Seminars in Ophthalmology

ISSN: 0882-0538 (Print) 1744-5205 (Online) Journal homepage: https://www.tandfonline.com/loi/isio20

Imaging Amblyopia: Insights from Optical

Coherence Tomography (OCT)

Eric D. Gaier, Ryan Gise & Gena Heidary

To cite this article: Eric D. Gaier, Ryan Gise & Gena Heidary (2019) Imaging Amblyopia: Insights
from Optical Coherence Tomography (OCT), Seminars in Ophthalmology, 34:4, 303-311, DOI:
10.1080/08820538.2019.1620810

To link to this article: https://doi.org/10.1080/08820538.2019.1620810

Published online: 03 Jun 2019.

Submit your article to this journal

Article views: 49

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=isio20
Seminars in Ophthalmology, 2019; 34(4): 303–311
© Taylor & Francis
ISSN: 0882-0538 print / 1744-5205 online
DOI: https://doi.org/10.1080/08820538.2019.1620810

Imaging Amblyopia: Insights from Optical

Coherence Tomography (OCT)
Eric D. Gaier1,2, Ryan Gise2,3, and Gena Heidary1,2

1
Department of Ophthalmology, Boston Children’s Hospital, Boston, MA, USA, 2Harvard Medical School,
Boston, MA, USA, and 3Neuro-Ophthalmology Service, Department of Ophthalmology, Massachusetts Eye
and Ear, Boston, MA, USA

ABSTRACT
Amblyopia refers to visual impairment resulting from perturbations in visual experience during visual devel-
opment, typically secondary to strabismus, uncorrected refractive error, and/or deprivation. Amblyopia has
traditionally been considered a cortical disease, but the depth of our understanding of this complex neurode-
velopmental condition is limited by our ability to appreciate structural pathophysiology in the visual pathway.
Recent advances in Optical Coherence Tomography (OCT) have facilitated numerous studies of the structural
changes in the retina and optic nerve, thereby expanding our appreciation for the pathogenesis of this
condition. In this review, we summarize findings from studies evaluating retinal, retinal nerve fiber layer,
and choroidal thickness changes in patients with amblyopia. Focusing on the largest and most recent studies,
we discuss common limitations and confounding variables in these studies. We summarize recent advances in
ocular imaging technology and reconcile the findings of early histological reports with those of structural OCT
in amblyopia.
Keywords: Amblyopia, optical coherence tomography, strabismus, anisometropia, choroidal thickness

INTRODUCTION changes associated with amblyopia. With these

advances, structural changes in the anterior visual
Amblyopia is a neurodevelopmental disorder that pathway, including the retina, are becoming increas-
results from abnormal visual experience during ingly recognized and studied. Both time and spectral
a critical period of visual development. Risk factors domain OCT have been used to study amblyopia. The
that typically contribute to abnormal visual experi- most widely employed subtype of OCT used clini-
ence include strabismus, uncorrected refractive error, cally and in the study of amblyopia is spectral domain
and/or deprivation (blockade of light reaching the OCT (SD-OCT), and studies using SD-OCT have eval-
retina). As a consequence, the neural connections uated retinal, nerve fiber layer, and choroidal thick-
between the retina and the brain do not develop nesses in amblyopic patients.
normally. In this regard, the pathogenesis of amblyo- In this review, we summarize findings from the
pia has traditionally been considered to be rooted in application of OCT to amblyopia. Due to the large
the visual cortex. The depth of our understanding of number and heterogeneity of studies falling in this
this complex neurodevelopmental condition is limited category, we focus on the largest and most recent
by our ability to appreciate structural pathophysiol- studies. We discuss common limitations and con-
ogy in the visual pathway. founding variables in these studies. Lastly, we
Recent advances in in vivo neuroimaging modal- examine more recent advances in ocular imaging
ities, including magnetic resonance imaging and opti- technology and reconcile the findings of early his-
cal coherence tomography (OCT), have facilitated tological reports with those of structural OCT in
numerous studies aiming to elucidate the structural amblyopia.

Received 1 May 2019; accepted 9 May 2019; published online 3 June 2019.
Correspondence Eric D. Gaier, Department of Ophthalmology, 300 Longwood Avenue, Fegan 4, Boston, MA 02115. E-mail:
[email protected]

303
304 E. D. Gaier et al.

MACULAR RETINAL THICKNESS eyes of 25 patients with anisomyopia and 31 patients

with ansiohypermetropia. Five studies made compari-
Many studies using OCT have examined retinal thick- sons with normal-sighted control eyes in addition to
ness in patients with amblyopia dating back to 2004.1 fellow eyes.9–11,13,16 Therefore, more recent studies of
Comparisons across several parameters have been retinal thickness in amblyopia are largely opposed in
made with fellow eyes (in cases of unilateral amblyo- their findings and conclusions with a greater number
pia) and the eyes of normal-sighted control subjects, (9/14) reporting no difference.
ideally age- and sex-matched. These studies are Increased central macular thickness is thought to
numerous with varied methodologies and conflicting represent delayed or perturbed pruning of intra-
results. retinal synapses in amblyopia. If this were the case,
In a comprehensive meta-analysis, Li et al.2 com- then one might expect segment-specific effects com-
piled 28 clinical studies of macular retinal thickness in paring amblyopic and non-amblyopic eyes. In their
amblyopia prior to 2015 that involved 408 patients meta-analysis, Li et al.2 did not find a significant dif-
with primarily anisometropic, strabismic, and mixed ference specific to the inner or outer macular thick-
amblyopia. They found an overall increase in foveal nesses. Kusbeci et al.9 went further to analyze the
minimum thickness (7 studies), mean foveal thickness ganglion cell complex, which includes the ganglion
(11 studies), and mean macular thickness (13 studies). cell layer and the inner plexiform layer, and they
Among studies examining macular difference found no difference between the eyes of 30 patients
between amblyopic and control eyes, there was with amblyopia (16 anisometropic, 14 strabismic) and
a combined increase in foveal minimum thickness (3 their fellow eyes or the eyes of 30 healthy controls.
studies). There was no difference in mean foveal (2 Chen et al.7 performed laminar retinal analysis on 18
studies) or macular thickness (6 studies) between eyes pediatric patients with unilateral anisometropic
with anisometropic versus strabismic amblyopia. amblyopia, comparing amblyopic eyes with the fel-
Although a meta-analysis helps to address the limita- low eyes and the eyes of 18 healthy control subjects.
tion of small sample sizes, it incorporates and ampli- They found statistically significant (with correction
fies systemic confounding by uncontrolled variables. for multiple comparisons) increases in the nasal and
Nevertheless, the conclusion was that amblyopia is inferior perifoveal retinal thicknesses, which were car-
associated with increased foveal and macular ried by increases in the nerve fiber layer and inner
thicknesses. nuclear layer thicknesses, respectively. Therefore, the
More recently, there have been 14 studies examining evidence across OCT studies to support this explana-
macular retinal thickness since 2015. Five studies have tion remains elusive.
demonstrated a difference in foveal and/or macular
thickness.3–7 Rajavi et al.3 found greater central and
1 mm perifoveal retinal thicknesses among 14 anisome- RETINAL NERVE FIBER LAYER (RNFL)
tropic and 14 strabismic amblyopic eyes compared to THICKNESS
fellow and the eyes of 28 control patients, and these
differences were greater with worsened depth of Many studies have examined the RNFL in amblyopia,
amblyopia. Kasem & Badawi4 found increased central dating back as early as 2004.1 Early OCT studies of the
and overall macular thickness in the amblyopic com- RNFL in amblyopic eyes were predicated on the
pared to the fellow eyes in 30 anisometropic, 22 strabis- hypothesis that amblyopia represents a mild optic
mic, and 12 deprivation amblyopia cases. Demircan neuropathy. In their meta-analysis, Li et al.2 also
et al.6 found a significant increase in the mean central incorporated studies examining the parapapillary
macular thickness in the amblyopic eyes compared to RNFL through 2015. Compiling data from 20 studies,
that of the fellow eyes of 53 patients with anisohyper- there was an overall thicker RNFL among amblyopic
metropia. This difference was present among patients in eyes compared to fellow eyes. However, only four
the younger group (5–12 years, N = 18) and not in the studies compared RNFL thicknesses in amblyopic
older group (13–42 years, N = 35). eyes to those of control eyes, finding no differences.
Nine studies conducted in this same time-frame Since 2015, 11 studies have specifically examined
found no difference in retinal thickness among amblyo- the RNFL in amblyopic eyes, and 2 of those found
pic eyes.8–16 Singh et al.12 conducted the largest of these a significant increase in RNFL thickness (not includ-
studies, examining 101 adult patients with anisometro- ing Chen et al.7 who described an increase in the nasal
pic amblyopia, including 31 patients with myopia, 42 perifoveal nerve fiber layer of anisometropic amblyo-
with hyperopia and 28 with significant astigmatism. pic eyes).4,12 Singh et al.12 specifically found an
They found no significant difference in central macular increase in inferior RNFL among patients with aniso-
thickness compared to the fellow eyes in all 3 anisome- hyperopic amblyopia only, but no difference in
tropic groups. Taskiran Comez et al.14 found no signifi- patients with anisomyopic or astigmatic amblyopia.
cant differences between the amblyopic and fellows Kasem & Badawi4 found an overall increase in the

Seminars in Ophthalmology
 Seminars in Ophthalmology 305

total RNFL thickness across the categories of amblyo- patients, 319 controls) with ages ranging from 4.5 to
pia they studied (anisometropic, strabismic and depri- 39.5 years among studies conducted in Turkey, Japan,
vational), with significant effects in the superior, and China. Axial length was inconsistently measured
inferior and nasal peripapillary sectors. Among and matched in control subjects, so they conducted
patients with anisometropic amblyopia, the inferior two separate analyses in which axial length was
sector was increased across myopic, astigmatic and accounted for and not considered. Both analyses
hyperopic eyes; otherwise, only patients with hyper- reached the same conclusion; there was an overall
opia showed differences in other sectors, including increase in choroidal thickness among amblyopic
superior, temporal and nasal sectors. Comparisons eyes compared to both fellow and control eyes. This
for these two studies were made to fellow eyes only effect was observed in subgroup analyses for ani-
without a normal-sighted control group. sometropic and strabismic amblyopia.
No differences in RNFL thickness among amblyo- Since the meta-analysis conducted by Liu et al., six
pic and non-amblyopic eyes were found in the major studies have used similar methods to study the
remaining nine studies.5,6,8–11,14–16 Of these studies, choroidal thickness in amblyopia. In a follow-up to the
two found increased retinal thickness among amblyo- Copenhagen Child Cohort 2000 Eye Study, Hansen et -
pic eyes (described above),5,6 and seven found no al.13 performed EDI-OCT on the 20 children with
differences.8–11,14–16 Yakar et al.15 found no difference amblyopia (among the 1335 in the study, ages 11–12
in global RNFL thicknesses between the amblyopic years) and found increased subfoveal choroidal thick-
and fellow eyes of 30 patients with anisometropic nesses in amblyopic eyes compared to the eyes of nor-
(hyperopic or astigmatic) amblyopia. Kantarci et al.16 mal-sighted children after adjusting for axial length.
examined 54 adult subjects (54 patients with anisome- Choroidal thickness in the amblyopic eyes was signifi-
tropic amblyopia and 52 control subjects) and found cantly increased compared with that of fellow eyes, but
no difference in RNFL thickness among amblyopic not after adjusting for axial length. Niyaz et al.18 used
eyes compared to fellow or control eyes. Several of EDI-OCT to compare the choroidal thicknesses of 90
these studies also examined optic disc parameters, children with various categories of amblyopic including
including optic disc area, and found no difference (20 anisometropic, 57 strabismic, and 13 mixed) with 30
between amblyopic and fellow/control eyes.5,7,10,14,15 control subjects, 10 of whom were hypermetropic. They
Therefore, while a composition of studies prior to found increased choroidal thicknesses in anisometropic
2015 suggests an increase in RNFL thickness in eyes compared to fellow and control eyes, but no sig-
amblyopic eyes, more recent data tend to favor no nificant difference in choroidal thickness for the other
difference; however, there is considerable methodolo- amblyopia categories. Al-Haddad et al.19 also used EDI-
gical heterogeneity among these studies. OCT to compare the amblyopic eyes of 50 children (30
anisometropic and 20 strabismic) to 50 control subjects
and also found an increased subfoveal choroidal thick-
CHOROIDAL THICKNESS ness. They found increased thicknesses in various par-
afoveal choroidal sectors compared to fellow and
Aside from retinal changes, OCT also has the capacity to control eyes, but without a statistical correction for
image deeper layers of the posterior pole, including the multiple comparisons. Oner & Bulut20 used EDI-OCT
choroid. The choroid plays a role in emmetropization, to analyze central choroidal thicknesses in 32 anisome-
and choroidal thickness has been shown to be respon- tropic children with amblyopia and reported signifi-
sive to refractive defocus. Therefore, there may be some cantly increased thicknesses in amblyopic eyes
changes in the choroid, potentially visualized using compared to fellow eyes. Therefore, historic and recent
OCT, that contribute to the development or represent studies using EDI-OCT are largely in agreement that
the pathophysiology of amblyopia. However, imaging amblyopic eyes demonstrate increased choroidal
resolution below the retinal pigment epithelium using thickness.
conventional SD-OCT is quite poor. To address this The composition of the choroidal OCT signal has
problem, enhanced depth imaging OCT (EDI-OCT) also been examined in amblyopia. Nishi et al.21 con-
has been employed in several studies to quantitatively ducted a study using EDI-OCT that was included in
image the choroid. Much like the results collected on the meta-analysis by Liu et al. described above. Their
retinal structure in amblyopia, published analyses of the group followed up on this work in 2017,22 with over-
choroid in amblyopia have varied considerably in their lap of 22 patients included in the original report, to
approach and conclusions. include a total of 40 children with anisohypermetro-
Liu et al.17 compiled patients examined in 11 dif- pic amblyopia and 103 age-matched controls. They
ferent studies reported between 2014 and 2016 in also found that the total choroidal area was signifi-
which EDI-OCT was employed to measure the sub- cantly greater in amblyopic eyes compared to that of
foveal choroidal thickness in amblyopic eyes. Their fellow eyes. Within the choroid, they found that the
analysis included a total of 768 participants (449 ratio of signal corresponding to choroidal lumen to

© 2019 Taylor & Francis

306 E. D. Gaier et al.

that of the stroma was greater in amblyopic eyes in overall RNFL thickness but did find a significant
compared to fellow and hyperopic control eyes. increase in thickness of the nasal RNFL sector in
They concluded that the increase lumen/stroma amblyopic eyes compared to both fellow and control
ratio could signify developmental “immaturity” of eyes. Kasem & Badawi4 also found that global para-
these eyes, but what cellular changes may underlie papillary RNFL thicknesses were not different
this difference are purely speculative at this point. between amblyopic and fellow eyes, but on sectoral
analysis, the superior and inferior sectors of the para-
papillary RNFL were significantly thicker in deprived
eyes compared to fellow eyes. Therefore, there are
FINDINGS IN DEPRIVATION AMBLYOPIA relatively fewer studies focused on deprivation
AND ANIMAL MODELS amblyopia, and these studies have yielded mixed
results with regard to differences in retinal and
A handful of recent OCT studies have specifically RNFL thickness. Those that do report a difference
examined patients with deprivation amblyopia, consistently describe increased thicknesses, the same
which often results from a structural anomaly of the direction of change as reported for anisometropic and
anterior segment of the eye such as congenital catar- strabismic amblyopia.
act. Deprivation amblyopia is less common than ani- Overall, OCT findings of increased retinal thick-
sometropic or strabismic amblyopia but tends to be ness in patients with amblyopia contradict histo-
more severe. Anisometropia and strabismus are diffi- pathologic studies in animal models of amblyopia,
cult to model in animals, so monocular deprivation in which cell densities have consistently been
(typically by reversible tarsorrhaphy) has become the reported as reduced after monocular deprivation. As
most widely studied animal model of amblyopia. As little as 1 week of monocular deprivation causes
such, comparisons between OCT studies of depriva- a reduction in cell density in all three nuclear layers
tion amblyopia are most directly comparable to find- in the rabbit retina (ganglion cell layer, inner nuclear
ings in histologic animal studies of amblyopia. layer, and outer nuclear layer) compared to both non-
Several studies have analyzed macular changes in deprived and control eyes.26 These effects increase
deprivation amblyopia. Al-Haddad et al.23 examined with longer durations of deprivation and are most
14 eyes with deprivation amblyopia secondary to pronounced in the ganglion cell layer. Interestingly,
cataract and 20 control eyes. They found increased there is a corresponding increase in cell density in all
central macular thickness compared to controls, and three layers of the fellow, non-deprived eye in all
they found that macular thickness correlated with three nuclear layers compared to control eyes.
the degree of amblyopia severity. Long et al.24 exam- Similar effects are seen in tree shrews,27 macaque
ined 53 children with unilateral cataracts 3 months monkeys,28 and chimpanzees.29 Other studies have
following cataract extraction, at which point patients shown overall thinning of the inner plexiform layer
had an average interocular visual acuity difference of the retina with deprivation.30,31
of 0.6 LogMAR (6 Snellen lines). They found no It is important to note that these experimental
difference in central macular retinal thickness com- models also carry their own limitations and potential
pared to fellow eyes. Kim et al.25 examined 14 chil- confounding factors. There may be technique-related
dren with unilateral amblyopia secondary to elements that drive the observed effect, such as poten-
cataracts and compared retinal thicknesses to those tially elevated intraocular pressure with tight lid
of 14 control subjects; they also found no difference. sutures and use of potentially toxic gentamicin during
Comparing segmented ganglion cell complex thick- surgery in the sutured eyes only. It is also possible
nesses (combined ganglion cell layer and inner that histologic processing selectively distorts retinal
plexiform layer) did not reveal any differences lamina to misrepresent that of the in vivo state.
between amblyopic and control eyes. As reported Nevertheless, these data provide consistent experi-
in the same study described above, Kasem & mental findings in animal models that contradict
Badawi4 examined 12 patients with derivational those observed in vivo in human deprivation amblyo-
amblyopia (9 resulting from congenital cataracts pia as assessed by OCT.
removed after the age of 2, and 3 cases of severe
ptosis). The mean visual acuity in these cases was
roughly 1.0 LogMAR (20/200 Snellen equivalent).
Deprived eyes showed an average increase in central CONFOUNDING FACTORS
macular thickness compared to fellow eyes, but
other parameters including average macular thick- Could the findings of increased retinal and/or RNFL
ness, and macular volume were not different. thicknesses represent an optical artifact? Indeed, differ-
Of those mentioned above, only two also compared ences in refractive error, including those imparted by
RNFL thicknesses between amblyopic and non- axial length, can induce a magnification effect that
amblyopic eyes. Kim et al.25 observed no difference directly impacts RNFL and optic disc measurements.32,33

Seminars in Ophthalmology
 Seminars in Ophthalmology 307

The effect of magnification with shorter axial lengths the direction of RNFL thickness changes compared
does produce an artifactually increased measurement of between hyperopic and myopic amblyopia is what
the RNFL, and correction of this effect largely eliminates one would expect if confounded by axial length. By
differences in RNFL thickness between refractive error contrast, Al-Haddad et al.38 found an increase in cen-
groups. Unfortunately, very few studies appropriately tral macular thickness in the amblyopic eyes of
account or correct for axial length when comparing patients with anisomyopia, but no difference in
amblyopic and fellow or control eyes. In the single RNFL thickness. Axial length was not measured or
study that found both increases in retinal and RNFL accounted for in their study.
thicknesses, Kasem & Badawi4 measured axial length If the difference in retinal thickness observed in
and analyzed the relationship between macular thickness amblyopic eyes is not related to refractive error, then
and axial length following adjustment for spherical we would also expect to see this difference present in
equivalent. They found that axial length and refractive patients with purely strabismic amblyopia. Of those
error (spherical equivalent) did correlate significantly, studies published recently, some studies show an
although with a significant degree of residual variance, increase in the macular retinal thickness,3,4 RNFL,4
presumably representing variation in corneal and lenti- and choroidal thickness,19 whereas other studies
cular refractive power. As adjusted axial length did not have found no difference in these parameters among
correlate with macular or RNFL thickness, they con- patients with strabismic amblyopia.8,9,18 These find-
cluded that the OCT results were not confounded by ings are largely consistent with these studies’ respec-
this variable. However, there were no adjustments tive findings in anisometropic amblyopia. Subtle
made in their OCT measurements, and the potential differences in refractive error may nevertheless drive
confounding effect of axial length and/or refractive which eye becomes amblyopic in the presence of stra-
error on their comparisons between amblyopic and fel- bismus (though not at standard thresholds of 1.50
low eyes are unknown. D of spherical anisometropia) and may therefore sys-
Kok et al.34 examined the relationship between tematically confound OCT analyses of strabismic
axial length and retinal thickness in amblyopic and amblyopia.
control eyes. In their comparative analysis (included Multiple OCT device types and models are used
in the meta-analysis performed by Li et al.2), they across these studies. OCT type was the largest con-
found no overall difference in the foveal or perifoveal tributor to variation in choroidal thickness among
retinal thickness among 36 children with unilateral patients with amblyopia in another meta-analysis.17
amblyopia (17 anisometropic, 11 strabismic, and 8 Of the six recent SD-OCT-based studies that found
mixed) compared to fellow eyes and 30 controls. increased retinal and/or RNFL thicknesses in amblyo-
There was a modest linear and inverse relationship pic eyes, three used Topcon devices (Tokyo, Japan;
between axial length and retinal thickness (shorter Oakland, NJ USA).3–5 Topcon devices were not used
axial length corresponding to increased retinal thick- in any study that showed no differences; primarily
ness) in the control group and not the amblyopic Zeiss (Dublin, CA, USA) or Heidelberg (Heidelberg,
group. However, the significant hyperopia in the Germany) devices were used. Therefore, OCT device
amblyopic group relative to the control group did may play a significant confounding role in assessment
not allow for direct comparison of this relationship of thickness among amblyopic and non-amblyopic
with similar axial lengths; it is unclear whether such groups, and could potentially contribute to the large
a relationship would persist if these non-amblyopic variation in results. While head-to-head studies com-
controls were matched for axial length. A similar paring OCT devices have shown considerable agree-
inverse relationship between retinal thickness and ment in normal-sighted subjects, no study has directly
axial length has been demonstrated among myopic compared these devices in the context of anisometro-
children.35,36 pia and/or amblyopia.
If the difference in retinal thickness observed in
amblyopic eyes was directly related to refractive
error, then we would expect there to be an inverse NEW AVENUES IN OCT IMAGING
effect of amblyopia on retinal and RNFL thickness in
anisomyopic amblyopia. Tekin et al.37 examined 42 OCT-Angiography (OCT-A) is a relatively new tech-
children with and without anisomyopic amblyopia nology that allows for motion-based imaging of the
and found no difference in central macular thickness. retinal and (peri)papillary microvasculature with
They did find a significantly thinner average RNFL micrometer precision. OCT-A has been used in the
thickness in the more myopic eye of both amblyopic study of several retinal and optic disc pathologies,
and non-amblyopic subjects. In the same study, the revealing associated changes in the micovasculature
relationships between both the degree of spherical of the posterior pole in vascular-based and what were
myopia and axial length separately and strongly traditionally considered non-vascular-based disor-
inversely correlated with RNFL thickness. Therefore, ders. Much like the literature concerning structural

© 2019 Taylor & Francis

308 E. D. Gaier et al.

OCT, results from OCT-A studies of amblyopia have achievable in these types of studies, these differences
been mixed. may not carry clinical relevance. Further study and
There have been 5 published OCT-A studies in advancement of our understanding of this new ima-
amblyopia to date.39–43 Of those that examined the ging technique will allow us to improve our interpre-
macular superficial capillary plexus, three have iden- tation of these results and guide future hypotheses
tified a reduction in vessel density associated with and study design.
amblyopia compared to fellow eyes and/or Swept source OCT (SS-OCT) uses a longer wave-
controls.39,42,43 One study found no difference in ves- length of light (1050 nm) that allows for deeper pene-
sel density in this segment.40 Among studies examin- tration of posterior structures and has recently been
ing the deep capillary plexus, two reported applied to the question of choroidal thickness differ-
a reduction in vessel density42,43 and one found no ences in amblyopic eyes. Using a small cohort of 13
difference.40 The foveal avascular zone was found to children (ages 6.2 ± 2.4 years) with anisohypermetro-
be of similar size in two-thirds studies40,43 and pic amblyopia, Araki et al.44 published the first and
enlarged in one study.39 Two studies examined the only study employing SS-OCT in amblyopia to date.
choriocapillaris using OCT-A, with one showing an They found an increase in the subfoveal and perifo-
increase in vessel density41 and the other showing no veal choroidal thicknesses using axial length as
difference.43 Currently, spectral domain-based OCT-A a covariate. Further study of larger and more diverse
is limited to imaging the choriocapilaris and cannot cohorts of patients with amblyopia using SS-OCT and
sufficiently image the choroid itself. swept source-based OCT-A will provide more
Conflicting results and conclusions among these detailed resolution to study choroidal structure in
studies are likely secondary to heterogeneity in refrac- amblyopia.
tive error and sample demographics. In addition, Other methods to image the microvasculature and
sampling and segmentation parameters differ signifi- blood flow of the posterior pole have been employed
cantly between devices and therefore could also con- to study amblyopia. Laser speckle flowgraphy has
tribute to differing results. Much like structural OCT, also been used to study amblyopia in two children
OCT-A is susceptible to technical limitations and arti- with anisohypermetropic amblyopia and showed
facts that may skew results. The two most significant comparative reductions in choroidal blood flow and
examples are axial refractive error and fixation increased choroidal thickness as measured with EDI-
instability; both can produce artifactually lower vessel OCT.45 Interestingly, the flow deficit and increased
densities in the amblyopic eye, and both would be thickness reversed with optical correction and patch-
expected to be worse with more severe amblyopia. ing therapy. Quantitative AF and split spectrum
Specific study of the influence of these potential con- amplitude-decorrelation angiography was employed
founders would be of benefit. In summary, OCT-A is by Guo et al.46 to study amblyopia in 22 patients with
a new and powerful mode of vascular imaging that unilateral amblyopia (13 anisometropic, 9 strabismic)
carries potentially confounding limitations as a tool to and found increased subfoveal chroroidal thickness,
study amblyopia. a blurry choriocapillary network and dark atrophic
It remains unclear why there would be a difference patches in a subset of amblyopic eyes. However, mag-
in microvascular density in amblyopia. Structural OCT nification effects secondary to differential refractive
studies find an increase in retinal thickness and chor- error and/or axial length were not accounted for in
oid if anything (although in their study Sobral et al.39 this study.
found a significantly reduced RNFL thickness corre- Dickmann et al.47 used microperimetry in conjunc-
sponding with the reported reduce vascular density), tion with OCT to report a reduction in macular sensi-
so one might expect a corresponding increase in the tivity in the amblyopic eyes of patients with
number of capillaries needed to serve these structures. strabismus and anisometropia. Of note, they also
Structural OCT studies have more consistently shown found no differences in retinal or RNFL thicknesses
an increase in choroidal density, and a corresponding overall, though the strabismic group did have signifi-
increase in the choriocapillaris density is consistent cantly increased macular thicknesses compared to fel-
with this.41 The vascular change might reflect develop- low eyes. Many OCT devices can be used to assess
mental abnormalities of the retina and choroid related fixation, and therefore eccentricity can be quantified.
to amblyopia. Whether these changes lead to or are Eccentric fixation has been shown to be present in
secondary to amblyopia is unclear. patients with amblyopia secondary to microtropia48
As alteration of the microvasculature is a relatively and patients with anisometropic amblyopia with and
new marker of posterior pole disease made possible without microtropia.49 Amblyopia is also associated
through OCT-A imaging, what changes represent with fixation instability,50–52 which may also pose
clinically relevant differences remain unclear at this a potential source of image degradation and error
point. Though statistically significant differences are that can potentially confound results.

Seminars in Ophthalmology
 Seminars in Ophthalmology 309

REVERSIBILITY et al.61 examined 44 children with amblyopia (20 ani-

sometropic, 5 strabismic, 12 mixed, 7 ametropic) fol-
Even if one considers the results from these various lowing treatment. After adjusting for differences in
approaches across multiple investigators to be largely axial length, thickness and volume of the parafoveal
in agreement, the findings discussed to this point are and perifoveal retina were similar between those suc-
merely correlative; they do not distinguish causal cessfully treated and those with residual amblyopia.
from non-causal changes with regard to amblyopia. There were also no differences in choroidal thickness.
It may be that the changes observed on OCT represent Reversal of differences in choroidal thickness
a morphological risk factor or structural consequence found on EDI-OCT has also been reported recently.
of amblyopia rather than amblyopia pathophysiology. Nishi et al.22 retrospectively studied 24 children with
One approach to distinguish whether these changes anisohypermetropic amblyopia and 23 control chil-
are directly related to amblyopia pathophysiology is dren of similar ages. Overall, choroidal thickness
to determine whether these structural changes in the was greater in amblyopic eyes before and after treat-
retina and choroid reverse with successful treatment. ment. Interestingly, they found that the change in
However, it is important to consider that while the choroidal thickness in response to optical correction
main outcome measure governing whether amblyo- depended on whether the choroid was thicker or
pia treatment is “successful” is visual acuity, several thinner than control eyes at baseline, but always
other functional visual deficits in amblyopia may per- toward normalization; thinner choroids thickened,
sist. Therefore, this approach does not address and thicker choroids thinned. There were no corre-
whether unchanged structural correlates represent sponding changes in axial length to explain this
non-causal associated changes or persistent changes observation. Aslan Bayhan et al.62 followed 40 chil-
directly involved in an incompletely treated disorder. dren with anisohypermetropic amblyopia through
Yoon & Chun53 measured the foveal thicknesses of refractive correction and/or occlusion therapy for 6
30 patients with unilateral anisometropic amblyopia months and compared choroidal thicknesses to 22
before and after occlusion therapy. Of their sample, 22 control children. After adjusting for refractive error
experienced successful resolution of amblyopia. There and axial length, they found a reduction in choroidal
were no differences in macular thickness prior to or thickness in amblyopic eyes with treatment, but the
after occlusion therapy, but foveal volume (not thick- statistical difference between amblyopic and control
ness) decreased significantly and consistently in eyes persisted.
amblyopic eyes following treatment. Nishi et al.54 In their study of 32 patients with anisometropic
examined 21 children with anisohypermetropic amblyopia, Oner & Bulut20 compared choroidal thick-
amblyopia and 25 age-matched control subjects mea- nesses before and after 6 months of occlusion therapy.
suring segments of retinal thickness. They found that They found the baseline significant increase in chor-
optical correction of amblyopia was associated with oidal thickness in the amblyopic eyes persisted fol-
an increase in outer segment thickness, normalizing lowing treatment with no changes. In their case series
the reduction in outer segment thickness measured at using SS-OCT, Araki et al.44 imaged the choroidal
baseline. The change in outer segment thickness thickness in their 13 children with anisohypermetro-
remained present at 1 year following treatment and pic amblyopia who were successfully treated using
correlated with the improvement in visual acuity of occlusion therapy. Choroidal thickness remained sig-
the amblyopic eye. Pang et al.55 focused on patients nificantly greater among amblyopic eyes compared to
with amblyopia secondary to high myopia and found fellow eyes following treatment. Therefore, there are
increase foveal and macular thicknesses in amblyopic conflicting reports on the reversibility of increased
eyes. Similar baseline results were found in a follow- retinal and choroidal thicknesses in amblyopic eyes
up study, in which 20 patients with anisomyopic with treatment.
amblyopia were prospectively followed through
amblyopic treatment with optical correction followed
by patching.56 They found that the average foveal SUMMARY/CONCLUSIONS
thickness significantly decreased, and the interocular
differences in foveal and macular thicknesses were OCT provides a powerful in vivo mode for imaging the
eliminated. retina and therefore can provide valuable insights into
By contrast, other studies have also found no dif- the structural changes associated with amblyopia.
ferences in macular and RNFL thicknesses before and Technological and study design limitations, along with
after occlusion therapy.57–59 Chen et al.60 conducted the employment of various devices likely account for
a cross-sectional study of 118 patients with amblyopia the large degree of heterogeneity in reported results.
that was untreated and treated and found no differ- Overall, studies that do show differences in retinal and
ences between amblyopia and fellow eyes or by treat- choroidal thicknesses tend to show increased thickness
ment condition after adjusting for axial length. Liu associated with amblyopia. A similar proportion of

© 2019 Taylor & Francis

310 E. D. Gaier et al.

studies suggests these differences are reversible with 11. Perez-Rico C, Garcia-Romo E, Gros-Otero J, et al.
standard optical and/or occlusion therapy. Since Evaluation of visual function and retinal structure in
adult amblyopes. Optom Vis Sci. 2015;92(3):375–383.
many of these studies are observational or correlative,
doi:10.1097/OPX.0000000000000492.
they do not allow us to definitively distinguish asso- 12. Singh N, Rohatgi J, Gupta VP, Kumar V. Measurement of
ciated changes or risk factors from changes central to the peripapillary retinal nerve fiber layer thickness and macu-
disease process. Nevertheless, these advancements lar thickness in anisometropia using spectral domain opti-
frame established findings and guide future studies in cal coherence tomography: a prospective study. Clin
Ophthalmol. 2017;11:429–434. doi:10.2147/OPTH.S123273.
translational science and animal models. There remains
13. Hansen MH, Munch IC, Li XQ, et al. Visual acuity and
a critical need for improving upon the biomarkers of amblyopia prevalence in 11- to 12-year-old Danish children
amblyopia to facilitate early detection and advance our from the Copenhagen Child Cohort 2000. Acta Ophthalmol.
understanding of this complex disorder of experience- 2018;97(1):29–35.
dependent neurodevelopment.63 14. Taskiran Comez A, Sanal Ulu E, Ekim Y. Retina and optic
disc characteristics in amblyopic and non-amblyopic eyes
of patients with myopic or hyperopic anisometropia. Turk
DECLARATION OF INTEREST J Ophthalmol. 2017;47(1):28–33. doi:10.4274/tjo.54289.
15. Yakar K, Kan E, Alan A, Alp MH, Ceylan T. Retinal nerve fibre
None of the authors have any financial interests or layer and macular thicknesses in adults with hyperopic ani-
sometropic amblyopia. J Ophthalmol. 2015;2015:946467.
conflicts directly relevant to this work. EDG is 16. Kantarci FA, Tatar MG, Uslu H, et al. Choroidal and peri-
a scientific advisor for Luminopia, a for-profit com- papillary retinal nerve fiber layer thickness in adults with
pany that develops technologies to treat amblyopia. anisometropic amblyopia. Eur J Ophthalmol. 2015;25
(5):437–442. doi:10.5301/ejo.5000594.
17. Liu Y, Dong Y, Zhao K. A meta-analysis of choroidal
thickness changes in unilateral amblyopia. J Ophthalmol.
REFERENCES 2017;2017:2915261. doi:10.1155/2017/2915261.
18. Niyaz L, Yucel OE, Ariturk N, Terzi O. Choroidal thick-
1. Yen MY, Cheng CY, Wang AG. Retinal nerve fiber layer ness in strabismus and amblyopia cases. Strabismus.
thickness in unilateral amblyopia. Invest Ophthalmol Vis Sci. 2017;25(2):56–59. doi:10.1080/09273972.2017.1318152.
2004;45:2224–2230. 19. Al-Haddad C, Fattah MA, Ismail K, Bashshur Z. Choroidal
2. Li J, Ji P, Yu M. Meta-analysis of retinal changes in uni- changes in anisometropic and strabismic children with uni-
lateral amblyopia using optical coherence tomography. Eur lateral amblyopia. Ophthalmic Surg Lasers Imaging Retina.
J Ophthalmol. 2015;25(5):400–409. doi:10.5301/ejo.5000583. 2016;47(10):900–907. doi:10.3928/23258160-20161004-02.
3. Rajavi Z, Sabbaghi H, Behradfar N, Yaseri M, Aghazadeh 20. Oner V, Bulut A. Does the treatment of amblyopia normalise
Amiri M, Faghihi M. Macular thickness in moderate to subfoveal choroidal thickness in amblyopic children? Clin Exp
severe amblyopia. Korean J Ophthalmol. 2018;32(4):312–318. Optom. 2017;100(2):184–188. doi:10.1111/cxo.12483.
doi:10.3341/kjo.2017.0101. 21. Nishi T, Ueda T, Hasegawa T, Miyata K, Ogata N.
4. Kasem MA, Badawi AE. Changes in macular parameters in Choroidal thickness in children with hyperopic anisome-
different types of amblyopia: optical coherence tomography tropic amblyopia. Br J Ophthalmol. 2014;98(2):228–232.
study. Clin Ophthalmol. 2017;11:1407–1416. doi:10.2147/OPTH. doi:10.1136/bjophthalmol-2013-303938.
S143223. 22. Nishi T, Ueda T, Mizusawa Y, et al. Effect of optical correc-
5. Kuhli-Hattenbach C, Koss MJ, Kohnen T, Fronius M. tion on subfoveal choroidal thickness in children with
A morphological study of amblyopic eyes in children fail- anisohypermetropic amblyopia. PLoS One. 2017;12(12):
ing to achieve normal visual acuity after electronically e0189735. doi:10.1371/journal.pone.0189735.
monitored long-term occlusion treatment. Graefes Arch 23. Al-Haddad C, Mehanna CJ, Ismail K. High-definition opti-
Clin Exp Ophthalmol. 2015;253(11):2021–2028. doi:10.1007/ cal coherence tomography of the macula in deprivational
s00417-015-3085-0. amblyopia. Ophthalmic Surg Lasers Imaging Retina. 2018;49
6. Demircan S, Gokce G, Yuvaci I, Atas M, Baskan B, (3):198–204. doi:10.3928/23258160-20180221-08.
Zararsiz G. The assessment of anterior and posterior ocular 24. Long E, Chen J, Liu Z, et al. Interocular anatomical and
structures in hyperopic anisometropic amblyopia. Med Sci visual functional differences in pediatric patients with uni-
Monit. 2015;21:1181–1188. doi:10.12659/MSM.893979. lateral cataracts. BMC Ophthalmol. 2016;16(1):192.
7. Chen W, Xu J, Zhou J, et al. Thickness of retinal layers in the doi:10.1186/s12886-016-0371-5.
foveas of children with anisometropic amblyopia. PLoS One. 25. Kim YW, Kim SJ, Yu YS. Spectral-domain optical coher-
2017;12(3):e0174537. doi:10.1371/journal.pone.0174537. ence tomography analysis in deprivational amblyopia:
8. Lekskul A, Wuthisiri W, Padungkiatsagul T. Evaluation of a pilot study with unilateral pediatric cataract patients.
retinal structure in unilateral amblyopia using spectral Graefes Arch Clin Exp Ophthalmol. 2013;251(12):2811–2819.
domain optical coherence tomography. J Aapos. 2018;22 doi:10.1007/s00417-013-2494-1.
(5):386–389. doi:10.1016/j.jaapos.2018.05.014. 26. Mwachaka PM, Saidi H, Odula PO, Mandela PI. Effect of
9. Kusbeci T, Karti O, Karahan E, Oguztoreli M. The evalua- monocular deprivation on rabbit neural retinal cell
tion of anatomic and functional changes in unilateral mod- densities. J Ophthalmic Vis Res. 2015;10(2):144–150.
erate amblyopic eyes using optical coherence tomography doi:10.4103/2008-322X.163770.
and pupil cycle time. Curr Eye Res. 2017;42(12):1725–1732. 27. Abbott CJ, Grunert U, Pianta MJ, McBrien NA. Retinal
doi:10.1080/02713683.2017.1349153. thinning in tree shrews with induced high myopia: optical
10. Altintas O, Gumustas S, Cinik R, Anik Y, Ozkan B, coherence tomography and histological assessment. Vision
Karabas L. Correlation of the measurements of optical Res. 2011;51(3):376–385. doi:10.1016/j.visres.2010.12.005.
coherence tomography and diffuse tension imaging of 28. Von Noorden GK, Crawford ML, Middleditch PR. Effect of
optic pathways in amblyopia. Int Ophthalmol. 2017;37 lid suture on retinal ganglion cells in macaca mulatta. Brain
(1):85–93. doi:10.1007/s10792-016-0229-0. Res. 1977;122:437–444.

Seminars in Ophthalmology
 Seminars in Ophthalmology 311

29. Chow KL, Riesen AH, Newell FW. Degeneration of retinal 46. Guo L, Tao J, Xia F, Yang Z, Ma X, Hua R. In vivo optical
ganglion cells in infant chimpanzees reared in darkness. imaging of amblyopia: digital subtraction autofluorescence
J Comp Neurol. 1957;107:27–42. and split-spectrum amplitude-decorrelation angiography.
30. Rasch E, Swift H, Riesen AH, Chow KL. Altered structure Lasers Surg Med. 2016;48(7):660–667. doi:10.1002/lsm.22520.
and composition of retinal cells in darkreared mammals. 47. Dickmann A, Petroni S, Perrotta V, et al. A morpho-functional
Exp Cell Res. 1961;25:348–363. study of amblyopic eyes with the use of optical coherence
31. Weiskrantz L. Sensory deprivation and the cat’s optic ner- tomography and microperimetry. J Aapos. 2011;15(4):338–341.
vous system. Nature. 1958;181:1047–1050. doi:10.1016/j.jaapos.2011.03.019.
32. Aykut V, Oner V, Tas M, Iscan Y, Agachan A. Influence of 48. Garcia-Garcia MA, Belda JI, Schargel K, et al. Optical
axial length on peripapillary retinal nerve fiber layer thick- coherence tomography in children with microtropia.
ness in children: a study by RTVue spectral-domain optical J Pediatr Ophthalmol Strabismus. 2018;55(3):171–177.
coherence tomography. Curr Eye Res. 2013;38 doi:10.3928/01913913-20171026-01.
(12):1241–1247. doi:10.3109/02713683.2013.820328. 49. Nakamoto Y, Takada R, Tanaka M, et al. Quantification of
33. Savini G, Barboni P, Parisi V, Carbonelli M. The influence eccentric fixation using spectral-domain optical coherence
of axial length on retinal nerve fibre layer thickness and tomography. Ophthalmic Res. 2018;60(4):231–237.
optic-disc size measurements by spectral-domain OCT. Br doi:10.1159/000493487.
J Ophthalmol. 2012;96(1):57–61. doi:10.1136/bjo.2010.196782. 50. Subramanian V, Jost RM, Birch EE. A quantitative study of
34. Kok PH, de Kinkelder R, Braaksma-Besselink YC, et al. fixation stability in amblyopia. Invest Ophthalmol Vis Sci.
Anomalous relation between axial length and retinal thick- 2013;54(3):1998–2003. doi:10.1167/iovs.12-11054.
ness in amblyopic children. J Aapos. 2013;17(6):598–602. 51. Shaikh AG, Otero-Millan J, Kumar P, Ghasia FF. Abnormal
doi:10.1016/j.jaapos.2013.09.005. fixational eye movements in amblyopia. PLoS One. 2016;11
35. Luo HD, Gazzard G, Fong A, et al. Myopia, axial length, (3):e0149953. doi:10.1371/journal.pone.0149953.
and OCT characteristics of the macula in Singaporean 52. Kelly KR, Cheng-Patel CS, Jost RM, Wang YZ, Birch EE.
children. Invest Ophthalmol Vis Sci. 2006;47(7):2773–2781. Fixation instability during binocular viewing in anisome-
doi:10.1167/iovs.05-1380. tropic and strabismic children. Exp Eye Res. 2018.
36. Lim MC, Hoh ST, Foster PJ, et al. Use of optical coherence doi:10.1016/j.exer.2018.07.013.
tomography to assess variations in macular retinal thick- 53. Yoon DH, Chun BY. Comparison of the thickness and
ness in myopia. Invest Ophthalmol Vis Sci. 2005;46 volume of the macula and fovea in patients with anisome-
(3):974–978. doi:10.1167/iovs.04-0828. tropic amblyopia prior to and after occlusion therapy.
37. Tekin K, Cankurtaran V, Inanc M, Sekeroglu MA, Korean J Ophthalmol. 2018;32(1):52–58. doi:10.3341/
Yilmazbas P. Effect of myopic anisometropia on anterior kjo.2016.0127.
and posterior ocular segment parameters. Int Ophthalmol. 54. Nishi T, Ueda T, Hasegawa T, Miyata K, Ogata N. Retinal
2017;37(2):377–384. doi:10.1007/s10792-016-0272-x. thickness in children with anisohypermetropic amblyopia.
38. Al-Haddad CE, Mollayess GM, Cherfan CG, Jaafar DF, Br J Ophthalmol. 2015;99(8):1060–1064. doi:10.1136/
Bashshur ZF. Retinal nerve fibre layer and macular thick- bjophthalmol-2014-305685.
ness in amblyopia as measured by spectral-domain optical 55. Pang Y, Goodfellow GW, Allison C, Block S, Frantz KA.
coherence tomography. Br J Ophthalmol. 2011;95 A prospective study of macular thickness in amblyopic
(12):1696–1699. doi:10.1136/bjo.2010.195081. children with unilateral high myopia. Invest Ophthalmol
39. Sobral I, Rodrigues TM, Soares M, et al. OCT angiography Vis Sci. 2011;52(5):2444–2449. doi:10.1167/iovs.10-5550.
findings in children with amblyopia. J Aapos. 2018;22 56. Pang Y, Frantz KA, Block S, Goodfellow GW, Allison C.
(4):286–289.e282. doi:10.1016/j.jaapos.2018.03.009. Effect of amblyopia treatment on macular thickness in eyes
40. Demirayak B, Vural A, Onur IU, Kaya FS, Yigit FU. with myopic anisometropic amblyopia. Invest Ophthalmol
Analysis of macular vessel density and foveal avascular Vis Sci. 2015;56(4):2677–2683. doi:10.1167/iovs.14-15532.
zone using spectral-domain optical coherence tomography 57. Miki A, Shirakashi M, Yaoeda K, et al. Retinal nerve fiber
angiography in children with amblyopia. J Pediatr layer thickness in recovered and persistent amblyopia. Clin
Ophthalmol Strabismus. 2018;56(1):55–59. Ophthalmol. 2010;4:1061–1064. doi:10.2147/OPTH.S13145.
41. Borrelli E, Lonngi M, Balasubramanian S, et al. Increased 58. Tugcu B, Araz-Ersan B, Erdogan ET, et al. Structural and
choriocapillaris vessel density in amblyopic children: a functional comparison of the persistent and resolved
case-control study. J Aapos. 2018;22(5):366–370. doi:10.1016/j. amblyopia. Documenta Ophthalmol Adv Ophthalmol.
jaapos.2018.04.005. 2014;128(2):101–109. doi:10.1007/s10633-013-9422-x.
42. Lonngi M, Velez FG, Tsui I, et al. Spectral-domain optical 59. Yassin SA, Al-Tamimi ER, Al-Hassan S. Macular and ret-
coherence tomographic angiography in children with inal nerve fiber thickness in recovered and persistent
amblyopia. JAMA Ophthalmol. 2017;135(10):1086–1091. amblyopia. Int Ophthalmol. 2015;35(6):833–842.
doi:10.1001/jamaophthalmol.2017.3423. doi:10.1007/s10792-015-0055-9.
43. Yilmaz I, Ocak OB, Yilmaz BS, Inal A, Gokyigit B, 60. Chen W, Chen J, Huang J, Xu J, Zhang F, Lu F. Comparison
Taskapili M. Comparison of quantitative measurement of of macular and retinal nerve fiber layer thickness in
foveal avascular zone and macular vessel density in eyes of untreated and treated binocular amblyopia. Curr Eye Res.
children with amblyopia and healthy controls: an optical 2013;38(12):1248–1254. doi:10.3109/02713683.2013.805233.
coherence tomography angiography study. J Aapos. 2017;21 61. Liu CH, Ong SJ, Huang CY, Wu WC, Kao LY, Yang ML.
(3):224–228. doi:10.1016/j.jaapos.2017.05.002. Macular thickness, foveal volume, and choroidal thickness
44. Araki S, Miki A, Goto K, et al. Effect of amblyopia treatment on in amblyopic eyes and their relationships to the treatment
choroidal thickness in hypermetropic anisometropic amblyo- outcome. J Ophthalmol. 2018;2018:1967621. doi:10.1155/
pia using swept-source optical coherence tomography. BMC 2018/1967621.
Ophthalmol. 2018;18(1):227. doi:10.1186/s12886-018-0894-z. 62. Aslan Bayhan S, Bayhan HA. Effect of amblyopia treat-
45. Hashimoto R, Kawamura J, Hirota A, Oyamada M, ment on choroidal thickness in children with hyperopic
Sakai A, Maeno T. Changes in choroidal blood flow and anisometropic amblyopia. Curr Eye Res. 2017;42
choroidal thickness after treatment in two cases of pedia- (9):1254–1259. doi:10.1080/02713683.2017.1315141.
tric anisohypermetropic amblyopia. Am J Ophthalmol Case 63. Hunter D, Cotter S. Early diagnosis of amblyopia. Vis
Rep. 2017;8:39–43. doi:10.1016/j.ajoc.2017.10.010. Neurosci. 2018;35:E013. doi:10.1017/S0952523817000207.

© 2019 Taylor & Francis