Vascular Biomarkers from Optical Coherence Tomography Angiography and Glaucoma: Where do we stand in 2021?

Abstract

Biomarkers of ocular blood flow originating from a wide variety of imaging modalities have been associated with glaucoma onset and progression for many decades. Advancements in imaging platforms including optical coherence tomography angiography (OCTA) have provided the ability to quantify vascular changes in glaucoma patients, alongside traditional measures such as retinal nerve fiber layer thickness (RNFL) and optic nerve head (ONH) structure. Current literature on vascular biomarkers, as measured by OCTA, indicates significant relationships between glaucoma and blood flow and capillary density in the retina and ONH. The data currently available, however, is highly diverse and lacks robust longitudinal data on OCTA vascular outcomes and glaucoma progression. Herein we discuss and summarize the relevant current literature on OCTA vascular biomarkers and glaucoma reviewed from December 14, 2020 through March 1, 2021. Associations between OCTA vascular biomarkers and clinical structural and functional glaucoma outcomes as well as differences between glaucoma patients and healthy controls are reviewed and summarized. The available data identifies significantly decreased flow density, flow index, and vessel density in the ONH, peripapillary vascular layer, and macula of glaucoma patients compared to controls. Whole image vessel density is also significantly decreased in glaucoma patients compared to controls and this outcome has been found to correspond to severity of visual field loss. OCTA vascular biomarkers alongside clinical structural outcomes may aid in assessing overall risk for glaucoma in patients.

INTRODUCTION:

Open-angle glaucoma (OAG) is a leading cause of irreversible blindness, affecting over 70 million people worldwide (Weinreb et al., 2014). Although our understanding of the disease has significantly evolved, it was traditionally characterized as an optic neuropathy with retinal ganglion cell death related to a high intraocular pressure (IOP) (Weinreb et al., 2014). However, despite reduction of IOP or it being within normal limits, many patients experience disease onset and continued progression. In certain patients, vascular changes in the retina and optic nerve head have been linked with glaucoma onset and progression of the disease (Weinreb and Harris, 2009). Despite its relevance, a lack of gold standard in ocular vascular imaging and historically invasive, highly technical, and fragmented approach have prevented the widespread adoption of vascular biomarkers in glaucoma management.

Due to OCTA’s ability to accurately detect clinical structural changes alongside ocular blood flow biomarkers, it offers a novel approach to aid in diagnosis and management of glaucoma. Within the current literature, studies have investigated using OCTA to detect various types of glaucoma, including exfoliation glaucoma, normal-tension glaucoma, angle-closure glaucoma, and OAG (Lommatzsch et al., 2019; Hou et al. 2020). This review summarizes the relevant literature on the use of OCTA vascular biomarkers in glaucoma and discusses their specific meaning and use in determining overall risk for glaucoma onset and progression.

MATERIAL AND METHODS:

Optical coherence tomography angiography (OCTA) is an imaging platform that is non-invasive with a dye-less modality to detect endoluminal blood flow changes within the retina and ONH (Kromer et al., 2019). Previously used imaging modalities to evaluate blood flow in the retina include highly invasive fluorescein angiography (FA) and indocyanine green angiography (ICGA) (de Carlo et al., 2015). While FA remains valuable in identifying choroidal and vascular neovascularization and leaks, it is not an ideal modality for localizing decreased blood flow to a specific layer of the eye and can even obscure certain pathologies (de Carlo et al., 2015). Other limitations include the invasive nature of the procedure (dye is injected systemically into a vein), cost, and time spent performing the procedure (de Carlo et al., 2015). There is also a small risk of anaphylaxis, as well as specific contraindications for ICGA (de Carlo et al., 2015). Advantages of OCTA include its non-invasive nature, the collection of both structural and vascular data simultaneously, and its ability to analyze the images in cross-section to better localize the pathology (de Carlo et al., 2015; Spaide et al., 2018). Classic Doppler OCT had major limitations in directly measuring blood flow due to blood flow being the perpendicular direction of the OCT beams (Spaide et al., 2018). Blood velocity can be determined but only through complex calculations, including the knowledge of the Doppler angle, and the angle between the light beam and blood vessel (Spaide et al., 2018). Comparatively, OCTA can capture the microvascular structure like a classic OCT device while also localizing the depth of the lesion and affected capillary layer, making it unique with this dual functionality (Spaide et al., 2018).

Specifically, OCTA compares sequential B-scans acquired at a single location within the eye (Wan & Leung, 2017). As static structures would not be expected to change, changes between sequential images are indicative of blood flow in the region of interest (Wan & Leung, 2017). The main vascular outcomes that can be measured include: vessel density (% of detected vessel area/total imaged area), flow index (average decorrelation signal reported as a number between 0 and 1), and blood flux index (mean flow intensity in vessel area divided by full dynamic range of blood flow signal intensity normalized to a number between 0 and 1) (Wan & Leung, 2017). There are additionally algorithms such as split-spectrum decorrelation angiography (SSADA) and optical microangiography (OMAG) based on OCTA B-scan images to approximate blood flow parameters (Wan & Leung, 2017). In 2017, available OCTA devices include models from Optovue, Spectralis, Heidelberg, Zeiss (Munk et al., 2017). In a head-to-head comparison, researchers found no significant differences in vessel density measurements, but found differences in number of vessel bifurcations to be statistically significant (p≤0.001) (Munk et al., 2017).

A comprehensive literature search was conducted from December 14, 2020 through March 1, 2021 on PubMed, Medline, Google Scholar, and associated digital platforms with the following keywords: OCTA, AngioOCT, vascular, blood flow, vascular imaging, retina, glaucoma, hemodynamics, biomarkers, vessel density, ocular perfusion pressure, perfusion, vascular and vessels. The associated articles were also searched and cross-referenced for relevant citations. There were no publication year restrictions. Review papers and non-English articles were excluded, as well as studies using OCTA only to capture structural measurements, and not vascular biomarkers of the eye. The age of subjects for included publications did not significantly differ. After applying the inclusion and exclusion criteria, 28 studies were eligible for this review. Studies are grouped into categories by locality of vascular measure, specifically peripapillary and circumpapillary and optic nerve head (ONH), macula and fovea, and choroid.

RESULTS:

Optic Nerve Head & Peripapillary and Circumpapillary Capillary Layer OCTA Vascular Biomarkers

OCTA vascular biomarkers of the optic nerve head and peripapillary and circumpapillary capillary layer are found in Table 1. Specifically, Jia et. al (2012) found that in preperimetric glaucomatous (PPG) eyes, whole disc flow index was significantly reduced (35%) compared to healthy controls.⁹ The flow index was also greatly reduced in the temporal ellipse of the disc (57%) in PPG eyes compared to controls (Jia et al., 2012). Jia and colleagues also found a significant decrease in vessel density in both the whole disc area and temporal ellipse of the disc by 34% and 57% respectively (Jia et al., 2012). Jia et al. (2014) using OCTA and a SSADA further demonstrated a decreased blood flow index in glaucomatous eyes in the area of the optic disc (Jia et al., 2014). Vessel density loss in glaucomatous eyes was also demonstrated in the optic nerve head by Lévêque et al. (2016) showing a decrease in both the total and temporal vessel density 24.7% and 22.88% respectively (Lévêque et al., 2016).

Table 1:

Peripapillary/circumpapillary/optic nerve head (ONH) optical coherence tomography angiography (OCTA) Vascular Biomarkers. OAG: open-angle glaucoma.

Author (year)	Patients	OCTA Scan	Results	P-value
Jia et al. (2012)	4 preperimetric glaucoma (PPG) patients 4 healthy subjects	Whole disc	Whole disc flow index was reduced by 35% in the PPG group	P=0.040
			Whole disc vessel area was reduced by 34% in the PPG group	P=0.045
		Regional disc perfusion based the temporal ellipse of disc	Temporal ellipse of disc flow index was reduced by 57% in the PPG group	P=0.010
			Temporal ellipse of disc vessel density was reduced by 57% in the PPG group	P=0.013
Jia et al. (2014)	11 glaucoma patients 24 healthy subjects	3 × 3 mm, using SSADA algorithm	Disc flow index was reduced by 25% in the glaucoma group	P=0.003
Lévêque et al. (2016)	50 glaucoma patients 30 normal subjects	3 × 3 mm ONH	Total ONH vessel density was reduced by 24.7% in glaucoma group compared with control group	P<0.0001
			Temporal disc vessel density was significantly reduced by 22.88% in the glaucoma group compared to the control group.	P=0.001
Liu et al. (2015)	12 OAG 12 Healthy subjects	3 × 3-mm ONH	Peripapillary flow density was significantly lower in glaucomatous eyes compared to control	P<.001
			Peripapillary vessel density was significantly lower in glaucomatous eyes compared to normal controls	P<.001
Chen et al. (2016)	42 OAG subjects 26 glaucoma suspect subjects 20 normal subjects	OMAG, OCT-A 2.4 × 2.4 mm	Eyes from OAG subjects and glaucoma suspects showed significantly lower blood flux index compared with normal eyes	P≤0.0015
Chen et al. (2020)	47 OAG 36 normal participants	ONH 4.5 × 4.5 mm	OAG patients had significantly more areas of low perfusion in the peripapillary retina compared to normal controls	P<0.001
			Focal perfusion loss was greater in OAG patients than normal controls	P<0.001
Chen et al. (2017)	26 OAG 27 healthy subjects	4.5 × 4.5 mm ONH 6.0 × 6.0 mm Macula	In OAG whole image vessel density was significantly lower than healthy controls in the macular and peripapillary areas	P<0.001
			Circumpapillary vessel density was significantly decreased compared to controls	P<0.001
Mammo et al. (2016)	5 subjects with unilateral glaucoma 3 glaucoma suspect patients 9 normal subjects	2 × 2 mm ONH	Vessel density in the radial peripapillary capillary network was significantly lower in glaucomatous eyes compared to matched regions in the non-affected eye, glaucoma suspect, and normal group	P<.001 for all groups
Kim et al. (2017)	13 patients (unilateral perimetric normal tension glaucoma, fellow eye preperimetric normal tension glaucoma) 9 healthy controls	4.5 × 4.5 mm ONH	Vessel density in the whole peripapillary region in eyes with perimetric glaucoma were significantly lower than preperimetric glaucomatous eyes	P=0.001
			Vessel density in the inferotemporal region in perimetric glaucoma eyes were significantly lower than preperimetric glaucomatous eyes	P<0.001
Triolo et al. (2017)	40 normal controls 40 glaucoma suspect 40 glaucoma	6 × 6-mm Fovea, ONH	Average Peripapillary vessel perfusion density was significantly decreased in the glaucoma group	P≤0.001
			Peripapillary vessel perfusion density was decreased in the superior and inferior quadrants in the glaucoma group	P≤0.001
Mansoori et al. (2017)	24 OAG 52 normal subjects	3.45 × 3.45 mm ONH using Bar-Selective Combination of Shifted Filter Responses (B-COSFIRE)	Radial peripapillary capillary density was decreased in the inferotemporal sector compared to normal controls	P=0.002
			Radial peripapillary capillary density was decreased in the superotemporal sector compared to normal controls	P=0.008
Akil et al. (2017)	20 mild OAG 20 pre-perimetric glaucoma 16 controls	3 × 3 mm ONH using 1050 nm wavelength swept source OCT	Optic nerve head vessel density was significantly decreased in pre-perimetric glaucoma patients compared to controls	P=0.0002
			Peripapillary density was significantly decreased in pre-perimetric glaucoma patients compared controls	P=0.007
			Optic nerve head vessel density was significantly decreased in OAG glaucoma compared to pre-perimetric glaucoma patients	P=0.002
			Peripapillary vessel density was significantly decreased in OAG patients compared to pre-perimetric glaucoma patients	P=0.001
Lu et al. (2020)	41 normal 44 pre-perimetric OAG (PPG) 42 early perimetric OAG (EG)	ONH 4.5 mm × 4.5 mm Radial peripapillary density measured using 750 μm wide elliptical annulus around disc	Inferior-temporal radial peripapillary capillary vessel density was lower in the EG group compared to controls.	P<.05
Yarmohammadi et al. (2016)	23 healthy subjects 37 glaucoma suspect 104 OAG	4.5 × 4.5-mm ONH	Circumpapillary vessel density was significantly lower in the OAG group compared to glaucoma suspect compared to healthy controls (55.1 vs. 60.3% vs. 64.2)	P<0.001
			Whole image capillary density was significantly lower in the OAG group compared to glaucoma suspect compared to healthy controls (46.2% vs. 51.3% vs. 56.6%)	P<0.001
Yarmohammadi et al. (2016)	31 healthy subjects 48 glaucoma suspect 74 glaucoma patients	4.5 × 4.5 mm ONH	Whole image vessel density was highest in healthy subjects followed by glaucoma suspects, then mild glaucoma, lastly moderate to severe glaucoma (55.5% vs. 51.3% vs. 48.3% vs. 41.7%)	P<0.001
		750-μm-wide elliptical annulus around the optic disc	Circumpapillary vessel density was highest in healthy controls and lowest in severe glaucomatous eyes (62.8% vs. 49.6%)	P<0.001
Yarmohammadi et al. (2018)	33 healthy subjects 33 OAG with a VF defect in one eye and normal VF in other eye	4.5 × 4.5 mm RNFL 3 × 3 mm Fovea	In OAG patients, whole image vessel density in the perimetrically unaffected eye (52.0%) was greater than in the perimetrically affected eye (48.8%) Whole image vessel density was greatest in normal subjects (55.9%)	P<0.001
Jesus et al. (2019)	40 healthy subjects 82 glaucoma patients	3 × 3 mm optic disc-centered	Significantly lower circumpapillary microvascular density in glaucoma group compared to healthy group	P<0.001

Glaucomatous vascular changes were also found in the peripapillary capillary layer. Liu et al. (2015) showed a significant decrease in both the flow density and vessel density of the peripapillary layer in glaucomatous eyes compared to healthy controls (Liu et al., 2015). Using OMAG, Chen et al. (2016) found a lower blood flux index in the peripapillary layer in glaucomatous eyes compared to healthy controls (Chen et al., 2016). Chen et al. (2017) also found decreased vessel density in both the macular and peripapillary areas as well as in the whole image in glaucomatous eyes (Chen et al., 2017). Furthermore, circumpapillary vessel density was significantly decreased in glaucomatous eyes (Chen et al., 2017). Chen et al. (2020) demonstrated that eyes with POAG had more areas of low perfusion in the peripapillary layer compared to controls and greater areas of focal perfusion loss (Chen et al., 2020). Notably, Mammo et al. (2016) found that vessel density was significantly decreased in the radial peripapillary layer in the glaucomatous eye of patients with unilateral glaucoma compared to the unaffected eye, glaucoma suspects, and healthy controls (Mammo et al., 2016). Similar results by Kim et al. (2017) showed patients with unilateral perimetric glaucoma had significantly greater vessel density loss in the peripapillary and inferotemporal capillary beds compared to the fellow eye with preperimetric glaucoma (Kim et al., 2017).

These results show that even within a single patient, OCTA vascular progression can be different for each eye. Triolo et al. (2017) compared 40 normal subjects, 40 glaucoma suspects, and 40 glaucoma patients, and found that peripapillary vessel perfusion density decreased in the superior and inferior quadrants in glaucoma patients (Triolo et al., 2017). The average peripapillary vessel perfusion density was also decreased (Triolo et al., 2017). Similar findings by Mansoori et al. (2017) showed a significant decrease in the radial peripapillary capillary density in the inferotemporal and superotemporal regions compared to normal healthy controls (Mansoori et al., 2017). Akil et al. (2017) also demonstrated a significantly greater vessel density loss in the optic nerve head and peripapillary region in OAG patients compared to pre-perimetric glaucoma patients, as well as a significant decrease between pre-perimetric glaucoma and healthy controls (Akil et al., 2017). Lu et al. (2020) also demonstrated a significant decrease in vessel density of the inferior-temporal radial peripapillary capillary layer in early perimetric OAG patients compared to controls (Lu et al., 2020).

Four papers were included in the review that examined circumpapillary vessel density and whole image capillary density (Yarmohammadi et al., 2016a; Yarmohammadi et al., 2016b; Yarmohammadi et al., 2018; Jesus et al., 2019). Yarmohammadi et al. (2016) found that circumpapillary vessel density and whole image capillary density were significantly lower in eyes with OAG compared to glaucoma suspect and healthy controls (Yarmohammadi et al., 2016a). These measurements had similar diagnostic accuracy as retinal nerve fiber layer thickness in differentiating between glaucoma and healthy eyes. Another study by Yarmohammadi and colleagues (2016) studied the associations between severity of visual field loss and vessel density, specifically circumpapillary and whole-image, measured by OCTA in subjects with OAG (Yarmohammadi et al., 2016b). Their findings showed that eyes with severe glaucoma had the lowest circumpapillary vessel density while healthy controls had the highest (49.6% vs. 62.8%, p<0.001) (Yarmohammadi et al., 2016b). Similarly, Jesus et al. (2019) found significantly lower circumpapillary microvasculature density (cpmVC) among glaucoma patients compared to healthy controls (p<0.001), as well as a reduction in cpmVD variation in glaucoma patients (Jesus et al., 2019). Yarmohammadi et al., 2016b also found that whole-image vessel density decreased as the severity of glaucoma increased (healthy, then glaucoma suspect, then mild glaucoma, lastly moderate to severe glaucoma). These findings are consistent with a study conducted by Yarmohammadi et al. (2018), which studied OAG patients with a visual field defect in one eye and normal visual field in the other eye, compared to healthy controls (Yarmohammadi et al., 2018). They found that whole-image vessel density in the unaffected eye (52.0%) was greater than in the affected eye (48.8%) in OAG patients and that whole image vessel density was greatest in normal subjects (55.9%) (Yarmohammadi et al., 2018). These results demonstrate a significantly decreased vessel density in glaucomatous eyes in the area of the optic nerve head, and peripapillary and circumpapillary vascular layer as measured by OCTA.

Macula, Fovea and Perifoveal OCT-A Vascular Biomarkers

Although glaucoma is most known to cause peripheral vision loss, damage can occur to retinal ganglion cells residing in the macula. Therefore, studying macula and fovea can provide crucial evidence for the detection of glaucoma (Takusagawa et al., 2017). Relevant studies measuring macula, fovea and perifoveal vascular biomarkers for glaucoma are listed in Table 2. Takusagawa et al. (2017) used OCTA to detect macular perfusion defects in 30 subjects with perimetric glaucoma and 30 age-matched normal participants. They found that the superficial vascular complex (SVC) and all-plexus retinal vessel density were lower in the glaucoma group compared to the normal group (p<0.001 for both) and that among all macular vessel density parameters studied, the SVC vessel density showed the best diagnostic accuracy (Takusagawa et al., 2017). However, Takusagawa et al. (2017) did not find a significant difference in vessel density within the intermediate or deep capillary plexuses. Choi et al. (2017) also studied macular vessel density, as well as the foveal avascular zone (FAZ), in 52 patients with OAG and 52 healthy controls (Choi et al., 2017).

Table 2:

Macula, Fovea and Perifoveal optical coherence tomography angiography (OCTA) Vascular Biomarkers. OAG: open-angle glaucoma.

Author (year)	Patients	OCTA Scan	Results	P-value
Takusagawa et al. (2017)	30 perimetric glaucoma 30 age-matched normal subjects	6 × 6 mm macular	Superficial vascular complex and all-plexus retinal vessel density were lower in the glaucoma group than the normal group	P<0.001 for both
Kromer et al. (2019)	30 OAG 21 healthy subjects	5.0 × 3.5 mm Perifoveal	Macular flow density was globally reduced in glaucoma patients in the superficial plexus	P=0.0203
			Macular flow density was nasally reduced in the superficial plexus	P=0.0003
			Macular flow density was globally reduced in the deep retinal plexus	P=0.0113
			Macular flow density was nasally reduced in the deep retinal plexus	P<0.0001
Baek et al. (2019)	20 OAG patients 19 healthy subjects	Macula 4.5 × 4.5 mm	Significantly greater diurnal changes in OAG in peripapillary retinal vessel density compared to healthy controls	P=0.013
			Significantly greater diurnal changes in OAG in macular RVD compared to healthy controls	P=0.042
Choi et al. (2017)	52 patients with primary open angle glaucoma 52 healthy participants	Macula 3 × 3 mm	Macular vessel density was lower in OAG patients compared to normal controls in the superficial layer	P=0.013
			Macula vessel density was lower in OAG in the deep layer	P<0.001
			Macula vessel density was lower in OAG in the whole retina	P=0.002
			Increased foveal avascular zone perimeter and decreased foveal avascular zone circularity index in glaucoma patients compared to control	P<0.001
Shoji et al. (2017)	38 healthy controls 30 glaucoma suspect 32 OAG	3.0 × 3.0 mm Macula	Macula vessel density was greatest in healthy eyes, then glaucoma-suspect, and lowest in OAG	P<.01
			Rate of change in macula vessel density was significantly faster in OAG (-2.23%/y) compared to glaucoma suspect eyes (0.87%/y)	P=.001
			Rate of change in macula vessel density was significantly faster in OAG compared to healthy eyes (0.29%/y)	P=.004
			Rate of vessel density loss in the superior and inferior regions of the macula was significantly faster in glaucoma patients compared with healthy controls or glaucoma suspect eyes	P<.05
Chao et al. (2019)	18 OAG 14 NTG 18 OHT 20 normal subjects	3.0 × 3.0 mm Macula	OAG and NTG had a significantly lower superficial vessel density compared to healthy controls	P<0.01
			NTG showed a significantly lower deep vessel density compared to healthy controls	P<0.01
			NTG group had a larger foveal avascular zone (FAZ) compared to the OHT group	P<0.01
			OAG group had lower flow area in the outer retina compared to the OHT group and control group	P<0.01
Lommatzsch et. al. (2018)	85 eyes with glaucoma 50 healthy controls	6 × 6 mm macula	Macular vessel density was lower in superficial retinal vascular plexus	P<0.0001
			Macular vessel density was lower in deep retinal vascular plexus	P=0.009
Lu et. al (2020)	41 normal 44 pre-perimetric OAG (PPG) 42 early perimetric OAG (EG)	6 × 6 mm macula 4.5 × 4.5 mm ONH Radial peripapillary density measured using 750 μm wide elliptical annulus around disc	Parafoveal vessel density was lower in the temporal quadrant in PPG compared to normal controls	P<0.044
			Perifoveal vessel density was lower in the PPG group compared to normal controls	P<0.05
			Both parafoveal and perifoveal vessel density was lower in the EG group compared to normal controls	P<0.001
Milani et. al (2020)	35 healthy controls 41 ocular hypertension (HTN) 24 OAG	3 × 3 mm Macula	Mean whole vessel density in the macular superficial capillary plexus was highest in normal controls, followed by HTN, followed by OAG––during the morning (51.337 vs. 50.781 vs. 45.726 respectively)	P<0.0001
			Mean whole vessel density in the macular superficial capillary plexus was highest in normal controls, followed by HTN, followed by OAG during the evening (51.504 vs. 50.812 vs. 46.069 respectively)	P<0.0001
			Mean parafoveal vessel density in the macular superficial capillary plexus was highest in normal controls, followed by HTN, followed by OAG during the morning (54.052 vs. 53.472 vs. 49.687 respectively)	P=0.0005
			Mean parafoveal vessel density in the macular superficial capillary plexus was highest in normal controls, followed by HTN, followed by OAG during the evening (54.037 vs. 53.300 vs. 50.154 respectively)	P=0.0022

Unlike Takusagawa et al. (2017), Choi and colleagues found that macular vessel density was significantly lower in OAG patients in the superficial layer, deep layer, and whole retina. (Choi et al., 2017). This finding is consistent with Kromer et al. (2019), who found that macular flow density was globally and nasally reduced in glaucoma patients in both the superficial and deep retinal plexus compared to healthy controls (Kromer et al., 2019). These results also align with Lommatzsch et al. (2018), who studied macular vessel density in glaucoma compared to healthy control eyes, and found that both superior and deep retinal vascular plexus densities were lower in glaucomatous eyes than healthy eyes (Lommatzsch et al., 2018). Chao et al. (2019) found a similar decrease in the superficial capillary vessel density in OAG patients and normal tension glaucoma patients (NTG) compared to healthy controls (Chao et al., 2019). Chao et al. (2019) also reported a decrease in the deep capillary layers in NTG compared to controls (Chao et al., 2019). Notably, a significantly larger FAZ was found in NTG compared to the ocular hypertension group (OHT) (Chao et al., 2019). Lu et al. (2020) compared perifoveal and parafoveal density in healthy controls, PPG patients, and early perimetric OAG patients (Lu et al., 2020). Lu et al. found a significant decrease in the vessel density of the temporal quadrant of the parafoveal layer in PPG patients compared to healthy controls (Lu et al., 2020). Additionally, PPG patients also had a significantly decreased perifoveal vessel density compared to controls (Lu et al., 2020). In OAG patients, there was a significant decrease in the vessel density in both the parafoveal and perifoveal capillary layers compared to healthy controls, demonstrating a correlation between vessel density loss and increased severity of disease (Lu et al., 2020). Milani et al. (2021) compared healthy controls, ocular hypertension patients, and OAG (Milani et al., 2021). Milani et al. (2021) found a significant decrease in whole vessel density and parafoveal vessel density in the macular superficial capillary plexus in OAG compared to ocular hypertension patients and normal controls (Milani et al., 2021). This decrease in vessel density was measured both in the morning and in the evening demonstrating a consistent decrease in vessel density without diurnal variation (Milani et al., 2021).

In terms of the FAZ, Choi et al. (2017) found diagnostic value in the FAZ circularity index and perimeter. Specifically, eyes with glaucoma presented with decreased FAZ circularity index and increased FAZ perimeter compared to healthy controls (Choi et al., 2017). In addition to measuring macula vessel density, Shoji et al. (2017) compared the rate of vessel density loss among glaucoma, glaucoma suspect, and healthy eyes (Shoji et al., 2017). Of note, the authors found that the mean rate of change in macula whole en-face vessel density was significantly faster in glaucoma eyes compared to glaucoma suspect eyes (0.85%/year, p=0.001) and healthy eyes (0.29%/year, p=0.004). This correlation remained consistent when comparing superior and inferior sectors of the macula (Shoji et al., 2017). Baek et al. (2019) specifically investigated the pattern and magnitude of diurnal variation in peripapillary and macular retinal vessel density in subjects with OAG using OCTA. The authors found that in patients with OAG, the magnitude of diurnal changes in peripapillary RVD (p=0.013) and macular RVD (p=0.042) were significantly greater than the healthy controls (Baek et al., 2019). The magnitudes of diurnal variations of intraocular pressure and mean ocular-perfusion pressure were also found to be greater in the OAG group compared to healthy controls (Baek et al., 2019). These results demonstrate that eyes with glaucoma present with not only reduced macular vessel density, but also greater changes throughout the day compared to healthy controls, as measured by OCTA. Contrasting this Verticchio Vercellin et al. recently did not find any statistically significant diurnal variation in ONH ocular blood flow measurements assessed by OCTA in OAG, OHT, and healthy subjects (Verticchio Vercellin et al., 2020).

Choroid Layer OCTA Vascular Biomarkers

Significant differences in OCTA vascular biomarkers were also observed in the choroidal vasculature between glaucomatous eyes and healthy controls. Tepelus et al. (2019) found that patients with NTG had lower perfusion density of the choriocapillaries compared with healthy controls (Tepelus et al., 2019). In addition, NTG eyes showed a decrease in vessel length density in other layers of the eye such as the superficial vascular plexus, optic nerve head, and peripapillary areas compared to controls (Tepelus et al., 2019). These changes are consistent with earlier discussions of glaucomatous changes above. Changes in the choroid OCTA vascular biomarkers were also seen in severe OAG with disc hemorrhage compared to OAG without disc hemorrhage. Park et al. (2018) showed greater choroid microvascular dropout in OAG patients with previous disc hemorrhage compared to OAG glaucoma patients without previous disc hemorrhage (Park et al., 2018). Park and colleagues also showed that within the disc hemorrhage group, progressive glaucoma patients (as defined by OCTA) had greater choroid microvascular dropout compared to glaucoma patients with stable glaucoma (77.3% vs. 10.5% respectively) (Park et al., 2018). Choroid microvascular dropout was also greater in progressive glaucoma patients in the non-disc hemorrhage group compared to stable glaucoma patients in the non-disc hemorrhage group (50% vs. 23.1%) (Park et al., 2018). These studies provide evidence for greater OCTA vascular biomarker changes observed in the choroid layer of glaucomatous eyes compared to healthy controls, and differences in patients with progressive OAG vs. stable OAG.

DISCUSSION:

OCTA has emerged as a novel, non-invasive imaging modality combining clinical structural outcomes alongisde vascular biomarkers in a single imaging modality. A wealth of pilot data has shown strong associations of OAG with lower OCTA assessed vascular biomarkers in the retina and ONH. However, data from large longitudinal studies of OCTA vascular biomarkers and glaucomatous progression remain missing and the use of vascular outcomes in predicting OAG progression has not yet been realized. Examining the current literature reveals a strong uniformity of lower OCTA assessed vascular biomarkers in OAG patients, with supporting pilot data describing relationships to visual field loss and glaucomatous structural changes.

Specifically, both Jia and Lévêque showed a significant decrease in total capillary density and blood flow in the area of the optic nerve head in glaucomatous eyes compared to normal controls indicating greater vascular abnormalities in eyes affected by OAG (Jia et al., 2012; Jia et al., 2014; Lévêque et al., 2016). Authors Liu, Chen, and Triolo, Teplus, and Akil showed significant decreases in vessel density and flow density in the peripapillary capillary layers of ONH in glaucomatous eyes and demonstrated greater areas of perfusion loss and low flow (Liu et al., 2015; Chen et al. 2016; Chen et al., 2017; Chen et al., 2020; Triolo et al., 2017; Akil et al., 2017; Lu et al., 2020; Tepelus et al., 2019). Similar significant decreases were identified in the circumpapillary vessel density among glaucomatous eyes compared to healthy controls, with the greatest decrease seen in severely glaucomatous eyes (Yarmohammadi et al., 2016a; Yarmohammadi et al., 2016b). Finally in patients with unilateral glaucoma, Kim and Mammo demonstrated greater vessel density changes in the affected eye compared to the fellow unaffected eye showing that vascular biomarkers on OCTA are useful for tracking progression of glaucoma for each eye independently (Mammo et al., 2016; Kim et al., 2017).

In the macula, decreases in vessel density in glaucomatous eyes were seen in a number of capillary layers including the superficial vascular complex, deep retinal capillary layer, and whole retina (Takusagawa et al., 2017; Choi et al., 2017; Chao et al., 2019; Milani et al., 2021). A decrease in flow density was also demonstrated by Kromer in both the superficial and deep capillary layers in the macula as a whole, as well as the parafoveal, perifoveal, nasal regions specifically (Kromer et al, 2019). There were also significant decreases in perifoveal and parafoveal vessel density in OAG patients compared to healthy controls (Lu et al. 2020)

In the choroidal layer, lower perfusion density was seen in the choriocapillaris layer in LTG patients compared to controls (Tepelus et al., 2019). Greater choroid microvascular dropout was greater in OAG patients with previous disc hemorrhage compared to patients without previous disc hemorrhage, and within groups patients with progressive glaucoma had greater microvascular dropout than patients with stable glaucoma (Park et al., 2018).

While OCTA has emerged as a powerful new imaging modality, it is important to acknowledge its limitations. First, OCTA specifically images the posterior retinal and choroidal microvasculature, which means that the complete ocular vascular system, including the retrobulbar vessels, cannot be analyzed. Second, it is possible for the superficial layer of retinal vasculature to obscure the deeper vessels of the retina when imaged using OCTA, which may make interpreting the images more challenging. Furthermore, a high-quality OCTA image relies on the patient to remain still and avoid blinking. In addition, the presence of artefacts can make the OCTA scans more difficult to analyze (Koustenis et al., 2017). Lastly, Monteiro-Henriques et al. (2021) found that several patient characteristics including smoking status, hyper or hypoxia conditions, and cardiovascular risk, among others, were related to ocular vascular changes (Monteiro-Henriques et al., 2021). As a result, vascular biomarkers measured by OCTA may need to be adjusted to take into account such variables, but further research is needed in this area. Despite this, many OCTA devices offer a metric of scan quality for each image, which can aid clinicians in which scans they can reliably analyze and interpret. In terms of research evaluating the uses of OCTA for vascular biomarkers in glaucoma, more prospective longitudinal studies are needed as well as larger sample sizes.

While vascular OCTA biomarkers provide valuable insight into glaucoma progression, standardization across different OCTA devices is an ongoing concern. In a study comparing the Optovue, Topcon, and Zeiss OCT-A devices, Lu et al. (2019) found significant differences in the superior capillary plexus, and the deep capillary plexus in the macula (Lu et al., 2019). There was no significant difference in the foveal avascular zone (Lu et al., 2019). Another study compared 7 different OCT-A devices in measuring the vessel density, fractal density, and foveal avascular zone (VD, FD, FAZ) (Corvi et al., 2018). The comparison included the Spectralis, Optovue, Angioplex, PlexElite, RS-3000 Advance, OCT-HS100, and Revo NX OCT-A devices (Corvi et al., 2018). There was a significant difference in the mean value produced by the 7 devices in both the superficial and deep vascular plexus (Friedman test p<.0001) (Corvi et al., 2018). The authors of the study concluded that comparing the measurements of the VD, FD, and FAZ among the 7 devices was not feasible (Corvi et al., 2018). This highlights the importance of the need for standardization among different devices and caution in making direct comparisons between devices (Corvi et al., 2018; Lu et al., 2019). In addition to inter-device variability, OCTA image data can differ based on the software, protocol, processing, and measurements generated (Spaide et al., 2018). There are currently a number of OCTA algorithms for analyzing OCTA images including but not limited to: optical angiography (OMAG), split-spectrum-amplitude-decorrelation angiography (SSADA), and full-spectrum amplitude-decorrelation angiography (FSADA) (Rodríguez et al., 2018). These algorithms are often unique to the various OCTA devices that are currently on the market (Rodríguez et al., 2018). The differences in these measurement methods make comparing OCTA data between protocols and devices difficult, and standardization of OCTA vascular markers is necessary going forward (Spaide et al.2018).

In conclusion, the ability for OCTA to quantify both clinical structural and blood flow changes to numerous vascular layers of the eye simultaneously makes it a promising tool for advancing OAG diagnosis and disease management. Vascular biomarkers from OCTA may be especially important in monitoring risk in persons with higher levels of vascular disease. OCTA is currently limited, however, by its variation in scan quality, influence of patient movement, potential for artifacts, lack of standardization in measurements and analysis, and differences among devices made by different manufactures that make direct comparison of OCTA data difficult. Future studies should be designed for understanding OCTA vascular biomarkers as predictors of disease progression by engaging in large and appropriately powered longitudinal studies of OCTA vascular biomarkers alongside traditional structural and functional measures of glaucoma progression.

Table 3:

Choroid Layer optical coherence tomography angiography (OCTA) Vascular Markers. OAG: open-angle glaucoma; ONH: optic nerve head.

Author (year)	Subjects	OCTA Scan	Results	P-value
Tepelus et al. (2019)	26 low tension glaucoma (LTG) patients 22 age-matched healthy subjects	6 × 6 mm2 macular layers	LTG group had lower perfusion density of the choriocapillaris compared with normal controls	P<0.001
			LTG group showed reductions in vessel length density compared with controls for the superficial vascular plexus and optic nerve head + peripapillary area	P=0.03 for both
Park et al. (2018)	82 OAG with disc hemorrhage 68 OAG without disc hemorrhage	4.5 × 4.5 mm ONH	Choroid microvascular dropout was greater in patients with glaucoma and previous disc hemorrhage than glaucoma patients without previous disc hemorrhage	P=0.025
			Within the Disc-Hemorrhage group, progressive glaucoma patients (as defined by OCTA) had significantly greater choroid microvascular dropout compared to stable patients (77.3% vs. 10.5%)	P<0.001
			Within the non-Disc Hemorrhage group, progressive glaucoma patients had significantly greater choroid microvascular dropout compared to stable patients (50% vs. 23.1%)	P<0.001