Age-related macular degeneration (AMD) has been
classified in two groups, neovascular and non-neovascular,
which historically has been synonymous with exudative
and non-exudative, respectively. Neovascular AMD occurs
when pathologic blood vessels, choroidal neovascularization
(CNV), arise from the choroid and extend above Bruch’s
membrane toward the outer retina. Often these vessels
are exudative in nature, resulting in leakage of fluid, lipid
exudate, or blood causing vision loss. The exudation
from CNV allows detection as leakage with dye based
angiography and as fluid with structural optical coherence
tomography (OCT). In 2013, Querques et al. used multimodel
imaging to detect treatment naïve quiescent CNV (1).
This diagnosis requires presence of (I) moderate reflective
material between an elevated retinal pigment epithelium
and Bruch’s membrane on spectral domain OCT; (II)
absence of fluid on OCT; (III) staining with fluorescein
angiography; and (IV) a plaque identified with indocyanine
green angiography. This study confirmed the existence of
non-exudative neovascular AMD, and that exudation is
not required for the presence of CNV. OCT angiography
(OCTA), which can detect CNV as moving blood cells in
the outer retinal slab (2) rather than relying on the presence
of leakage on FA or fluid on OCT, is an ideal imaging
modality to further study non-exudative CNV.
In 2015, Palejwala et al. first described the use of OCTA
to detect non-exudative CNV (3). In this study, 32 fellow
eyes of patients with neovascular AMD were scanned using
OCTA. Two cases of clinically silent non-exudative CNV
were found as flow in the outer retinal slab without exudation
on structural OCT and without leakage on fluorescein
angiography. One case was followed longitudinally and over
8 months, the CNV vessel area enlarged by 20%, however,
exudation never developed. Several other studies have
subsequently further characterized non-exudative CNV
using OCT angiography (4-7).
The recent study by Carnevali et al. aimed to describe
features of treatment naïve quiescent CNV (TNQ-CNV)
and estimate the detection rate by OCTA (7). This was an
observational case series comparing TNQ-CNV diagnosed
by traditional imaging methods to those which could be
imaged on commercially available OCTA, AngioPlex and
AngioVue. A group of 22 eyes of 22 patients with drusenoid
pigment epithelial detachment (PED) without vascular
network on ICGA were used as negative controls. OCTA
detected CNV in 18/22 study eyes and there were no false
positive eyes. They concluded there was sensitivity of 81.8%
(18/22 patients) and 100% specificity. The most common
morphology was irregular, foveal sparing, with a nonvisible
core and well-defined margin.
While OCT angiography is a powerful tool, it is important
to be aware of confounding projection artifact to avoid falsely
identifying CNV. As light passes through moving red blood
cells in the superficial retina, a flickering shadow is cast onto
the deeper structures and is misinterpreted as moving red
blood cells by OCT angiogram algorithms (8-11). One must thus be cognizant of this principle during evaluation for
CNV with OCTA. An excellent illustration is provided by
Zheng et al. (11), who compare slabs containing segments of
RPE (RPE-fit) versus slabs which follow the RPE contour;
the apparent vessels in the former are in fact projection
artifact from retinal vessels. The authors suggest projection
artifact may have been interpreted as vascularized drusen in
a separate study (5). The use of new algorithms to suppress
projection artifact and appropriate sectioning is therefore
critical to minimizing false positives in CNV detection (9).
Carnevali et al. are to be commended for their consideration
of potential projection artifacts. To avoid falsely identifying
drusenoid PEDs as CNV, they applied algorithms to
suppress projection artifact and manually adjusted
automatic segmentation software to separate the capillary
plexus, outer retinal layers and the choriocapillaris. They
then analyzed only the choriocapillaris slab.
CNV is typically diagnosed by FA and ICG which is
invasive, time consuming and carries the risk of an adverse
reaction. OCTA has the advantage of imaging faster at
higher resolution, in 3 dimensions, and in different layers.
Surveillance of non-exudative quiescent CNV in this
noninvasive way is appropriate because monitoring can be
frequent without need for invasive procedures, given no
required treatment.
Findings reported by Carnevali et al. and others provide
useful information on non-exudative CNV, but the real
utility will be to observe over time to learn its natural
history which may remain quiescent for long periods
of time, regress or wax and wane. It is likely that some
quiescent CNV is a precursor to exudative CNV, but it
could potentially also represent more mature vessels with
competent endothelial cell junctions, or be a separate entity
from exudative CNV. Only when features of non-exudative
quiescent CNV are understood that can be used to predict
development of exudative CNV will the technology be
useful to influence prognosis and treatment. Further
longitudinal studies are needed to establish what constitute
high risk features. These studies could also shed light on
other questions such as the effect of anti-VEGF on nonexudative
CNV in the fellow eye.
