OCTA of Early and Intermediate AMD
The early and intermediate phases of AMD are characterized by drusen and/or RPE and CC changes. The location of drusen above the CC, between Bruch’s membrane and the RPE, has led to speculation that overlying RPE and outer retinal changes occur due to nutrient deprivation. Drusen may form a type of barrier that interferes with the diffusion of oxygen and nutrients supplied by the CC. Furthermore, studies have suggested that certain sites are more prone to drusen formation due to prior changes in vascular dynamics. CC dysfunction, indicating insufficient choroidal perfusion, may guide and dictate the area of drusen formation. Indeed, FA and indocyanine green angiography show prolonged choroidal filling in dry AMD, suggesting underlying microvascular deficiency. However, it remains a matter of debate whether these vascular changes cause or result from drusen formation.
Fig. 1. SD-OCT and OCTA images of a 77-year-old male with dry AMD. a The structural OCT B-scan highlights three drusen lesions (yellow arrows). b On the corresponding en face OCTA image of the CC, loss of flow (yellow arrows) is seen under these three lesions. c Note that the corresponding regions on the OCT intensity en face do not suggest pronounced signal attenuation, confirming that the dark areas of no flow on the OCTA are indeed CC loss, as opposed to shadowing.
OCT has allowed for some light to be shed on the structural CC changes involved in disease progression. Drusen, RPE changes, and outer retinal alterations have been well visualized, especially with enhanced depth imaging OCT [33] and high-resolution spectral-domain (SD) OCT. This technology has allowed for the quantification and automated detection of drusen and choroidal and outer retinal structural changes [34, 35]. For example, choroidal thinning has been shown to correlate with AMD pathogenesis, progression, and prognosis [2, 36–38].
OCTA, as an extension of OCT, has further allowed for the analysis of associated microvascular changes. OCTA allows for depth-resolved visualization of retinal and choroidal vasculature. It acquires repeated B-scans at a given location and assesses differences in phase and intensity between consecutive scans caused by erythrocyte movement, using these decorrelation signals to generate a vascular map co-registered with structural data [39, 40]. En face images generated from axial cuts in the volumetric cube scan allow for the evaluation of various retinal layers, ranging from the internal limiting membrane to the choroid. The thickness and axial position of these slabs can also be altered to best visualize certain pathologies, offering a direct view of the suspected area of disease. The 840-nm wavelength of SD-OCTA limits our visualization of choroidal vasculature, as scattering by media opacities limits penetration below the RPE. This becomes significant when assessing eyes with dry AMD, as drusen contribute to light attenuation, thereby making it difficult to discern whether an underlying decrease in OCTA signal is due to limited signal penetration or is an indication of pathological decreased CC flow. However, this distinction can be made by comparing the en face OCT intensity image, or comparing a cross-sectional B scan image at the area of interest to the en face OCTA image. An area of shadowing, due to signal attenuation, will appear dark on both en face images as well as on the B scan image, while an area of decreased blood flow will appear dark only on the en face OCTA image and normal on the OCT intensity en face and B scan image [9]. However, the longer 1,050-nm wavelength of SS-OCTA allows for increased depth penetration with less scattering and interference, and therefore improved imaging of the choroid and CC [16].
OCTA images of the CC in normal eyes depict a dense, homogenous, and regular vascular pattern. Compared to age-matched normal eyes, OCTA images of eyes with early AMD have shown a reduction in CC density (Fig. 1). This finding supports histologic quantification studies that have shown a correlation between increased drusen density and decreased CC vascular density [8–10]. Focal areas of CC loss may also occur. As these areas become larger, and advance into areas of atrophy, the area originally occupied by the CC may become occupied by displaced underlying larger choroidal vessels [9]. Hypotheses for this lack of flow seen in the CC include drusen-mediated vascular reduction, decreased CC vessel caliber, decreased CC flow as opposed to a complete lack of flow, or a relationship between vessel walls and drusen (rather than lumen) prompting drusen formation at choriocapillary pillars [3]. Irrespective of the cause-effect uncertainty between drusen formation and CC vascular depletion, drusen have been used as an indirect marker for CC dysfunction [3, 8].
Prior belief held that anatomical thinning of the choroid began after there was progression of disease and RPE dysfunction, giving the impression that the choroid was spared during early AMD. However, the discovery of CC changes associated with drusen, as discussed above, has suggested that choroidal changes may begin earlier than considered [2]. Further studies comparing early and intermediate AMD have found conflicting results with respect to alterations in the superficial and deep retinal plexuses. Some studies suggest that the retinal vasculature is unaffected, while others have shown that the superficial vessel density was decreased in intermediate AMD eyes, suggesting that intraretinal vascular depletion starts at the intermediate stage [2]. These intraretinal vascular changes correlate with thinning of the choroid and of the inner retinal layer, and may be a late response to reduced oxygen demand [2].
OCTA of Late AMD
The advanced presentation of dry AMD is GA (or CRORA), which is characterized by a large well-defined area of loss of the RPE, overlying photoreceptors, and the CC [31]. This atrophy allows for the direct visualization of underlying larger choroidal vessels. OCTA has shown loss of CC flow in these regions and even displacement of underlying choroidal vessels into these CC voids (Fig. 2). Moreover, there also appears to be loss of choroidal vessels in the areas in the immediate perimeter of the GA.
Interestingly, SS-OCTA has demonstrated that while there is true loss of flow in the areas of the CC underlying GA, those areas in the perimeter of the GA which appear to have a lack of flow in the CC are actually areas of slow flow [16, 40]. SS-OCTA has allowed for the development of a technique to detect relative blood flow: variable inter-scan time analysis (VISTA) [40, 41]. Analysis of differences between consecutive OCT B-scans at a specific location provide the OCTA flow signal. If the flow within a particular