At the completion of this CPD activity, optometrists will understand the clinical utility of optical coherence tomography angiography (OCT-A). Including:

OCT-A has emerged as a pivotal imaging technology. But despite its clinical utility for detailed viewing of the retinal and choroidal microvasculature, many optometrists have been slow to adopt it. DR MALI OKADA explores its benefits and the limitations, as well as conditions where OCT-A provides insight that surpasses traditional optometric imaging devices.

Dr Mali Okada MMed, FRANZCO Royal Victorian Eye and Ear Hospital, Melbourne, Australia Centre for Eye Research Australia, University of Melbourne, Melbourne, Australia

Op hthalmic imaging, specifically retinal imaging technology, has seen remarkable advances over the last two decades. From basic 30-55 degree colour fundus photography and fundus fluorescein angiogram (FFA), clinicians now have access to a suite of multimodal imaging techniques including: optical coherence tomography (OCT), fundus autofluorescence, ultra-wide field 200-degree fundus imaging, and more recently, optical coherence tomography angiography (OCT-A). 

In particular, OCT-A has been instrumental in allowing repeated non-invasive cross-sectional imaging of the eye. The technology builds upon standard OCT by using motion-contrast scans to produce detailed volumetric maps of both retinal and choroidal microvasculature.1 

Comparison of OCT-A versus FFA

In contrast to conventional FFA, OCT-A imaging is non-invasive and because no dye injection is required, risks of allergy or adverse reactions (such as nausea, vomiting and dye extravasation around the cannuala) for the patient are avoided. 

OCT-A is fast, reproducible and typically takes seconds to complete (compared to five to 10 minutes for a standard FFA).2 This is particularly useful in the paediatric population and those with poor vascular access such as diabetic or elderly patients. It can also be used in pregnant women where traditional invasive FFA is a relative contraindication.3

Consequently, it can be frequently repeated and used for both diagnostic and monitoring purposes in the office including in optometry practices, which are restricted in their ability to do FFA. 

Aside from ease-of-use, the major advantage of OCT-A is its ability to provide high-resolution detail as well as depth information down to the micron level. It can show the relationship between the vessels to its surrounding tissue architecture through the B-scan slice. Compared to FFA, which typically can only show the superficial capillary plexus (SCP), OCT-A can visualise and separate out the deeper vascular layers as well as the choriocapillaris (Figure 1).4 As such, OCT-A technology has been used to provide insight into abnormalities in retinochoroidal vascular structure such as neovascularisation as well as disorders of perfusion.

As with all technology however, OCT-A does have some inherent limitations. Compared to FFA, it can be more prone to imaging artifacts. These include signal to noise artifacts from media opacities (such as cataracts), movement artifact (such as patient blinking or saccades) and projection artifacts (for example: from intraretinal lipid exudates).2 Segmentation errors can also provide a false representation of the vasculature, particularly if viewed in a single enface image.5 This can be a problem particularly if there is significant anatomical changes such as large pigment epithelial detachments or posterior staphylomas in high myopes which can obscure any underlying choroidal neovascular membranes (CNVM). 

Additionally, because OCT-A relies on motion to generate structural images, it does not provide any assessment of leakage. This aspect can be beneficial in cases where the fluorescein dye in FFA leaks and obscures underlying structures but can be a drawback if flow information is required (for example, in assessing patency of thrombosed retinal macroaneurysms with slow flow). It also cannot show the point of leakage of lesions such as focal leaks in central serous retinopathy (CSR) that may be amenable to laser treatment and is less useful for detecting polypoidal lesions than indocyanine green angiography.  

Furthermore, although newer experimental models of OCT-A can provide wider field of view beyond the macula, as yet, it cannot match the 200 degrees provided by ultrawide field devices such as the Optos FFA when assessing peripheral pathology as such as peripheral ischaemia or leakage (Figure 2). 

In these situations above, where a lesion is suspected on OCT but not evident on OCT-A or is in an area not readily imaged by OCT-A in the periphery, it would be advisable to still obtain or refer to an ophthalmologist or clinic for a traditional FFA. 

Tips and step by step process for OCT-A interpretation

There are different ways to analyse OCT-A images, but some general points to consider when performing and interpreting scans include:

Acquisition areas of modern OCT-A devices range from 3 x 3 mm to 12 x 12 mm and even wider field using montage functions of newer models (15 x 9 mm).6 It is important to choose the appropriate scan protocol depending on the indication and likely pathology as there is a trade-off between increased field of view and lower resolution of fine vasculature. 

For example, for choroidal neovascularisation (CNV) in age related macular degeneration (AMD), a scan protocol of 3 x 3 mm is ideal as it can help detect smaller lesions and segmentation can be more finely tuned. In contrast, for diabetic retinopathy assessment, ultra-wide field scans of 12 x 12 mm provide a broader view of retinal ischaemia, delineating areas of capillary non-perfusion and neovascularisation elsewhere. 

Although the OCT-A device will automatically segment enface images into various pre-defined layers (superficial capillary plexus, deep capillary plexus (DCP), avascular layer, RPE layer, choriocapillaris layer), the algorithm used to detect boundaries may not always represent the true anatomical layer.7 

When pathology is present (for example: in high myopia or eyes with large pigment epithelial detachment), the normal architecture can be severely disrupted and manual readjustment of the segmentation lines may be required to visualise the appropriate layer. In some cases where this is not possible, awareness of the limitations of the segmentation needs to be considered when interpreting the scan output. 

Similarly, other artifacts such as projection artifacts can give the impression of vessels (typically from more superficial layers) being erroneously visualised in deeper layers. One strategy that can minimise this is to subtract the SCP from the enface image of the DCP.5

Although OCT-A images are typically viewed from its enface output, it is helpful to remember that a key benefit of OCT-A technology is it provides a volumetric and depth-resolved image. The retina is not flat but a three-dimensional structure, and images should be viewed with this in mind to fully appreciate the interrelation between various vessels and tissue structures. While sophisticated 3D rendering is now possible using OCT-A,8 viewing the angiographic analysis as a fly-through video from superficial-to-deep in one continuum can help with this spatial interpretation. Vascular lesions should also be correlated with their corresponding image on structural OCT scan. 

Summary of suggested steps when approaching OCT-A:

Indications for OCT-A imaging in macular disorders

Although OCT-A imaging can be performed in any macular disorder, there are some conditions where this technology is particularly high yield and beneficial.

The role of OCT-A is perhaps best highlighted in patients with neovascular AMD (nAMD) as it can accurately image CNV lesions (Figure 3). This can be used for diagnostic purposes but also to monitor response to therapy with anti-vascular endothelial growth factor (anti-VEGF) injections over time.9 As there is no obscuration from dye leakage, quantitative and qualitative measurements of the lesion can be made. In particular, CNV lesions can show regression in linear size as well as pruning of the peripheral capillary tufts with treatment.10

More recently, OCT-A has identified a new entity, ‘non-exudative’ or ‘quiescent’ CNV – lesions that are detectable on OCT-A as abnormal vascular networks above Bruch’s membrane but show no evidence of leakage either on OCT or FFA (Figure 4). This is of particular relevance in patients with a history of CNV in one eye, with studies showing subclinical CNVM in up to 14% of fellow eyes with presumed intermediate only AMD.11 For these patients, the risk of conversion to exudative or active disease at one year was 15 times higher than those without subclinical lesions.11 As such, closer follow up and patient counselling for symptoms is recommended in this cohort. 

OCT-A can also be a useful adjunct to help differentiate polypoidal choroidal vasculopathy (PCV) lesions from nAMD. Studies have shown when combined with the structural OCT image, OCT-A can reveal localised increased flow in the subretinal pigment epithelium space, providing high sensitivity and specificity for diagnosing PCV without requiring the use of FFA or indocyanine green angiography (ICGA).12 

Other secondary causes of choroidal neovascularisation 

The development of OCT-A imaging has also been of benefit in screening for secondary CNV lesions. For example, in patients with central serous retinopathy, adult vitelliform dystrophy or macular telangiectasia type 2, there can be fluid or hyporeflective spaces on structural OCT and leakage or pooling of the dye on FA can be an inherent part of the macular disorder. 

It can often be difficult to tell if a patient is progressing from natural history or from the development of a secondary CNV. In this situation, OCT-A can be used to screen for presence of a CNV complex. Similarly, patients with posterior uveitis can have active inflammatory lesions as well as a secondary CNV. As both can leak on FFA and ICGA, OCT-A may be used to help differentiate these two lesions, an important distinction given the contrasting treatment approaches.13 

Diabetic retinopathy and retinal vascular ischaemia

Aside from CNV detection, OCT-A has also shown significant promise in assessing retinal capillary ischaemia in many retinal vascular disease such as diabetic retinopathy and retinal vein occlusions. In patients with diabetic retinopathy (DR), 3 x 3 mm macula OCT-A scan provides accurate assessment of the foveal avascular zone (FAZ), demonstrating FAZ enlargement and perifoveal capillary dropout in both the SCP and deep capillary plexus.14 

This is particularly useful in patients with seemingly mild or early disease, where OCT-A can reveal significant pathology such as vascular remodelling and capillary drop out.15 It can also demonstrate other features of DR including quantifying microaneursyms and neovascularisation (NV) at the disc and elsewhere without the need for traditional dye angiography. Although used mostly as a research tool, newer wide field OCT-A can be used to both detect and monitor more peripheral NV lesions and areas of ischaemia without the need to montage images.16  

Retinal imaging technology has and continues to evolve rapidly. Although currently OCT-A may not be required for daily management of most macular disorders, it can provide helpful insight above and beyond traditional imaging and its role is likely to expand as we continue to understand the wealth of information this imaging technology provides. 

Optical coherence tomography angiography: hype or hope?

An optometrist’s role in the diabetic care team

Unilateral glaucoma or historic non-arteritic anterior ischaemic optic neuropathy (NAION)?