6 SD P Value POAG 8. PCD were compared among the POAG. vasculature. ROI after the removal of major blood vessels. HVF MD.76 6 0. producing an annular ROI with a width of segments these two boundaries on the OCT scans and performs 0. Inc. OCT Angiography Peripapillary Analysis in Glaucoma IOVS j Special Issue j Vol. each grayscale OCT-A image (304 3 304 pixels) was number of pixels associated with perfused capillaries over the resized by a factor of six (1824 3 1824 pixels).51 It also calculates Capillary Density (%) the mean decorrelation and subtracts this value from each frame.43 6 8. This Statistical Analysis final binary image contains only the perfused peripapillary Linear regression analysis was used to investigate the capillaries (Fig.35 NTG 8. (%) NTG Patients. The OCT-A images of all patients RNFL thickness measurement that is currently employed in the were reviewed to ensure proper segmentation of OCT scans.04 6 9. and OCT-A Qualitative Analysis: Color-Coded Perfused IOP. After the removal of major blood vessels on the raphy has excellent repeatability and reproducibility indices. respectively.5. 57 j No.43 6 1.and 4. RNFL. The repeatability and reproducibility of the Optovue Avanti This binary image was used as a mask (Fig. only a single eye from each was calculated by the number of pixels associated with patient was selected at random and included for data analysis to perfused capillaries over the number of pixels in the annular avoid potential interocular correlation between eyes. Then.17 0. and Total 40 26 intermediate densities are represented on a spectrum of yellow to green. Study HVF MD.91 0. Downloaded From: http://iovs.. lm.33 6 7. This 3. age.

01‡ Normal 41. (Table 3). 27% moderate. and 26 normal patients 6 1.03 6 1.32 6 1. Annular PCD.52% for NTG. moderate. version 23. Males represented 46% of all significant differences between group means for both 3.43 6 8. Mean age was 66.94 mm glaucoma groups were limited.45). groups by 1-way analysis of variance (ANOVA). 57 j No. 37. IBM. Annular PCD for the 3.05‡ NTG 36. 3.05 was considered statistically significant (1-way ANOVA. P ¼ 0. also similar across all groups (1-way ANOVA. and 30% severe.24 6 6.5-mm was included.20 6 3.5-mm Scan.05‡ 34. 15% moderate.96% 41. 4. The only statistically significant changes were statistically significantly different between POAG (14. Field Severity Scale (Table 2). This distribution was both 3. with 50% early.18% <0.24 6 6.5-mm scans was significantly reduced in both similar across all groups (1-way ANOVA. P ¼ 0. determined by chart review. A total of 62% of patients were Hispanic. after the removal of both the inner 1. moderate. Qualitative analysis was also performed to compare dorzolamide-timolol.and 4.81% * P value of glaucoma group versus normal. All analysis was performed with medications in addition to brimonidine. area and the major blood vessels. while NTG patients were treated with an average of performed for the comparisons of the glaucoma parameters 1.and patients.5-mm scans (1-way ANOVA.82 mm Hg) groups (unpaired t-test. and 42. Study Group Mean 6 SD P Value* P Value† Mean 6 SD P Value* P Value† POAG 33. a The comparison of annular PCD in POAG.01‡ <0. Trabeculectomy (83%) was more common The charts of 137 patients imaged with OCT-A were reviewed. In the NTG group 54% statistical software (SPSS. In the POAG between POAG and NTG groups.76% <0. and normal total of 92 patients were included.40 6 6.5. CDR.20 6 3. After excluding patients based on the criteria listed above. Global PCD.85 6 between early and moderate glaucoma in both POAG (Mann- 3.45 A single eye from 40 POAG. Tukey’s HSD test). OCT Angiography Peripapillary Analysis in Glaucoma IOVS j Special Issue j Vol.57 mm Hg). 37. Mean IOP on the day of imaging (Ta) was also increased. Annular PCD for the 4. and normal groups Whitney U test.001).01‡ <0. NTG.81% for normal patients. and 23% severe. an equal distribution of early. The NTG image.88 6 2.arvojournals.01‡ <0. Table 1).95-mm circular group was similarly distributed.49 6 3. was also performed (Table 5). Post hoc pairwise Table 1). 4. Corp.40 mm Hg). Study Group Mean 6 SD P Value* P Value† Mean 6 SD P Value* P Value† POAG 33. P < 0. than tube shunt surgery (17%). and Normal Patients Annular PCD.08 eye pairwise comparisons between groups.76 6 1.5-mm Scan. 26 NTG.5-mm Scan.5. P ¼ 0. While the sample sizes in the early. 15% in combination with NY.29 years overall. Downloaded From: http://iovs. NTG (12. † P value of POAG versus NTG.04) groups (Table 6) (15..0.01‡ <0. average RNFL thickness.51% for NTG. Primary open-angle glaucoma significant difference (HSD) test was applied for post hoc patients were treated with an average of 2.75 6 3. Tukey’s HSD test). a global PCD analysis of each group was 55% early. Global Perfused Capillary Density (PCD) Values in POAG. We found similar trends and statistical significances in the There was no significant difference between mean HVF MD. and Normal Patients Global PCD. and CCT in the glaucoma groups 1E).01) and NTG (P ¼ 0.05 was considered statistically significant (1-way ANOVA. Mean peak IOP (Tmax). Annular Perfused Capillary Density (PCD) Values in POAG.53% <0.org/pdfaccess. A total of 22 POAG.14 eye drops (unpaired t-test.13 6 6.56% 42. Four (10%) POAG patients and two (8%) NTG patients underwent previous RESULTS glaucoma surgery. Tukey’s honest also varied between the groups. moderate. Annular PCD represents the and 7 normal patients were excluded.90 6 2. P ¼ 0. 16 NTG.88 6 1. Treatment and between early and severe glaucoma in the POAG group (P TABLE 5.01‡ 37. Armonk. 17% in used to compare PCD values between early.40 6 6.34.45 6 1.46 The distribution in the POAG Using our custom software.05 6 10.98 mm Hg.66.76% for POAG. Unpaired t-tests were drops. 9 j OCT615 TABLE 4. and severe was significantly different between POAG (26. and severe glaucoma Annular PCD was also significantly reduced in the POAG group patients in the POAG and NTG groups based on the HAP Visual when compared to the NTG group (P < 0. NTG.75 6 3. NTG.10% <0. and 31% both medications in addition to color-coded PCD maps to average RNFL thickness. Nine (22%) POAG patients and four (15%) NTG patients underwent prior cataract extraction.5-mm Scan. patients is displayed in Table 4. 20% African comparisons between groups indicated that annular PCD for American.51% <0.48 6 3.01‡ 36.ashx?url=/data/journals/iovs/935468/ on 03/27/2018 . and combination with dorzolamide-timolol. There was POAG and NTG groups when compared to normal (P < 0.05‡ NTG 37.53% for POAG.5-mm scans was visually significant cataract (29%). 33.99 6 1.56% for normal patients.49% <0. global PCD values when compared to the annular PCD (Fig.01). P demonstrated decreased annular PCD as glaucoma severity < 0. ‡ P value < 0.99 6 1. 1-way ANOVA. USA). There were statistically ANOVA. 4. Mann-Whitney U test was group 15% used a prostaglandin analogue alone. brimonidine. 3.01‡ Normal 42. 14% Caucasian.01). and 10% Asian.01).23% <0. used a prostaglandin analogue alone.01). and 68% both these severe glaucoma subgroups. using ANOVA was found to be similar among all groups (1-way and 42. P ¼ 0. ‡ P value < 0. and scans was 34. † P value of POAG versus NTG. both POAG and NTG patients Hg) and NTG (18.16% * P value of glaucoma group versus normal.52% <0.05‡ 33. A majority of patients percent of pixels associated with perfused capillaries over the were excluded due to unreliable visual fields (71%) or poor total number of pixels in the annular ROI after removal of the image quality secondary to poor fixation or the presence of a major blood vessels.69 6 2.01). P < 0.31 6 6.

50 in ¼ 0. (A–D) Graphs display the results of the linear regression analysis used to compare annular PCD in 4.42 and r ¼ 0. 57 j No. respectively).5-mm scan was r ¼ 0.and 4. P angiography is a technique first described by Jia et al.org/pdfaccess.74 (P < 0.65 6 3. respectively) or IOP (r ¼ 0. areas lacking in NTG was not statistically significant (P ¼ 0.10. red blood cells flowing in retinal and choroidal vessels. Moderate 32.50.5-mm scans. This the perifoveal and peripapillary regions in diseased correlation was weaker for NTG patients. The comparison between early and severe glaucoma and NTG patients as glaucoma advances.02) in the NTG group.5. 1).01) for the POAG group and r ¼ 0. < 0.24 6 6. from ILM to Bruch’s membrane.25 6 5.42 The present FIGURE 2. P ¼ 0. Annular Perfused Capillary Density (PCD) by Glaucoma Stage 0. 9 j OCT616 TABLE 6.66% 39. mean HVF MD.01). P ¼ 0.50 reported that quantitative analysis of OCT-A can differentiate (P ¼ 0.63%* 36. including overall RNFL thickness. In both the POAG and RNFL peripapillary capillaries of glaucomatous patients com- NTG groups there was no significant correlation between pared to age-matched normals are similar to recently published annular PCD values and age (r ¼ 0.47 and r ¼ 0.19).5.32.arvojournals.69 (P < 0.76% 37. and HVF PSD.52% The linear correlation between annular PCD values.01 and r ¼ 0. and mean OCT RNFL thickness * Statistically significant difference compared to early glaucoma (P suggests that as glaucoma progresses.5-mm Scan 4. again with the POAG group showing stronger correlation than including the fine capillary networks.001). respectively.and 4. Analysis of the color-coded PCD maps showed that when comparing color maps in early. annular PCD decreases. mean HVF PSD.01) for the NTG group.01) and r ¼ 0.94.42 6 2. P ¼ 0. respectively). HVF PSD.56% mm annular PCD values with 3.01). Graphs correlating 4.52%* 36.22.46 (P ¼ 0.05% through 2D. Downloaded From: http://iovs.18 and r ¼ 0. age.01) in the POAG group peripapillary vasculature.5-mm annular PCD values.5- Early 38.05.61 (P < 0.75 6 3. There was a moderate linear relationship between average RNFL thickness and annular PCD for POAG and normal DISCUSSION patients reaching statistical significance in both 3. Total 34. comparison to the other groups.5-mm scans to other clinically relevant information used when assessing glaucoma patients.12. HVF MD. Pearson correlation for HVF PSD glaucomatous eyes from normal eyes in analyzing the entire in the 4. NTG Annular PCD. We also analyzed the correlation between POAG Annular PCD.02.01 6 3.16 6 2. P ¼ 0. and IOP are shown in Figures 2A Severe 26.5-mm images without intravenous contrast. P < 0.5-mm Scan strong correlation (r ¼ 0.ashx?url=/data/journals/iovs/935468/ on 03/27/2018 . moderate. Previous studies demonstrate the use of OCT with a SSADA mm scans (Fig. P ¼ data analyzing the entire peripapillary retina. likely as a perfused capillaries become larger as glaucoma progresses result of a small sample size and higher standard deviation in (Fig.42–44. RNFL. Similar correlations Our findings of decreased annular and global PCD values in the were demonstrated in the 3.72 (P < 0. Pearson correlation for the POAG group was algorithm in the detection of microvascular changes in both r ¼ 0. Mann-Whitney U test). 3. it did still eyes. Pearson correlation for HVF MD in the 4.94%* HVF MD.89 6 6. The algorithm uses intrinsic motion contrast to detect also showed moderate correlation with annular PCD values.5-mm annular PCD values. which showed a very Glaucoma Severity 4.53–55 Split-spectrum amplitude-decorrelation reach statistical significance (r ¼ 0.5.40. Previous studies have scans was r ¼ 0.10.50. and creates angiographic the NTG group.66). 3). however.47. Humphrey visual field MD and HVF PSD 2012. and severe POAG < 0. OCT Angiography Peripapillary Analysis in Glaucoma IOVS j Special Issue j Vol.45.

can that PCD declines as glaucoma progresses (Table 6. and a severe NTG patient. which offers the advantage of using qualitative data use a single operator. 9 j OCT617 FIGURE 3.61–67 In this study we chose to mapping. as our POAG group was within-visit repeatability as well as high between-visit repro- using a larger number of drops overall.org/pdfaccess.42 eyes in the clinical setting. OCT-A detecting glaucoma with similar sensitivities when compared appears more reproducible and reliable.ashx?url=/data/journals/iovs/935468/ on 03/27/2018 .43.57 This device is the difficulty encountered when imaging patients difference could be related to the different pathophysiological with poor vision or unstable fixation. a severe POAG. Our current study demonstrates its value for the quantitative To our knowledge.52 One limitation of the may have effects on ocular perfusion pressure. which may affect processes proposed in normal. these measurements are overall density of perfused vessels. which consistent with other reports. along with the quantitative PCD value. Our findings were previously reported and are percentage of patients on medications such as timolol. but in a clinical practice OCT-A can easily to quickly and easily differentiate normal versus glaucomatous be used by multiple operators with good reproducibility. While the thickness maps.5-mm OCT-A images and color-coded perfused capillary density maps in a normal. help in the detection of disease. It Optical coherence tomography angiography has high could be related to medication effect. All images show the temporal peripapillary region on the left.40 eyes.arvojournals. this is the first study to compare and qualitative evaluation of perfused peripapillary capillaries perfused peripapillary capillary densities in POAG and NTG in glaucoma. Its ease of use to OCT RNFL thickness. First column: grayscale OCT-A images generated using SSADA algorithm. regions lacking perfused capillaries ex- mild. Our findings support their report. In addition.40–44. Third column: corresponding color-coded perfused capillary density maps. panded as in more advanced stages of glaucoma.and high-tension glaucoma. Both POAG and NTG patients demonstrated color-coded PCD maps are as easy to interpret as OCT RNFL decreased PCD values compared to normal patients. it appears fast method that.42 suggested that changes in OCT-A peripapillary techniques that attempt to measure blood flow using laser vessel density may be a reliable diagnostic parameter for Doppler flowmetry or laser speckle flowgraphy. moderate. repeatability and reliability. This translates into a very high interoperator reproducibility study also introduces the concept of OCT-A color-coded PCD compared to other devices.56. it quantitative data.42 We recently described the use of this novel values and generating color-coded PCD maps of perfused OCT-A mapping technique in perifoveal OCT-A images for peripapillary capillaries in the evaluation of glaucomatous quantitative and qualitative grading of diabetic retinopathy. Figs. We found that the qualitative PCD maps offer a sample size of our within-group analysis is limited. In comparing the early. Comparisons of 4. the color maps Downloaded From: http://iovs. this difference is less likely to be related to disease stage.33–36.58–60 However. moderate. OCT Angiography Peripapillary Analysis in Glaucoma IOVS j Special Issue j Vol. Last column: superimposed image of the color maps and the inverted OCT-A images. Color-coded topographic maps with While OCT-A is not capable of directly measuring blood flow. 3. and advanced glaucoma cases in each group. with POAG patients having a lower patients. have offers insights into the perfusion status of the peripapillary been used successfully by clinicians to evaluate glaucoma region by identifying perfused vessels and quantifying the patients for many years. Optical coherence tomography angiography using OCT-A. Compared to other noninvasive Liu et al. Given the similar distribution of glaucoma patients. such as OCT RNFL thickness analysis. 4). Second column: thresholded binary images containing only the perfused capillaries after the removal of major blood vessels. including a larger ducibility. 57 j No. and severe PCD value than NTG patients. study demonstrates the utility of the Optovue SSADA algorithm limited to detecting evidence of more advanced stages of and our custom image analysis approach in calculating PCD disease. Both POAG and NTG patients We also see a significant difference in PCD values between the have a decrease in perfused vessels compared to normal POAG and NTG groups.

Patients with improve longitudinal analysis. patients must have good advantage of using the fixed annular ROI is the ability to fixation and relatively good central acuity. This suggests that even when image size varies. OCT-A was capable of capturing ed across all groups. This may allow for improved longitudinal In summary.41 and Hispanic patients. Other factors that may influence PCD include IOP. Current ethnicities. The phy is an easily interpretable. are unlikely to have time. noninvasive tool relatively small sample size and retrospective nature limit the that uniquely permits visualization of perfused peripapillary power of our findings.and 4.and 4. Second column: quadrant-specific perfused capillary density of the corresponding OCT-As on the left. This was also a cross-sectional analysis capillaries. Future refinements in the software may serve (Table 5). further refine- peripapillary capillary images after the removal of both the ments of both the SSADA algorithm and our custom image major blood vessels and the inner 1. In enough to detect dynamic changes in perfused peripapillary addition. making it an attractive imaging modality for Downloaded From: http://iovs.org/pdfaccess.arvojournals. Mean RNFL thicknesses are indicated in the center. we found research is necessary to determine the effects of these variables similar PCD values when applying the annular ROI to these on PCD values.ashx?url=/data/journals/iovs/935468/ on 03/27/2018 . easily identify regions of decreased perfused vessel density.5 3 3.5-mm peripapillary scans. Further scans.5-mm scans density scaling. it is possible that PCD varies between both 3.5-mm OCT-As. Interestingly. glaucoma patients.5 3 4. slightly decentered images or various scan sizes obtained over such as cataracts or macular lesions. and patients were not followed over time.001. global PCD analysis of the perfused RNFL reliable or reproducible results. comparable results. OCT Angiography Peripapillary Analysis in Glaucoma IOVS j Special Issue j Vol.5 and 4.5-mm annular PCD values required for the OCT-A imaging process. 57 j No. Last column: circumpapillary RNFL thickness plots. reliable. P < 0.5. 4. Further glaucomatous patients from normal patients. so while OCT-A has This high-resolution imaging not only is capable of differenti. software capabilities have expanded to allow 4. and may offer research is necessary to determine if this translates to the new insights into early diagnosis and a better understanding of newer 6-mm images as well.and 4. 9 j OCT618 FIGURE 4. 2A).and 6-mm medications. fast.5-mm quantitative information provided by OCT-A can differentiate scans. POAG. The 3. All OCT-A images show the temporal peripapillary region on the left. we cannot comment on the ating acquired changes in diseased states.5. Comparisons of peripapillary perfused capillary density and circumpapillary RNFL thickness in the same normal. our study was composed mostly of African American vessels in transient states of hyperoxia. Mean perfused capillary densities are indicated in the center. Optical coherence tomography angiogra- There are a number of limitations to our study. this study demonstrates that the qualitative and analysis of patients imaged with either 3. In addition. or a combination of scans acquired over time. and other systemic vascular conditions. Fig. but is also sensitive device’s effectiveness in assessing disease progression.5-mm scans.94. While ethnicities were evenly distribut- At the time of our study. our to assist OCT-A in becoming a standardized imaging modality custom image analysis approach is capable of producing for the diagnosis and management of glaucoma. In order to obtain correlated very strongly (r ¼ 0. First column: annular peripapillary capillary area superimposed with 4. When comparing the 3. especially in the setting of advanced glaucoma or poor acuity due to other conditions. good repeatability and reliability. and NTG patients shown in Figure 2. Another limitation is the acquisition time images (Table 4).5-mm scans.5. The OCT-A image of the left eye has been flipped horizontally. Third column: quadrant-specific RNFL thickness of the corresponding RNFL thickness plots on the right.5. One images with adequate signal strength.95-mm circular area (inside analysis approach may help improve the concept of perfusion the disc) also produced similar results in 3.

Podhajsky P. pressure. and systemic hypertension: the Blue Mountains Eye Study. Wu SY. Open-angle glaucoma Acknowledgments and ocular hypertension: the Long Island Glaucoma Case- control Study. Medeiros FA. 31. Proyecto VER. Hypertension. Musch DC. initial glaucoma treatment study: baseline visual field and test- Ophthalmology. et al. v. Wu SY. Hussain RM. Graefes Arch Clin Exp Ophthalmol. B. Podhajsky P.113:216–221. et al. 267. Marrus Glaucoma Research Fund and the hypotension increases the risk of glaucoma progression.18:345–353.B. Allergan (C). Zimmerman MB. Ophthalmology. et al. Katz J. 1990. and further research is 954–959. Congdon NG. 9. Chui. Connell AM. Arch Ophthalmol.213:76–96. West SK. glaucoma patients. Charlson ME. MacKenzie CR. T. 1999. 24. None. 2001. Can J Ophthalmol. A randomized trial 5. 21. consistent. 1989. and retinal circulation on optic nerve head and retinal nerve 29. Leske MC.14:166–172. 2009. 20. Am J Ophthalmol.43:328–336.e2. et al. et al. Nocturnal systemic from the David E. Arch Ophthalmol. Eye and Ear Infirmary of Mount Sinai.What patterns identify patients at risk for 4.107:1287–1293. Lietman T. et al.7:93–98. Rechtman E. Optic disc fibre layer structure in patients with open-angle glaucoma over progression and rates of visual field change in treated an 18-month period. Bandi JR. Kaiser HJ. Drance SM. De Moraes CG. Friedman DS.90:262– Ophthalmologica. Risk factors for open. Incident open-angle OCT-A to investigate different patterns of glaucomatous glaucoma and blood pressure. Systemic blood pressure in Panarelli. et al.2014:514948. Charlson ME. et al. discussion 1311–1314. USA.117:603–624. 9 j OCT619 assessing glaucoma.A. Abadia B. Ophthalmol. Arch Ophthalmol. Agemy. OCT Angiography Peripapillary Analysis in Glaucoma IOVS j Special Issue j Vol. Gold JP. 2008. Wijsman K. automated perimetry in healthy and glaucoma patients. de Moraes CG.P. Association of blood Biomed Res Int. Alencar LM. Scripsema. domain optical coherence tomography for glaucoma. Prevent Blindness. Clarity (C). 1995. Harris A. outcomes after coronary artery bypass. and characterization of glaucomatous nerve glaucoma in a population-based study of Hispanic subjects: changes according to pathophysiological features. Inc. localized visual field loss before and after disc 12. R. The role of retrobulbar 2011. Warheit-Roberts L. and intraoperative hemodynamic predictors of postoperative Ophthalmic Epidemiol. Am J Ophthalmol. Ophthalmic Epidemiol. Chen TC.113:918–924. L. Glaucoma risk factors observed in the Baltimore hemorrhage.50:4727–4733. Opticology (F). et al. Spatially based assessment. Am J Ophthalmol. Ambulatory blood Disclosure: N. Important causes of noncardiac surgery. 2012. Graham SL. Ann Surg. J Supported in part by an unrestricted grant from Research to Glaucoma. Hayreh SS. Sidoti.120: changes in the peripapillary retina. MacKenzie CR. Zimmerman MB. et al. Garcia. Liebmann CA. None. S.212:567–580.142:144–145. 2009. Downloaded From: http://iovs. Tsai. Open-angle glaucoma – an epidemiologic overview. 2013. Artes PH. Diagnostic capability of spectral- subjects: the Thessaloniki eye study. 8. and primary open-angle glaucoma. Ophthalmol Clin 26. et al. Vascular risk factors 27.ashx?url=/data/journals/iovs/935468/ on 03/27/2018 . Am J 142:60–67. Sung KR. et al. Ocata Medical (C). 33.118:60–65. Lee AJ.127:1250–1256. Br J Ophthalmol. 17. Bonomi L. Quigley HA. P. Calvo P. Charlson ME.210:637–648. Arch Ophthalmol. None. Risk factors 30. angle glaucoma.44:2613– 7.50: and intraobserver reproducibility of the subjective determina- 5266–5274. New York. 2002. Nocturnal arterial hypotension and its role in optic nerve head and ocular References ischemic disorders. Herman Peters Pediatric Glaucoma Research Fund of the New York Ophthalmology.C.F. Acta Ophthalmol. Nemesure B. de Boer JF. Invest Ophthalmol Vis Sci. Tielsch JM. Curr Opin Ophthalmol. visual impairment in the world today. Invest Ophthalmol Vis Sci.153:815–826. et al. et al. Mitchell P.D. The preoperative 2. The role of optic nerve postoperative complications? Ann Surg. Y. Gold JP. Budenz DL. Invest Ophthalmol Vis Sci. Leske MC. Twenty-four hour ocular 2620. et al. et al. S. Marraffa M. et al.102:61–69. 2015. Lee S. 16. Regeneron (C) 22. None. NY.M. Krawitz. Chauhan BC. blood pressure.154:702–711. Incidence and rates of for visual field progression in the low-pressure glaucoma visual field progression after longitudinally measured optic treatment study. Ferreras A. et al. Mo. Marchini G.org/pdfaccess. Zangwill LM. Kagemann L. 2006. Topouzis F. Wu H. et al.110:1302–1311. Prata TS. Hayreh SS. None. worldwide in 2010 and 2020. perfusion pressure fluctuation and risk of normal-tension 28. Broman AT. pressure monitoring in glaucoma. et al. Nicolela MT. glaucoma. P. The prevalence of monitoring. None.D. None. et al. Leske MC. Association of blood pressure status with the optic spectral-domain optical coherence tomography and standard disk structure. 2004. Flammer J. Gold JP. blood flow in the pathogenesis of glaucoma. myocardial infarction or ischemia in patients undergoing 3. Open-angle glaucoma suggestions and comments. Interobserver agreement glaucoma progression. 2003. Gillespie BW. Nano Retina (C). 1995. J. These findings are the first step in using 15. 1995. 2118. necessary to determine the role of OCT-A in the detection. Leske MC. The number of people with glaucoma arterial hypotension in optic nerve head ischemic disorders. Coleman AL. retest variability.13:319–326.121:2004–2012. pressure status with the optic disk structure in non-glaucoma 34. Charlson ME. 231:677–680. Tobe LA.B. Lin. 14. 2007. 2009. Graf T.99:609–612. 2006. Ehrlich R. De Moraes CG. Relationship between 13. Arch Ophthalmol. 2003. et al. Park SB. None. pressure. functional loss in glaucoma from progressive optic disc 11. Am J Ophthalmol. The collaborative for primary open angle glaucoma: the Egna-Neumarkt Study. Optovue (C). Wu SY. Eye Survey. The Barbados Eye Study.arvojournals. Rodriguez J. disc change in glaucoma.116:2110– 10. Tanna AP. 23. None.119:1819–1826. Link A. A population. Harris A. 57 j No. 1996. Improvement of North Am. Rosen.3:85–96. Jonas JB. Oph- Bavier. The authors thank the two anonymous reviewers for their helpful 18. De Moraes CG. 2014. Measuring and comparing intraoperative high versus low mean arterial interpreting ocular blood flow and metabolism in glaucoma. Greenfield DS.290:2057– 25.91:e86–e91. et al. Liebmann JM. None. et al. 2009. Ophthalmology. perfusion damage. JAMA. J Thorac Cardiovasc Surg. Intraoperative 2060. None. Siesky B. 2005. et al.Y. Sommer A. Park SC. Role of nocturnal 1. 32. Harris A. R.A. Williams-Russo P. tion of glaucomatous visual field progression. J. 1994. 2014. 6. Rochtchina E. Xu. Prediction of 1995. Harris A. None.. Sommer A. thalmology. Br J Ophthalmol. 1993. 2012. Quigley HA. Liebmann JM.K. PAS receives support 19. 1996. The nocturnal dip. 2006. Guire KE. 2000.

e1. Leung CK. Williams ZY. Am J 1557–1564. So KF.110:2185–2191. Biomed Opt pressure and perfusion pressure in patients with glaucoma or Express. et al. Lasta M. Kagemann L. Br J Ophthalmol. Singh K.121:1322–1332. Moghimi S. PLoS One. placebo- 40. Funaki S. Schuman JS. reproducibility of a two-dimensional mapping of the optic 49. et al. Jiang C. Scripsema N. Tokayer J. Graefes Arch Clin Exp Ophthalmol. et al. perfusion areas in peripapillary retina. Tanga L. 2014.51171–51176. et al. Ophthalmology. Invest Ophthalmol Vis Sci. 2014. Anderson DR.160:35–44. Retinal nerve fiber layer 1337–1340. Yaoeda K. et al. Agemy S. Shahlaee A. 1993:52–61. Yu M. Jia Y. 2016. et al. Aizawa N. 2012. Br J Ophthalmol. Eur J Ophthalmol. Parrish RK II. imaging with spectral-domain optical coherence tomography: mology. et al. Ophthal. imaged by fluorescein angiography and optical coherence 66. 2015. et al. et al. Weinreb RN. Effects of topical 39.ashx?url=/data/journals/iovs/935468/ on 03/27/2018 . JAMA Ophthalmol.56:3212–3217. Shirakashi M. Bagheri N. Optic nerve head and retinal angiography of peripapillary retinal blood flow response to nerve fiber layer analysis: a report by the American Academy hyperoxia. 134:538–546. Ishibazawa A. 2015. St. Chung HS. ocular hypertension: a randomized. 56. 67.94:103–104. Polak K. Meta-analysis of timolol on raphy in normals and diabetic retinopathy patients. phy. Jia Y. 42. Iester M.arvojournals. 9 j OCT620 35.118:241–248. Tadarati M. Lee VW. Oakley JD. Macular vascular abnormalities short-wavelength perimetry abnormalities. density mapping using optical coherence tomography angiog. et al. patients with sickle cell disease. et al. et al. Rossetti L. Morrison JC. Jia Y. 2015. Br J optical coherence tomography volumes on a per A-scan basis Ophthalmol. 2012. 2015.98:1368–1373. J Glaucoma. Opt Express. Spaide RF. Shah C. Retinal nerve fiber layer angiography of optic disc perfusion in glaucoma. Motion correction in flowmetry: factors affecting blood flow measurement.114:1937–1949. Quantitative optical coherence speckle flowgraphy. Optical coherence tomography 58. Michelson G. Wei E. Chinese: an optical coherence tomography angiogram study. Optical coherence tomography ogy. 57 j No. Retinal vascular perfusion controlled clinical trial.5% on circadian IOP. Stewart JA. 2002. tive pilot study. 2007.87:875–880. coherence tomography angiography of vascular abnormalities 2011. Altieri M. Asano T. Wang X.82:131–136. Acta Ophthalmol. Takahashi A. et al. et al. et al. 2012. Reproducibility tomography angiography. blood angiography of optic nerve head blood flow. JAMA Ophthalmol. 2011. Kunikata H. Sridhar J. Nagaoka T. et al.119:1858–1866. Clin Ophthalmol. Louis: The CV Mosby Co. Wilson R.253: flowgraphy and scanning laser Doppler flowmetry. 2004. Chiba N. Mwanza JC. Klancnik JM Jr. angiography of optic nerve head and parafovea in multiple 61. Nassiri N.35:2353–2363. Ophthalmol- 44. of Ophthalmology. Ophthalmol. Downloaded From: http://iovs. Twelve-hour reproducibility sclerosis. et al. Quantitative optical flowgraphy-NAVI in patients with glaucoma.88:1266–1269. thickness measured with optical coherence tomography is 54.53: 52. Heidelberg retinal 51. Boden C. 2015. OCT Angiography Peripapillary Analysis in Glaucoma IOVS j Special Issue j Vol.org/pdfaccess. Cooney MJ. Spain R.129:734–739. 2012. 2003. Lee PW. a E2395–E2402. Retinal vascular layers nerve head perfusion.133:1045–1052. Bowd C. Optical coherence tomography 60. 2015.112: 64. Split-spectrum amplitude. et al.133: 37. decorrelation angiography with optical coherence tomogra. Structure-function using orthogonal scan patterns. Ophthalmology. Optical coherence 59.20:4710–4725. Shiga Y. Jiang C. 2015. En face optical related to visual function in glaucomatous eyes. Optical coherence 55.11:488–492. Gamell L. Correlation between optic disc individuals.20:1035–1041. Yokoyama Y. Liu L. 2010. Budenz DL. by optical coherence tomography in glaucoma suspects with 53. Am J Ophthalmol. 43. et al. Tokayer J. Bailey ST. Retina. et al. Biomed Opt Express.3:3127–3137. Bailey ST. JAMA Ophthalmol.55: tomography angiography of choroidal neovascularization in 3899–3904. 2014. et al. Scott AW. Wang X. coherence tomography and angiography of talc retinopathy. diurnal and nighttime intraocular pressure and blood pressure. et al. Proc Natl Acad Sci U S A. Luksch A. age-related macular degeneration. Ability of cirrus HD- tomography angiography of the peripapillary retina in OCT optic nerve head parameters to discriminate normal from glaucoma. 2009. Ko T. Wilson DJ.133:45–50. Doyle A. Quantitative OCT bimatoprost 0. Oddone F. Mayer MA. 2003. et al. Macular perfusion in healthy 2740–2748. ogy. et al. microcirculation in the optic nerve head by laser speckle phy study. Liu L. et al. Jia Y. Jia Y. Lin SC. Mok KH. 2014. J Glaucoma. Jampel HD. of the Heidelberg retinal flowmeter in determining low 50. Am J Ophthalmol. Takusagawa HL. Kraus MF. Jia Y. 2000. patterns of retinal nerve fiber layer progression. Retinal nerve fiber layer measurement Invest Ophthalmol Vis Sci. 2012. 41. 2015. 1998. Hwang TS. Jonescu-Cuypers CP. Ophthalmol. Potsaid B. 57. Intraobserver 1435–1444. Jia Y. 62. double masked. novel blood flow index in the human retina derived from laser 48. perfusion and glaucomatous severity in patients with open. Harris A. of retinal and optic nerve blood flow parameters in healthy 45.01% and timolol 0. 36. Reproducibility of glaucoma. Invest Ophthalmol Vis Sci. 2002. Relative flow volume. El Beltagi TA. et al. 2015. achromatic perimetry. Clinical decisions in 63.121: 65. retinal circulation measurements obtained using laser speckle 47. Han IC. Jia Y. 38. identified by optical coherence tomographic angiography in 12:45–49. Optical coherence tomography measurement of nerve fiber layer thickness and tomography angiography in diabetic retinopathy: a prospec- the likelihood of a visual field defect.10:e0140601. et al. 46. Ophthalmology.56:3287–3291. et al. et al. in the living human eye. Yu J. Harris A.3: relationships between spectral-domain OCT and standard 1182–1199. Invest Ophthalmol Vis Sci. Acta Ophthalmol.e1. Hodapp EA. Nilforushan N. Wang X. Wang X. Tan O. Measurement of angle glaucoma: an optical coherence tomography angiogra. glaucomatous eyes. Pechauer AD. 2015.