Secondary Logo

Journal Logo

Review Article

The New Era of Retinal Imaging in Hypertensive Patients

Tan, Wilson∗,†; Yao, Xinwen PhD∗,‡,§; Le, Thu-Thao PhD¶,||; Tan, Bingyao PhD∗,‡,§; Schmetterer, Leopold PhD∗,‡,§,||,∗∗,††,‡‡; Chua, Jacqueline BOptom PhD∗,||

Author Information
Asia-Pacific Journal of Ophthalmology: March-April 2022 - Volume 11 - Issue 2 - p 149-159
doi: 10.1097/APO.0000000000000509
  • Open

Abstract

Systemic hypertension remains the leading contributor to the global burden of disease and global all-cause mortality,1 and microcirculation plays an important role in the pathophysiology of hypertension.2 In the early phase of hypertension, arteries, and arterioles frequently constrict because of a variety of nervous or endocrine, and autocrine mechanisms.2,3 In the chronic phase, these functional changes of vascular constriction become more structural, as distinguished by inward hypertrophy of arteries and arterioles accompanied by the ongoing rarefaction of arterioles and capillar-ies.2,3 These structural alterations not only result in the chronic elevation of high blood pressure, but also lead to impaired tissue perfusion to various end organs, such as the heart, brain, and kidneys, substantially increasing the risk of morbidity and mortality.

The management of hypertension involves the reduction of blood pressure levels but can be enhanced by assessing and prognosticating end-organ damage. Current assessment of endorgan damage, however, relies on biochemical parameters or ultrasound / magnetic resonance / computed tomography imaging. Direct visualization of the microcirculation is largely lacking, because of limited resolution of the imaging techniques. In this review, we will focus on the clinical applicability of advanced ocular imaging techniques, such as optical coherence tomography angiography (OCTA) and adaptive optics (AO) technology relevant to the management of systemic hypertension.

FUNDUS PHOTOGRAPHY

In the retina, where blood vessels are directly visible, the changes related to hypertension can be viewed by clinicians during ophthalmological examination and assessed as hypertensive retinopathy. The current systems of grading hypertensive retinopathy, the Keith-Wagner-Barker or Wong-Mitchell classification systems (Table 1), are based on the classification of hypertensive retinopathy into different severities, depending on the detection of retinal signs, such as arteriolar narrowing, arteriovenous nicking, and dot-blot hemorrhages4 (Fig. 1). Except in its severe stage, hypertensive retinopathy alone is not expected to cause vision loss. Therefore, the significance of hypertensive retinopathy has progressively shifted to a marker and prognostic factor for end-organ damage, namely in the heart, brain, and kidneys.5

Table 1 - Current Classification Systems for Hypertensive Retinopathy
Keith-Wagner-Barker Classification System93 and Associated Clinical Features Wong-Mitchell Classification System94 Clinical Features and Associated Classification System Systemic Associations Proposed by Wong and Mitchell94 for the Wong-Mitchell
Grade 1 Generalized arteriolar narrowing Mild Generalized/focal arteriolar narrowing, arteriovenous nicking, opacity of arteriolar wall (“copper wiring”), or combination of these signs Modest association with risk of stroke. However, association has only been specifically observed for generalized arteriolar narrowing and arteriovenous nicking.95Modest association with coronary heart disease. However, association has only been specifically observed for focal/ generalized arteriolar narrowing, and arteriovenous nicking.96,97 Modest association with death. However, association has only been specifically observed for focal/generalized arteriolar narrowing, and arteriovenous nicking, in middle-aged persons.98
Grade 2 Focal narrowing and arteriovenous nicking
Grade 3 Grade 2 plus exudates, hemorrhages, and cotton wool spots Moderate Soft/hard exudates, hemorrhages (flame, dot, blot), microaneurysms, or combination of these signs Strong association with risk of stroke. However, association has only been specifically observed for microaneurysms, soft exudates, blot hemorrhages and flame-shaped hemorrhages.95Strong association with cognitive decline. However, association has only been specifically observed for microaneurysms, retinal hemorrhage, and soft exudates.99Strong association with death from cardiovascular causes. However, association has not been demonstrated with any specific parameter.98
Grade 4 Grade 3 plus optic disc swelling Malignant Moderate grade plus optic disc swelling Strong association with death. However, no studies support this conclusion.

F1
Figure 1:
Arteriovenous nicking (blue arrow), cotton wool spots (white arrows), flame-shaped (black arrow), and dot-blot hemorrhages (yellow arrow), and microaneurysms (black arrows) in an eye with moderate hypertensive retinopathy.

Potential Biomarkers Obtained From Fundus Photographs

Over the last 3 decades, epidemiological studies from multiple countries such as the US,6,7 the UK,8 Denmark,9 Holland,10 Norway,11 Australia,7,12 Japan,13 Singapore,14 and China15 have used standardized fundus photographs and semi-automated image processing technologies to quantitatively measure retinal vascular geometric parameters, including vessel diameter, focal narrowing, fractal branching architecture, and tortuosity.16 The earlier studies were performed predominantly in patients with hypertension and diabetes because microcirculation had increasingly been recognized for its role in these diseases at that time. These preliminary studies essentially corroborated previous microvascular observations in animal models of hypertension that showed arteriolar narrowing resulting from intimal thickening and medial hyperplasia, hyalinization, and sclerosis of arteriolar walls.17

Furthermore, the narrowing of arterioles may be a hallmark of the early stages of hypertension. Indeed, a meta-analysis of long-term epidemiological follow-up studies observed that “normotensive” persons with generalized retinal arteriolar narrowing were more likely to develop hypertension.18 This confirmed the idea that retinal imaging could provide a new source of information. Because the retinal microvascular changes can already be visualized in preclinical stages of hypertension, retinal imaging could have predictive power for the development and related risk of end-organ damage.

Apart from the development of hypertension, several studies6,12 have observed that retinopathy signs such as generalized arteriolar narrowing and arteriovenous nicking have been associated with both current and past blood pressure levels, suggesting that these signs could be particularly helpful in assessing persistent microvascular changes from hypertension. The quantitative assessment of retinal microvascular geometry characteristics from these imaging techniques shows high reproducibility and can be used repeatedly in the same person for follow-up studies.19

Use of Fundus Photographs for End-Organ Disease Prognostication

Studies have also reported relationships of retinal microvascular changes with end-organ damage (Table 1). In terms of the heart, retinal arteriolar narrowing and venular widening were found to have links with coronary heart disease in females,20,21 and long-term mortality risk and ischemic stroke in both males and females.21 A review of these studies can be found in a 2019 meta-analysis paper.22 They observed that retinal arteriolar, but not venular caliber, was consistently correlated with large arterial function and large arterial structure, particularly in patients with cardiometabolic disease. One notable conclusion of this meta-analysis was that preclinical changes in the microcirculation and large arteries have some shared but mainly unique pathways associated with cardiovascular disease. This suggests that retinal microvascular changes have predictive power for the development of cardiovascular risk.

A recent review conducted on the associations between retinal microvascular signs and heart diseases23 suggested that the retinal microvasculature can provide essential data about concurrent cardiac disease status and predict future risk of cardiac-related events. An interesting finding of this review is that vessel diameters, particularly narrower arterioles in combination with wider venules, correlate with the incidence of acute coronary syndrome, especially in women, and can predict its development. This gender discrepancy supports the hypothesis that microvascular dysfunction plays a greater role in the pathogenesis of coronary heart disease in women than in men. Another important finding from this review is that apart from vessel diameters, other structural characteristics of the retinal vessels, such as branching angles and tortuosity, are associated with heart disease and mortality.

Apart from cardiovascular risks, retinal imaging can also provide information on the early signs of brain-related risks, such as the development of (vascular) dementia.24 Several studies have demonstrated that the presence of arteriolar narrowing and venular widening is associated with the incidence of stroke,25,26 and the incidence of lacunar strokes.27,28 Retinal hemorrhages were also associated with cerebral hemorrhages.29,30 Furthermore, venular widening has also been observed to have links with incident dementia,31 prevailing dementia,32,33 and prevailing Alzheimer disease34 independent of age, blood pressure, and traditional risk factors. Indeed, a recent review by Rim and coworkers35 found that both qualitatively and quantitatively assessed parameters on fundus photography were consistently associated with clinical cerebrovascular disease, including clinical stroke, cerebral hemorrhage, and stroke mortality. These parameters were also found to be associated with subclinical cerebrovascular disease, including subclinical cerebral large artery infarction, lacunar infarction, and white matter lesions identified on magnetic resonance imaging. These patterns of association between retinal changes with brain-related risks and diseases are not unexpected because the retinal vessels share similarities with the brain in terms of vascular development and likelihood abnormalities.24

There were fewer studies investigating the association between retinal microvascular signs and kidney diseases in hypertensive patients. However, these studies had conflicting findings. Although 1 study observed retinal arteriolar narrowing in association with incident chronic kidney disease,36 another found both retinal arteriolar narrowing and increased venular diameter were associated with incident chronic kidney disease.37 Other studies, however, did not find associations with chronic kidney disease38,39 nor decline in renal function.40

Limitations of Fundus Photographs

Although the nonmydriatic fundus camera allows the larger-scale application, the predictive value of mild hypertensive retinopathy for end-organ damage in hypertension remains modest (Table 1). In the search for methods to detect subclinical retinopathy damage early, the color fundus photography with advanced software to analyze the retina vessels seems to be more useful.41 However, fundus photographs remain limited to the larger vessels of the superficial layers of the retinal circulation, namely the arteries and veins. It does not allow the assessment of smaller vessels, ie, capillaries or deep retinal vascular layer and the choroid.

OPTICAL COHERENCE TOMOGRAPHY ANGIOGRAPHY

The optical coherence tomography angiography (OCTA) is a functional extension of optical coherence tomography (OCT) which images the retinal vasculature by detecting the movement of red blood cells within the vessels. By comparing repeated OCT B scans, it is possible to image the vessels by computing differences among the scans. It has provided us with high-resolution, depth-resolved images of both the superficial and deep retinal vascular layer and the choroid (Fig. 2). The OCTA is also highly attractive for use in large cohort studies and for consecutive follow-up visits as it does not require the administration of intravenous dye.42 A recent meta-analysis by our group reviewed the studies on quantitative OCTA parameters in hypertensive patients and concluded that certain OCTA parameters could provide objective information about preclinical microvascular changes from systemic hypertension.43

F2
Figure 2:
Clinical and optical coherence tomography angiography (OCTA; 3 x 3 mm area; PLEX Elite 9000, Carl Zeiss Meditec) imaging of the left eye. A, Color fundus photograph showing a normal eye. B-D, Images of superficial retinal layer. E-G, Deep retinal layer. H–J, Choriocapillaris layer. OCTA images of the respective plexuses were exported from the review software (B, E, and H), and retinal capillary density and choriocapillaris flow deficits parameters were automatically extracted using a customized MATLAB algorithm (The MathWorks, Inc, Natick, MA; Version R2020b). Foveal avascular zone area demarcated in blue (D and G). The white arrows in I and J indicate the presence of a single flow deficit, which appeared as black in the OCTA image (I), and white in the processed images (J). OCTA indicates optical coherence tomography angiography.

Potential Biomarkers Obtained From Optical Coherence Tomography Angiography

The vessel density of the superficial capillary plexus (SVD), vessel density of the deep capillary plexus (DVD), and superficial foveal avascular zone (FAZ) are OCTA parameters that have been extensively explored in hypertension studies. However, as OCTA measures blood flow by motion contrast, the decrease in vessel density in the retinal and/or choroidal capillaries could be due to a low rate of blood flow or capillary occlusions. Therefore, OCTA can only provide quantitative measurements of vessels with rate of blood flow exceeding minimum detection threshold, and measurements generated may not reflect the true extension of vessels.

In the SVD (Fig. 3), Hua and coworkers44 showed that hypertensive patients with poor blood pressure control had significantly lower SVD as compared to healthy controls. In addition, 2 studies45,46 showed that the SVD of hypertensive patients of at least 5 years’ duration had lower measurements when compared to controls, whereas Peng and coworkers47 found that the SVD was significantly decreased in patients with hypertensive retinopathy when compared to controls. Several studies, on the other hand,48,49 did not find a significant difference in SVD between the eyes of hypertensives and controls.

F3
Figure 3:
Retinal capillary density maps of the macular region showing the superficial retinal microvasculature of hypertensive patients (A-D) with poorly-controlled blood pressure (BP) and (E-H) with well-controlled BP. Hypertensive patients with poorly-controlled BP generally have sparser capillary density at the superficial (B; 18.02 mm−1) and deep (D; 13.19 mm−1) layers as compared to patients with well-controlled BP who have denser capillaries at the superficial (F; 22.62 mm−1) and deep (H; 18.98 mm−1) layers. The perfusion heat map further indicates the presence of large nonperfused areas (marked by red arrows) that tend to be seen in patients with poorly-controlled BP (B and D) as compared to patients with well-controlled BP (F and H). Foveal avascular zone area (in black) was excluded from the calculation of capillary density.

As for the DVD (Fig. 3), several studies48–50 found that the DVD was significantly reduced in the macula of hypertensive eyes. In addition, Peng and coworkers4 found that the DVD was reduced in hypertensive patients when compared to controls, independent of the presence of hypertensive retinopathy. However, Hua and coworkers46 found that there was no difference in DVD between hypertensive patients of least 5 years’ duration and controls.

In the FAZ, Donati and coworkers48 observed a significant increase in the FAZ area in hypertensive patients compared to healthy subjects. Two studies46,51 found that the FAZ area was significantly larger in patients with more than 10 years of hypertension when compared to controls. Similarly, Lim and coworkers45 observed that the FAZ area was significantly greater in patients with more than 5 years of hypertension when compared to controls.

Several additional OCTA parameters have been explored, such as fractal dimension and peripapillary vessel calibers. For example, Xu and coworkers50 observed that the fractal dimension of the retinal capillary plexuses was significantly reduced in the macula of hypertensive eyes. Meanwhile, hypertensive eyes had marginally narrower peripapillary arteriolar caliber.50

Other emerging OCTA biomarkers would be the choriocapillaris (Fig. 4), especially with the advent ofswept-source OCTA which offers distinct advantages over the previous spectral-domain OCTA technology, allowing researchers to view the choriocapillaris region underneath the retinal pigment epithelium with much greater clarity. Using the swept-source OCTA, Chua and coworkers52 reported that in normal controls there were many small choriocapillaris flow deficits, whereas in uncontrolled hypertension there was a progressive reduction in the number of choriocapillaris flow deficits with an increasing mean area of choriocapillaris flow deficits. In addition, the study observed that individuals with uncontrolled systemic hypertension had the most significant choriocapillaris flow deficits compared to well-controlled hypertensives and normal controls.

F4
Figure 4:
Swept-source optical coherence tomography angiography (3 x 3 mm area) and color-coded maps indicating regions of flow deficits (B; color-coded) of a choriocapillaris layer of a hypertensive with poorly-controlled blood pressure (BP) (A and B) and a hypertensive with well-controlled BP (C and D). The color-coded maps show the presence of larger-sized choriocapillaris flow deficits in a hypertensive with poorly controlled BP (B; labeled as yellow). The presence of flow deficits can be seen as areas of dark regions in the angiogram (A and C) and its sizes are color-coded (B and D).

OCTA parameters may have the potential to create new hypertensive retinopathy classification systems. A recent study by Liu and coworkers53 proposed a novel 3-stage hypertensive retinopathy classification system using parameters assessed on OCTA, in place of the classic Keith-Wagener-Barker classification system. The OCTA classification system was found to have correlations with renal damage but was unable to find a correlation between the stages of the OCTA classification system proposed and cardiovascular events. Whilst novel, the correlation between the proposed classification system and renal damage was cross-sectional in nature, and not appropriate for prognostication. Future studies or meta-analyses can consider using long-term follow-up studies to develop classification systems that have prognostic ability; such systems may have greater clinical utility.

Effect of Antihypertensive Medications Using Optical Coherence Tomography Angiography

Blood pressure levels have been found to be correlated with several OCTA parameters. Xu and coworkers50 observed that the skeletal and vessel density of the DVD was correlated negatively with mean arterial pressure, and Hua and coworkers44 observed that blood pressure was significantly correlated with the SVD, DVD, and the inside disc capillary density. Similarly, Peng and coworkers47 showed that blood pressure was found to be correlated with DVD. In addition, Chua and coworkers52 showed that choriocapillaris flow deficits were highly dependent on blood pressure control and less on systemic hypertension status, as flow patterns were similar between persons with well-controlled systemic hypertension and healthy controls. This is consistent with the findings of Hua and coworkers44 that the SVD did not significantly differ between patients with well-controlled blood pressure of more than 10 years and patients with well-controlled blood pressure of 5 to 10 years, suggesting that having a good blood pressure control maintained the structural integrity of the microvasculature in the SVD. Taken together, these findings support the possible beneficial role of blood pressure control in preventing, maintaining, or reversing these preclinical microvascular changes.

Furthermore, it has been proposed that different classes of antihypertensive treatments have differing effects at the micro-vascular levels.54 OCTA thus offers the opportunity to observe these effects at the microvascular level noninvasively, and the insights could possibly change practice patterns in terms of medication prescription for different subgroups of patients. For example, Peng and coworkers47 reported that patients under monotherapies were shown to have lower SVD than patients taking angiotensin-converting enzyme inhibitors (ACE inhibitors or ACE-I)/angiotensin receptor blocker and calcium channel blocker combination therapy, and patients taking only calcium channel blocker had significantly lower DVD. In addition, analyzing digitized red-free monochrome images of fundus photographs, Hughes and coworkers55 reported that amlodipine- and lisinopril-based treatment was significantly associated with the reduction of arteriolar narrowing. Similarly, Thom and coworkers56 observed using fundus photography that amlodipine-based treatment was associated with a lesser arteriolar narrowing than atenolol-based treatment. OCTA imaging may give greater insights into these findings and would be an interesting area to explore.

Use of Optical Coherence Tomography Angiography for End-Organ Disease Prognostication

Several studies have explored the association of OCTA parameters and hypertensive end-organ damage in the kidneys and brain. With regards to the kidneys, 5 studies48,52,53,57,58 have described relationships between OCTA parameters and impaired kidney function based on microalbuminuria and estimated glomerular filtration rate levels. Although both studies reported a reduced retinal capillary density and lower estimated glomerular filtration rate in persons with systemic hypertension, Chua and coworkers58 reported a change in the superficial vascular layer, whereas Frost and coworkers57 reported changes in the deep retinal vascular layer. In contrast, however, Donati and coworkers48 reported that microalbuminuria levels were not significantly correlated with SVD, DVD, or FAZ. In the choriocapillaris, Chua and coworkers52 showed that changes in the choriocapillaris microvasculature were associated with kidney function. As for the brain and cognition, Wang and coworkers59 observed a significant decrease in SVD in patients with cerebral small vessel disease, and that the hypoperfusion was associated with cerebral magnetic resonance imaging markers and cognitive function. Future longitudinal research is required to establish whether OCTA metrics improve end-organ disease prognostication over existing markers, or provide an easier method compared to less accessible or more costly tests.

Limitations of Optical Coherence Tomography Angiography Studies

There have been inconsistencies on whether OCTA parameters significantly differ between hypertensive patients and healthy controls. Several reasons explain the discordant findings. First, the small sample size, ranging from 2860 to 16947 hypertensive patients without the considerations of relevant confounders. Second, published OCTA studies in hypertensive patients relied primarily on classifications of brachial blood pressures of which targets remain controversial. In recent years, more reliable markers such as fibrosis to monitor the progression of hypertension have been proposed.61,62 The associations of these new hypertensive markers with OCTA parameters have not been investigated. Lastly, eyes with pathology, such as hypertensive retinopathy, can significantly alter retinal anatomy and cause algorithms to misidentify boundaries. Mislabeling of retinal layers and consequent vessel density measurement error have been reported even in one-third of healthy eyes,63 and of all the published studies, only the studies by Xu50 and Donati48 checked for segmentation errors.

ADAPTIVE OPTICS IMAGING

Several authors have been interested in the pathogenic mechanisms of hypertension, such as the influence of microvasculature changes in early life,64 or genetic determinants of microvascular structure and function in hypertension.65 A major challenge is examining early or subtle changes in vascular development as determined by genetic or environmental factors, and how such changes could lead to sustained hypertension in adulthood. For this reason, imaging microscopic features of the retinal vasculature using AO systems could prove useful in such studies.

The application of AO transforms an ophthalmoscope into a microscope, allowing greater in vivo resolution of the retinal microvasculature to an extent that is not possible with conventional clinical imaging modalities.66 The AO imaging technology uses deformable mirrors to compensate for the optical imperfections, also known as aberrations in the eye,67 allowing retinal vessels to be imaged at near histological levels.68 Given the remodeling of arterioles may be an early process underlying end-organ damage due to hypertension,69 the noninvasive nature of AO imaging makes it highly attractive for follow-up studies on pathogenic mechanisms and therapeutic interventions in hypertension.

The AO ophthalmoscopy can be performed using 2 modalities, either the AO flood illumination (AO-FIO) or AO scanning laser (AO-SLO). The AO-FIO system is a fundus photography system at its core, integrated with AO hardware to counteract the optical aberration. It adopts widefield (or flood) illumination and an area sensor to capture a 2-dimensional en face image in 1 single shot. On the contrary, the AO-SLO system is essentially a confocal scanning system where the incident light is tightly focused on 1 spot of the retina, with the help of the AO hardware, and the backscattered light is collected by a single-pixel detector through a confocal pinhole. A 2-dimensional en face image is created when the incident light is raster-scanned over the retina. Compared to AO-SLO, the AO-FIO systems are inherently less susceptible to motion artifacts owing to their fast acquisition speed. However, without the confocal pinhole to reject the out-of-focus signals, AO-SLO produces images with less signal-to-noise ratio.70 This review will focus on AO-FIO as more studies are published. The AO-SLO systems were mainly custom built71 and have only been made commercial recently.

Potential Biomarkers Obtained From Adaptive Optics Imaging

A series of vascular biomarkers such as the inner diameter, the outer diameter, and the parietal thickness can be imaged using the AO-FIO (Fig. 5). These biomarkers are commonly measured in arterioles because the venule wall is often not visible on AO-FIO images (Fig. 5). In a recent meta-analysis, Bakker et al71 summarized the results of studies included in PubMed and Scopus databases as of July 9, 2020, on retinal biomarkers obtained from AO imaging in patients with hypertension. They concluded that a significantly smaller inner diameter3,68,72–74 (Fig. 6) and larger parietal thickness3,68,72,73 can be observed in hypertensive patients as compared to healthy controls, resulting in an increased wall-to-lumen ratio3,68,72,73,75 without an increase in wall cross-sectional area.3,68,72–76 This finding supports the prevailing theory of arteriolar remodeling in hypertension77 with microscopic level scrutiny. Their overall conclusion from this systematic review is that although the AO vascular biomarkers are altered in patients with hypertension, there is a need to standardize the AO imaging protocols and validate these procedures for the longitudinal monitoring of hypertension.

F5
Figure 5:
Vascular imaging biomarkers as imaged from the adaptive optics image (RTX1, Imagine Eyes, France).
F6
Figure 6:
In an eye with mild hypertensive retinopathy, comparing fundus photographs (A and B) and adaptive optics image (RTX1, Imagine Eyes, France) showing arteriovenous nicking (blue arrow) with focal arteriolar narrowing (yellow arrow) (C).

Effect of Antihypertensive Medications Using Adaptive Optics Imaging

Several AO studies have observed that blood pressure management in hypertension can lead to the reversal of retinal microvascular changes.73,78 For instance, a significant increase was observed in inner diameter and reduction in parietal thickness after antihypertensive treatment,78 suggesting a reversal of eutrophic remodeling as a result of blood pressure changes. With regards to specific medication classes, however, AO imaging studies have presented discordant findings. Gallo and coworkers74 found no significant difference between the effects of different classes of antihypertensive treatment on retinal microvascular remodeling, whereas a multivariate analysis by Rosenbaum and coworkers73 showed that hypertension drug regimen was not an independent predictor of any retinal anatomical indices. In contrast, De Ciuceis and coworkers79 found a significant reduction in the wall-to-lumen ratio of hypertensive patients after undergoing combined lercanidipine and enalapril treatments over a period of 24 weeks.

Use of Adaptive Optics Imaging for End-Organ Disease Prognostication

The wall-to-lumen ratio of retinal arterioles has shown a strong correlation with the media-to-lumen ratio of subcutaneous vessels,3 a potent predictor of cardiovascular and cerebrovascular events in multivariate analyses80 but which requires invasive biopsy of gluteal tissue for analysis. There is currently, however, a paucity of studies examining the relationship between AO imaging biomarkers and end-organ damage in hypertensive cohorts. The only study to do so thus far failed to demonstrate a significant relationship between the wall-to-lumen ratio of patients with and other meaningful end-organ injuries.81 It was however a small study of 27 patients and only included patients with malignant hypertension.

Limitations of Adaptive Optics Imaging

Despite its capability to capture high-resolution images of the retinal vasculature, AO imaging is not quite ready for routine clinical care. This is because it is tedious to capture AO images and can be exhausting for patients. Furthermore, as the region of analysis in AO imaging is often limited to a single vessel in a limited field of view, usually a segment of the superotemporal arterioles,3,68,73–75,79 it is difficult to be certain that the changes in that arteriole are reflective of all other retinal vessels. In the future, large field-of-view montages may be attractive and will potentially yield more reliable information on the microcirculatory status.

Moreover, the optics of the AO imaging is highly sensitive to media opacities, fixation stability, and tear film quality. Likewise, image processing, montaging, and analysis require custom software and technical expertise due to the lack of automated techniques. Technological imaging advances in speed, tracking, and software improvement will improve the scan quality in challenging eyes and produce reports quickly in a fast-paced clinical setting.

In addition, the magnification of vessels on AO imaging is dependent on the axial length of the eye being imag ed.82 Specifically, measurements that rely on absolute values, such as the inner diameter and outer diameter, are impacted by the magnification effect of the eye and can cause bias when comparing values between subjects; however, parameters that compute a dimensionless ratio, such as the wall-to-lumen ratio, are not affected by this issue.

FUTURE DIRECTIONS

The advent of deep learning using artificial intelligence creates exciting possibilities for the future of hypertension diagnosis, cardiovascular risk factor screening, and prediction of cardiovascular events.83 Artificial intelligence has the potential to predict the presence of hypertension using fundus photographs. For instance, several models developed have achieved an area under the curve of 0.65 to 0.77 and an accuracy of 60.9% to 68.8% in predicting hypertension in different populations living in China.84,85 These models may be especially useful in detecting patients with white coat hypertension (higher blood pressure readings at the doctor's office than other settings) or masked hypertension (normal blood pressure readings at doctor's but chronically high blood pressure outside of the clinic). Deep learning may allow clinicians to rapidly and noninvasively detect these underdiagnosed groups of patients and give them the appropriate treatment that they would otherwise not receive.

Apart from predicting hypertension status, Google researchers Poplin and coworkers86 showed using White and Hispanic populations that using deep learning on retinal images alone was sufficient to predict several cardiovascular risk factors including age, gender, smoking status, blood pressure, and body mass index. The same deep learning algorithm was also able to predict the onset of major adverse cardiovascular events within 5 years. Similarly, using more than 70,000 images from multiple countries and multiple ethnicities, Cheung and coworkers87 developed a deep learning model that could predict multiple cardiovascular risk factors from retinal vessel calibers. Lastly, Rim and coworkers88 developed a deep learning model demonstrating superior performance in predicting the presence of coronary artery calcium, a preclinical marker of atherosclerosis that is strongly associated with risk of clinical cardiovascular disease.89 The use of deep learning is exciting as it can help make cardiovascular risk screening of a large population both technically and economically feasible.

CONCLUSIONS

Since the late 19th century, the significance of retinopathy signs as indicators of systemic morbidity and mortality has long been recognized. Furthermore, performing ophthalmologic examinations in patients with hypertension has been endorsed by international hypertension management guidelines.90,91 However, retinal assessment nearly disappeared from routine clinical practice towards the end of the 20th century.

Major technological advances in ocular imaging techniques have improved the way the retinal microcirculation can be assessed, offering a unique opportunity to further our understanding of eye–body relationships and support the development of novel diagnostic and prognostic tools through noninvasive means. Furthermore, microvascular imaging techniques such as the OCTA can be deployed in community primary care settings to identify people at risk of systemic diseases and follow their progression during treatment. This setting allows the collection of large data sets for the study of “Oculomics,” the association of ophthalmic biomarkers with systemic health and disease.92 Such microvascular damage, if demonstrable, may provide a biomarker of at–risk individuals with the greatest likelihood for progressive cardiac, neurological, and renal function decline.

In summary, although the clinical and prognostic value of early microvascular retinal abnormalities assessed either by OCTA and AO imaging remains lacking, the availability of novel standardized quantitative imaging analysis techniques may shed new insights on the challenging task of identifying hypertensive individuals at risk of end–organ damage. Furthermore, the combination of deep learning with OCTA and AO imaging techniques could help provide better risk stratification of hypertensive patients.

REFERENCES

1. Forouzanfar MH, Afshin A, Alexander LT, et al. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990-2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet 2016; 388:1659–1724.
2. Levy BI, Schiffrin EL, Mourad JJ, et al. Impaired tissue perfusion: a pathology common to hypertension, obesity, and diabetes mellitus. Circulation 2008; 118:968–976.
3. De Ciuceis C, Agabiti Rosei C, Caletti S, et al. Comparison between invasive and noninvasive techniques of evaluation of microvascular structural alterations. J Hypertens 2018; 36:1154–1163.
4. Chua J, Cheung YLC, Schmetterer L. Sheyman Alan, Fawzi Amani A, et al. Hypertensive fundus changes. Retinal Vascular Disease.. Singapore: Springer Singapore; 2020; 85-97.
5. Cheung CYL, Ikram MK, Sabanayagam C, et al. Retinal microvasculature as a model to study the manifestations of hypertension. Hypertension 2012; 60:1094–1103.
6. Sharrett AR, Hubbard LD, Cooper LS, et al. Retinal arteriolar diameters and elevated blood pressure: the Atherosclerosis Risk in Communities Study. Am J Epidemiol 1999; 150:263–270.
7. Wong TY, Klein R, Klein BE, et al. Retinal vessel diameters and their associations with age and blood pressure. Invest Ophthalmol Vis Sci 2003; 44:4644–4650.
8. Sharp PS, Chaturvedi N, Wormald R, et al. Hypertensive retinopathy in Afro-Caribbeans and Europeans. Prevalence and risk factor relationships. Hypertension 1995; 25:1322–1325.
9. Munch IC, Kessel L, Borch-Johnsen K, et al. Microvascular retinopathy in subjects without diabetes: the Inter99 Eye Study. Acta Ophthalmol 2012; 90:613–619.
10. Stolk RP, Vingerling JR, de Jong PT, et al. Retinopathy, glucose, and insulin in an elderly population. The Rotterdam Study. Diabetes 1995; 44:11–15.
11. Bertelsen G, Peto T, Lindekleiv H, et al. Sex differences in risk factors for retinopathy in non-diabetic men and women: the Tromso Eye Study. Acta Ophthalmol 2014; 92:316–322.
12. Leung H, Wang JJ, Rochtchina E, et al. Impact of current and past blood pressure on retinal arteriolar diameter in an older population. J Hypertens 2004; 22:1543–1549.
13. Fukushima S, Nakagami T, Suto C, et al. Prevalence of retinopathy and its risk factors in a Japanese population. J Diabetes Investig 2013; 4:349–354.
14. Bhargava M, Cheung CY, Sabanayagam C, et al. Prevalence and risk factors for retinopathy in persons without diabetes: the Singapore Indian Eye Study. Acta Ophthalmol 2014; 92:e602–e609.
15. Peng XY, Wang FH, Liang YB, et al. Retinopathy in persons without diabetes: the Handan Eye Study. Ophthalmology 2010; 117:531–537. 537.e1-2.
16. Cheung CY, Ikram MK, Sabanayagam C, et al. Retinal microvasculature as a model to study the manifestations of hypertension. Hypertension 2012; 60:1094–1103.
17. Tso MO, Jampol LM. Pathophysiology of hypertensive retinopathy. Ophthalmology 1982; 89:1132–1145.
18. Ding J, Wai KL, McGeechan K, et al. Retinal vascular caliber and the development of hypertension: a meta-analysis of individual participant data. J Hypertens 2014; 32:207–215.
19. Sherry LM, Wang JJ, Rochtchina E, et al. Reliability of computer-assisted retinal vessel measurementin a population. Clin Exp Ophthalmol 2002; 30:179–182.
20. Gopinath B, Chiha J, Plant AJ, et al. Associations between retinal microvascular structure and the severity and extent of coronary artery disease. Atherosclerosis 2014; 236:25–30.
21. Seidelmann SB, Claggett B, Bravo PE, et al. Retinal vessel calibers in predicting long-term cardiovascular outcomes: the Atherosclerosis Risk in Communities Study. Circulation 2016; 134:1328–1338.
22. Liu M, Wake M, Wong TY, et al. Associations of retinal microvascular caliber with intermediate phenotypes of large arterial function and structure: a systematic review and meta-analysis. Microcirculation 2019; 26:e12557.
23. Allon R, Aronov M, Belkin M, et al. Retinal microvascular signs as screening and prognostic factors for cardiac disease: a systematic review of current evidence. Am J Med 2021; 134:36–47.e7.
24. Cheung CY, Ikram MK, Chen C, et al. Imaging retina to study dementia and stroke. Prog Retin Eye Res 2017; 57:89–107.
25. Wieberdink RG, Ikram MK, Koudstaal PJ, et al. Retinal vascular calibers and the risk of intracerebral hemorrhage and cerebral infarction: the Rotterdam Study. Stroke 2010; 41:2757–2761.
26. Wong TY, Kamineni A, Klein R, et al. Quantitative retinal venular caliber and risk of cardiovascular disease in older persons: the cardiovascular health study. Arch Intern Med 2006; 166:2388–2394.
27. Ikram MK, Witteman JCM, Vingerling JR, et al. Retinal vessel diameters and risk of hypertension. Hypertension 2006; 47:189–194.
28. Yatsuya H, Folsom AR, Wong TY, et al. Retinal microvascular abnormalities and risk oflacunar stroke. Stroke 2010; 41:1349–1355.
29. Baker ML, Hand PJ, Wong TY, et al. Retinopathy and lobar intracerebral hemorrhage: insights into pathogenesis. Arch Neurol 2010; 67:1224–1230.
30. Gobron C, Erginay A, Massin P, et al. Microvascular retinal abnormalities in acute intracerebral haemorrhage and lacunar infarction. Rev Neurol 2014; 170:13–18.
31. de Jong FJ, Schrijvers EM, Ikram MK, et al. Retinal vascular caliber and risk of dementia: the Rotterdam study. Neurology 2011; 76:816–821.
32. Baker ML, Marino Larsen EK, Kuller LH, et al. Retinal microvascular signs, cognitive function, and dementia in older persons: the Cardiovascular Health Study. Stroke 2007; 38:2041–2047.
33. Qiu C, Cotch MF, Sigurdsson S, et al. Cerebral microbleeds, retinopathy, and dementia: the AGES-Reykjavik Study. Neurology 2010; 75:2221–2228.
34. Cheung CY, Ong YT, Ikram MK, et al. Microvascular network alterations in the retina of patients with Alzheimer's disease. Alzheimers Dement 2014; 10:135–142.
35. Rim TH, Teo AWJ, Yang HHS, et al. Retinal vascular signs and cerebrovascular diseases. J Neuroophthalmol 2020; 40:44–59.
36. Yau JW, Xie J, Kawasaki R, et al. Retinal arteriolar narrowing and subsequent development of CKD Stage 3: the Multi-Ethnic Study of Atherosclerosis (MESA). Am J Kidney Dis 2011; 58:39–46.
37. Yip W, Ong PG, Teo BW, et al. Retinal vascular imaging markers and incident chronic kidney disease: a prospective cohort study. Sci Rep 2017; 7:9374.
38. McGowan A, Silvestri G, Moore E, et al. Evaluation of the retinal vasculature in hypertension and chronic kidney disease in an elderly population ofIrish nuns. PLoS One 2015; 10:e0136434.
39. Phan K, Au C, Mitchell P, et al. Chronic kidney disease and the severity of coronary artery disease and retinal microvasculature changes: a cross- sectional study. J Thorac Dis 2016; 8:2111–2114.
40. Grunwald JE, Pistilli M, Ying GS, et al. Retinopathy and progression of CKD: the CRIC study. Clin J Am Soc Nephrol 2014; 9:1217–1224.
41. Sun C, Wang JJ, Mackey DA, et al. Retinal vascular caliber: systemic, environmental, and genetic associations. Surv Ophthalmol 2009; 54:74–95.
42. Chalam KV, Sambhav K. Optical coherence tomography angiography in retinal diseases. J Ophthalmic Vis Res 2016; 11:84–92.
43. Tan W, Yao X, Le TT, et al. The application of optical coherence tomography angiography in systemic hypertension: a meta-analysis. Front Med (Lausanne) 2021; 8:778330.
44. Hua D, Xu Y, Zhang X, et al. Retinal microvascular changes in hypertensive patients with different levels of blood pressure control and without hypertensive retinopathy. Curr Eye Res 2021; 46:107–114.
45. Lim HB, Lee MW, Park JH, et al. Changes in ganglion cell-inner plexiform layer thickness and retinal microvasculature in hypertension: an optical coherence tomography angiography study. Am J Ophthalmol 2019; 199:167–176.
46. Hua DH, Xu YS, Zeng XB, et al. Use of optical coherence tomography angiography for assessment of microvascular changes in the macula and optic nerve head in hypertensive patients without hypertensive retinopathy. Microvasc Res 2020; 129:103969.
47. Peng Q, Hu Y, Huang M, et al. Retinal neurovascular impairment in patients with essential hypertension: an optical coherence tomography angiography study. Invest Ophthalmol Vis Sci 2020; 61:42.
48. Donati S, Maresca AM, Cattaneo J, et al. Optical coherence tomography angiography and arterial hypertension: a role in identifying subclinical microvascular damage? Eur J Ophthalmol 2021; 31:158–165.
49. Sun C, Ladores C, Hong J, et al. Systemic hypertension associated retinal microvascular changes can be detected with optical coherence tomography angiography. Sci Rep 2020; 10:9580.
50. Xu Q, Sun HY, Huang X, et al. Retinal microvascular metrics in untreated essential hypertensives using optical coherence tomography angiography. Graefes Arch Clin Exp Ophthalmol 2021; 259:395–403.
51. Lee WH, Park JH, Won Y, et al. Retinal microvascular change in hypertension as measured by optical coherence tomography angiography. Sci Rep 2019; 9:156.
52. Chua J, Le TT, Tan B, et al. Choriocapillaris microvasculature dysfunction in systemic hypertension. Sci Rep 2021; 11:4603.
53. Liu Y, Li J, Pan J, et al. Morphological changes in and quantitative analysis of macular retinal microvasculature by optical coherence tomography angiography in hypertensive retinopathy. Hypertens Res 2021; 44:325–336.
54. Levy BI, Ambrosio G, Pries AR, et al. Microcirculation in hypertension. Circulation 2001; 104:735–740.
55. Hughes AD, Stanton AV, Jabbar AS, et al. Effect of antihypertensive treatment on retinal microvascular changes in hypertension. J Hypertens 2008; 26:1703–1707.
56. Thom S, Stettler C, Stanton A, et al. Differential effects of antihypertensive treatment on the retinal microcirculation. Hypertension 2009; 54:405–408.
57. Frost S, Nolde JM, Chan J, et al. Retinal capillary rarefaction is associated with arterial and kidney damage in hypertension. Sci Rep 2021; 11:1001.
58. Chua J, Chin CWL, Hong J, et al. Impact of hypertension on retinal capillary microvasculature using optical coherence tomographic angiography. J Hypertens 2019; 37:572–580.
59. Wang X, Wei Q, Wu X, et al. The vessel density of the superficial retinal capillary plexus as a new biomarker in cerebral small vessel disease: an optical coherence tomography angiography study. Neurol Sci 2021; 42:3615–3624.
60. Terheyden JH, Wintergerst MWM, Pizarro C, et al. Retinal and choroidal capillary perfusion are reduced in hypertensive crisis irrespective of retinopathy. Transl Vis Sci Technol 2020; 9:1–7.
61. Diez J, Querejeta R, Lopez B, et al. Losartan-dependent regression of myocardial fibrosis is associated with reduction of left ventricular chamber stiffness in hypertensive patients. Circulation 2002; 105:2512–2517.
62. Goh VJ, Le TT, Bryant J, et al. Novel index of maladaptive myocardial remodeling in hypertension. Circ Cardiovasc Imaging 2017; 10:e006840.
63. Ghasemi Falavarjani K, Habibi A, Anvari P, et al. Effect of segmentation error correction on optical coherence tomography angiography measurements in healthy subjects and diabetic macular oedema. Br J Ophthalmol 2020; 104:162–166.
64. Liew G, Wang JJ, Duncan BB, et al. Low birthweight is associated with narrower arterioles in adults. Hypertension 2008; 51:933–938.
65. Liu YP, Kuznetsova T, Thijs L, et al. Are retinal microvascular phenotypes associated with the 1675G/A polymorphism in the angiotensin II type-2 receptor gene? Am J Hypertens 2011; 24:1300–1305.
66. Liang J, Williams DR, Miller DT. Supernormal vision and high-resolution retinal imaging through adaptive optics. J Opt Soc Am A Opt Image Sci Vis 1997; 14:2884–2892.
67. Rizzoni D, Docchio F. Assessment of retinal arteriolar morphology by noninvasive methods: the philosopher's stone? J Hypertens 2016; 34:1044–1046.
68. Koch E, Rosenbaum D, Brolly A, et al. Morphometric analysis of small arteries in the human retina using adaptive optics imaging: relationship with blood pressure and focal vascular changes. J Hypertens 2014; 32:890–898.
69. Park JB, Schiffrin EL. Small artery remodeling is the most prevalent (earliest?) form of target organ damage in mild essential hypertension. J Hypertens 2001; 19:921–930.
70. Paques M, Meimon S, Rossant F, et al. Adaptive optics ophthalmoscopy: application to age-related macular degeneration and vascular diseases. Prog Retin Eye Res 2018; 66:1–16.
71. Bakker E, Dikland FA, Van Bakel R, et al. Adaptive optics ophthalmoscopy: a systematic review of vascular biomarkers. Surv Ophthalmol 2021; 67:369–387.
72. Gallo A, Mattina A, Rosenbaum D, et al. Retinal arteriolar remodeling evaluated with adaptive optics camera: relationship with blood pressure levels. Ann Cardiol Angeiol 2016; 65:203–207.
73. Rosenbaum D, Mattina A, Koch E, et al. Effects of age, blood pressure and antihypertensive treatments on retinal arterioles remodeling assessed by adaptive optics. J Hypertens 2016; 34:1115–1122.
74. Gallo A, Dietenbeck T, Giron A, et al. Non-invasive evaluation of retinal vascular remodeling and hypertrophy in humans: intricate effect of ageing, blood pressure and glycaemia. Clin Res Cardiol 2021; 110:959–970.
75. Mehta RA, Akkali MC, Jayadev C, et al. Morphometric analysis of retinal arterioles in control and hypertensive population using adaptive optics imaging. Indian J Ophthalmol 2019; 67:1673–1677.
76. Meixner E, Michelson G. Measurement of retinal wall-to-lumen ratio by adaptive optics retinal camera: a clinical research. Graefes Arch Clin Exp Ophthalmol 2015; 253:1985–1995.
77. Intengan HD, Schiffrin EL. Vascular remodeling in hypertension. Hypertension 2001; 38:581–587.
78. Gallo A, Rosenbaum D, Kanagasabapathy C, et al. Effects of carotid baroreceptor stimulation on retinal arteriole remodeling evaluated with adaptive optics camera in resistant hypertensive patients. Ann Cardiol Angeiol 2017; 66:165–170.
79. De Ciuceis C, Salvetti M, Paini A, et al. Comparison of lercanidipine plus hydrochlorothiazide vs. lercanidipine plus enalapril on micro and macrocirculation in patients with mild essential hypertension. Intern Emerg Med 2017; 12:963–974.
80. Rizzoni D, Agabiti Rosei C, De Ciuceis C, et al. New methods to study the microcirculation. Am J Hypertens 2018; 31:265–273.
81. de Nattes T, Saad R, Buob D, et al. Retinal arteriolar occlusions and exudative retinal detachments in malignant hypertension: more than meets the eye. Am J Hypertens 2021; 34:30–33.
82. Legras R, Gaudric A, Woog K. Distribution of cone density, spacing and arrangement in adult healthy retinas with adaptive optics flood illumination. PLoS One 2018; 13:e0191141.
83. Tseng RMWW, Rim TH, Cheung CY. Grzybowski A, et al. Artificial intelligence using the eye as a biomarker of systemic risk. Artificial Intelligence in Ophthalmology.. Cham: Springer; 2021; 243-255.
84. Zhang L, Yuan M, An Z, et al. Prediction of hypertension, hyperglycemia and dyslipidemia from retinal fundus photographs via deep learning: a cross-sectional study of chronic diseases in central China. PLoS One 2020; 15:e0233166.
85. Dai G, He W, Xu L, et al. Exploring the effect of hypertension on retinal microvasculature using deep learning on East Asian population. PLoS One 2020; 15:e0230111.
86. Poplin R, Varadarajan AV, Blumer K, et al. Prediction of cardiovascular risk factors from retinal fundus photographs via deep learning. Nat Biomed Eng 2018; 2:158–164.
87. Cheung CY, Xu D, Cheng CY, et al. A deep-learning system for the assessment of cardiovascular disease risk via the measurement of retinal- vessel calibre. Nat Biomed Eng 2021; 5:498–508.
88. Rim TH, Lee CJ, Tham YC, et al. Deep-learning-based cardiovascular risk stratification using coronary artery calcium scores predicted from retinal photographs. Lancet Digit Health 2021; 3:e306–e316.
89. Adelhoefer S, Uddin SMI, Osei AD, et al. Coronary artery calcium scoring: new insights into clinical interpretation-lessons from the CAC Consortium. Radiol Cardiothorac Imaging 2020; 2:e200281.
90. Williams B, Poulter NR, Brown MJ, et al. British Hypertension Society guidelines for hypertension management 2004 (BHS-IV): summary. BMJ 2004; 328:634–640.
91. Chobanian AV, Bakris GL, Black HR, et al. The Seventh report of the joint national committee on prevention, detection, evaluation, and treatment of high blood pressure: the JNC 7 report. JAMA 2003; 289:2560–2572.
92. Wagner SK, Fu DJ, Faes L, et al. Insights into systemic disease through retinal imaging-based oculomics. Transl Vis Sci Technol 2020; 9:6.
93. Keith NM, Wagener HP, Barker NW. Some different types of essential hypertension: their course and prognosis. Am J Med Sci 1974; 268:336–345.
94. Wong TY, Mitchell P. Hypertensive retinopathy. N Engl J Med 2004; 351:2310–2317.
95. Wong TY, Klein R, Couper DJ, et al. Retinal microvascular abnormalities and incident stroke: the Atherosclerosis Risk in Communities Study. Lancet 2001; 358:1134–1140.
96. Duncan BB, Wong TY, Tyroler HA, et al. Hypertensive retinopathy and incident coronary heart disease in high risk men. Br J Ophthalmol 2002; 86:1002–1006.
97. Wong TY, Klein R, Sharrett AR, et al. Retinal arteriolar narrowing and risk of coronary heart disease in men and women. The Atherosclerosis Risk in Communities Study. JAMA 2002; 287:1153–1159.
98. Wong TY, Klein R, Nieto FJ, et al. Retinal microvascular abnormalities and 10-year cardiovascular mortality: a population-based case-control study. Ophthalmology 2003; 110:933–940.
99. Wong TY, Klein R, Sharrett AR, et al. Retinal microvascular abnormalities and cognitive impairment in middle-aged persons: the Atherosclerosis Risk in Communities Study. Stroke 2002; 33:1487–1492.
Keywords:

adaptive optics; deep learning; hypertension; OCTA; optical coherence tomography angiography

Copyright © 2022 Asia-Pacific Academy of Ophthalmology. Published by Wolters Kluwer Health, Inc. on behalf of the Asia-Pacific Academy of Ophthalmology.