s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0

Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/survophthal

Review article

Virtual reality and augmented reality— emerging

screening and diagnostic techniques in
ophthalmology: A systematic review

Marco King In Ma a, Chinmoy Saha b, Stephanie Hiu Ling Poon b,

Rachel Sze Wan Yiu, PhD c, Kendrick Co Shih, FCOphth(HK) b,
Yau Kei Chan, PhD b,∗
a LiKa Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, China
b Department of Ophthalmology, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, China
c Department of Ophthalmology, Grantham Hospital, Hong Kong, China

a r t i c l e i n f o a b s t r a c t

Article history: In health care, virtual reality (VR) and augmented reality (AR) have been applied extensively
Received 13 December 2021 for many purposes. Similar to other technologies such as telemedicine and artificial intelli-
Revised 7 February 2022 gence, VR and AR may improve clinical diagnosis and screening services in ophthalmology
Accepted 11 February 2022 by alleviating current problems, including workforce shortage, diagnostic error, and under-
Available online 15 February 2022 diagnosis. In the past decade a number of studies and products have used VR and AR con-
cepts to build clinical tests for ophthalmology, but comprehensive reviews on these studies
Keywords:
are limited. Therefore, we conducted a systematic review on the use of VR and AR as a di-
Virtual reality
agnostic and screening tool in ophthalmology. We identified 26 studies that implemented
Augmented reality
a variety of VR and AR tests on different conditions, including VR cover tests for binocu-
Ophthalmology
lar vision disorder, VR perimetry for glaucoma, and AR slit lamp biomicroscopy for retinal
Diagnosis
diseases. In general, while VR and AR tools can become standardized, automated, and cost-
Screening
effective tests with good user experience, several weaknesses, including unsatisfactory ac-
Monitoring
curacy, weak validation, and hardware limitations, have prevented these VR and AR tools
Glaucoma
from having wider clinical application. Also, a comparison between VR and AR is made to
Perimetry
explain why studies have predominantly used VR rather than AR.
Cover test
© 2022 Elsevier Inc. All rights reserved.

and robotic surgeries.2 , 36 By 2023, the market for VR and AR

1. Introduction
technologies in health care application will potentially reach
nearly US$20 billion.A VR technologies create a simulated
In health care, virtual reality (VR) and augmented reality
world, providing a seemingly real and possibly interactive ex-
(AR) have been applied for various purposes, including post-
perience within the virtual space.13 , 22 In contrast, AR tech-
stroke cognitive rehabilitation, surgical training, laparoscopic

∗
Corresponding author: Yau Kei Chan, PhD, Department of Ophthalmology, Li Ka Shing Faculty of Medicine, University of Hong Kong.
E-mail address: jchanyk@hku.hk (Y.K. Chan).

0039-6257/$ – see front matter © 2022 Elsevier Inc. All rights reserved.
https://doi.org/10.1016/j.survophthal.2022.02.001
 s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0
nologies overlay digital information, such as images, onto the
real physical world to enhance our experience.13 , 22
Clinical diagnosis and screening services in ophthalmol-
ogy have encountered problems, including diagnostic er-
rors and workforce shortage. On one hand, in this spe-
cialty “error in diagnosis” is among the top 3 reasons be-
hind medical professional liability claims, compromising pa-
tients’ health and resulting in multimillion-dollar indem-
nity.18 , 47 These errors stemmed from cognitive bias and in-
accurate clinical information, and clinicians often need de-
tailed examinations to identify an elusive diagnosis.18 , 44 On
the other hand, while developing countries are currently
plagued by a severe shortage of ophthalmologists, in the fu-
ture developed countries will need to overcome a shortfall as
well.40
Clinicians and scientists have investigated on multiple new
digital technologies to alleviate the aforementioned issues.
Telemedicine breaks the geographical barriers and hence
makes distribution of health care service more efficient and
equitable.22 As a diagnostic modality for diabetic retinopa-
thy, telemedicine is used in the United States and was proven
cost-effective in a developing country setting.38 , 43 It may also
improves access to cytomegalovirus retinitis screening, a de-
bilitating condition found most commonly in immunocom-
promised patients.43 Meanwhile, machine learning and arti-
ficial intelligence can analyze ophthalmic images in the di-
agnostic and screening process, allowing an earlier detec-
tion of various ophthalmic diseases including glaucoma, dia-
betic retinopathy, and age-related macular degeneration.22 A
good example is that artificial intelligence and machine learn-
ing can now diagnose papilledema and other optic disk ab-
normalities from fundus photographs accurately,1 , 30 with a
study showing that their deep learning system is compara-
ble to expert neuro-ophthalmologists.9 Besides, VR and AR
are popular digital technologies that emphasize real-time in-
teraction between human and simulated objects and may
be helpful in diagnosis and screening in ophthalmology as
well.5
In recent years, research groups and companies have de-
veloped several clinical and commercialized VR and AR ap-
plications for the field of ophthalmology, including visual re-
habilitation, home monitoring, and aids for the visually im-
paired.22 A recent review offered a broad overview on the
use of VR and AR across different aspects of the field of
ophthalmology, demonstrating that VR and AR are suitable
for this specialty, with most studies focused on training, di-
agnosis/screening, and therapy;19 however, as it aimed to
provide a general picture with a wide scope, discussion on
each application was brief. Among the 3 main applications–
training, diagnosis/screening, and therapy–previous reviews
have only thoroughly discussed the use of VR or AR in edu-
cation and surgical training.6 , 20 , 48 Meanwhile, clinical reviews
on VR/AR diagnosis and screening tools are limited in cur-
rent literature. Hence, we have conducted a systematic re-
view on the use of VR and AR as a screening and diagnos-
tic tool in ophthalmology. By discussing the relevant transla-
tional research and limitations, we aim to update ophthalmol-
ogists and other professionals on recent developments and
application of VR and AR in diagnosing and screening for eye
diseases.
1517
2. Methods
The Preferred Reporting Items for Systematic Reviews and
Meta-Analyses (PRISMA) 2020 Statement33 were followed
when conducting and reporting this review. PRISMA 2020 ex-
planation and elaboration34 was deployed as a guidance when
reporting. The screening process was based on a list with
766 records generated by the systematic and final search on
PubMed on 16th July, 2021. The keywords “((Virtual reality) OR
(Augmented reality) OR (Mixed reality)) AND ((ophthalmology)
OR (eye) OR (ophthalmic) OR (ocular) OR (orbit)) AND ((Diag-
nosis) OR (Diagnosing) OR (Diagnostic) OR (Detection) OR (De-
tect) OR (Screening) OR (screen) OR (Examination) OR (exam-
ine) OR (Severity) OR (assessment) OR (assess) OR (test))” and
filter “English” were applied.
The same inclusion and exclusion criteria have been used
consistently throughout the process. Articles included shall
involve original findings regarding the use of VR or AR in di-
agnosis and screening in ophthalmology, discuss application
on any ophthalmic diseases or conditions, and be written in
English. Articles without original findings regarding our topic
of interest, or studies purely discussing disease disability and
rehabilitation would be excluded.
M.K.I.M. and C.S. independently conducted the search,
screened abstracts, and reviewed full texts. Consensus was
reached to resolve any disagreement. In the process of selec-
tion and extraction, Y.K.C. was the supervisor and was con-
sulted on any unresolved disagreement.
3. Results
We screened 766 publications and finally included 26 articles
(Fig. 1, PRISMA flow diagram based on a template provided in
PRISMA 2020 statement33 ). We have summarized the impor-
tant quantitative findings in Tables 1–5. While most studies
with validated or designed specific VR or AR tests focus on 1
condition, 2 studies covered multiple ophthalmic conditions
and focused on hardware testing or design.26 , 51 To avoid re-
peating the latter 2 studies in each subsection corresponding
to 1 eye condition below, they will be discussed in the “Other
diseases or conditions” section.
3.1. Glaucoma
Several studies have proposed the use of VR on screening and
diagnosing glaucoma. In most studies different VR devices
were deployed to implement perimetry, which is a form of
systematic visual field testing.46 In the remaining studies re-
searchers used VR to conduct visual evoked potentials (VEPs)-
based tests. All studies have used the Humphrey Field Ana-
lyzer (HFA) or clinical record of patients to validate the ac-
curacy of the VR tests. In this section, we first introduce VR
perimetries according to their applications, starting from di-
agnosis, screening, and finally visual field testing or monitor-
ing. Then, we discuss studies using visually evoked potentials
with VR Table 1. provides a summary of the quantitative find-
ings and important characteristics of these studies.
 1518
Table 1 – VR glaucoma tests.

Year Proposed Main Objective Subjects Investigated measure Comparison Key findings Reference
Application technique

2021 Diagnosis of mild To investigate the 25 healthy VR head-mounted HFA Correlation between the mean sensitivity Razeghinejad
and moderate performance subjects, and 26 perimetry with measured by VR perimetry and that by HFA et al39
glaucoma of a new VR head-mounted patients with eye-tracking function, result
visual perimetry mild or moderate implementing a central - All quadrants: significant correlation in both
in normal and glaucoma glaucoma 24–2 algorithm patients and control group
patients - Global: significant correlation in both normal
(P = 0.001) and glaucoma (P < 0.001) groups.
For discriminating glaucoma from normal, VR

s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0
perimetry mean sensitivity had a greater (0.98)
receiver operating characteristic curve than HFA
(0.93) mean sensitivity (P = 0.06).
Significantly longer operation time for both
groups of patients.
2021 Visual field test To describe a new VR-based 3 healthy Same as above N/A Mean testing time (simultaneous testing of both Montelongo
and at-home platform for evaluating subjects (Razeghinejad et al39 ), eyes): 8 minutes and 53 seconds et al31
glaucoma visual fields. with the use of a Successful test completion and generation of
monitoring. protocol equivalent to threshold values at each testing locations in
HVF 24–2. the assessed visual field.
2000 Screening of To compare the 7 normal VR-based scanning laser Standard Based on clinical classification of glaucoma Plummer
moderate and effectiveness of scanning subjects and 16 entoptic perimetry, with automated severity: et al37
severe glaucoma laser entoptic perimetry glaucoma different analytic HVF Sensitivity
(and retinal for screening asymptomatic patients approach algorithms [(1) perimetry - high for moderate and severe scotomas
disease) glaucomatous damage by cluster or pointwise (0.71–0.9)
within the central 60° of deviation; (2) by eye or - low for normal/early/suspect subjects
vision. by subject] (0.27–0.67)
Specificity: 0.78-1
Among all algorithms of the VR perimetry,
pattern cluster and by subject analysis had the
best predictive power.
2020 Screening of To investigate a VR 62 glaucoma VR head-mounted HFA 24–2 The area under the receiving Mees et al28
moderate and head-mounted visual field patients and 95 perimetry (presented SITA operator characteristic curve:
advanced test as a possible subjective healthy control stimuli in the same Standard test - 0.77 ± 0.06 for mild glaucoma
glaucoma field test for glaucoma subjects positions as the HFA - 0.86 ± 0.04 for moderate and advanced
screening and eventually 24–2 SITA Standard test) glaucoma
glaucoma monitoring. The VR perimetry was unreliable in identifying
visual field defects that matched HFA
Patient’s experience
- 93% preferred the VR test to the HFA.
- The VR test was easier to use (P < 0.001) and
more comfortable (P < 0.001) than the HFA.
- 60% willing to use the VR perimetry at home

(continued on next page)

Table 1 (continued)

Year Proposed Main Objective Subjects Investigated measure Comparison Key findings Reference
Application technique

2018 Home-based To evaluate a home-based 10 patients in Suprathreshold Humphrey Area under Receiver operating characteristic Tsapakis
telemedicine visual field examination glaucoma algorithm for central 24 perimetry curve: 0.762–0.837 (P < 0.001) et al50
glaucoma method using a PC monitor department degree of visual field, Sensitivity: 0.637–0.942

s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0
screening or VR glasses presented by LCD Specificity: 0.735–0.497
monitor Same technique can be implemented by
smartphone VR, and comparable results would
be expected
2014 Evaluation of To report 59 patients (glau- VR Head-mounted HFA II Results of measurements Wroblewski
visual field development and clinical coma/glaucoma perimetry - Minimal systematic differences between the 2 et al54
testing of a compact, suspect) implementing an modes
head-mounted, and algorithm simulating - a systematic shift (of 4–6 dB) to lower
eye-tracking perimetry full threshold 24–2 test sensitivities for VR devices, mostly in high dB
(with a manual mode range.
and a visual grasp Usability Test: VR perimetry preferred over HFA II
mode) because it is more comfortable and easier to use
2020 Remote visual To implement and monitor 1 glaucoma Ocular kinetic perimetry HVF The VR-OKP technique showed correspondence Deiner et al14
field test at-home VR oculo-kinetic patient (Case (OKP) technique with HVF and the result is repeatable.
perimetry during COVID-19 report) remotely installed on
VR headset
2017 Visual field To present and evaluate a 10 patients in VR perimetry testing for Humphrey High correlation coefficient (r = 0.808, P < 0.0001) Tsapakis
examination VR visual field examination glaucoma central 24 degree of perimetry between VR test and Humphrey perimetry et al49
method. department visual field
20063 , Assessment of To prove the feasibility of 10 healthy mfVEP recording under N/A3 ; Feasibility of dichoptic VR stimulation and Arvind et al3 , 4
20074 visual field loss3 ; investigated measure,3 and volunteers3 ; 28 VR dichoptic Monocular mfVEP recording is confirmed.3
diagnosis of early determine its effectiveness glaucoma stimulation mfVEP and Identification of all defects found by
glaucoma4 in detecting glaucomatous patients, 30 HFA4 HFA/monocular mfVEP, as well as early defects
defects.4 healthy subjects4 undetected in those comparison techniques4

AR = augmented reality; HFA= Humphrey’s field analyzer; HVF = Humphrey visual field; mfVEP= multifocal visual evoked potential; VR= virtual reality.

1519
1520 s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0

Fig. 1 – Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart based on the template
provided by the PRISMA 2020 statement.1 The PRISMA flow chart showed that we started with 766 articles identified from
PubMed and ultimately included 26 articles after abstract screening and full text review. (Reprinted from Page et al33 with
permission of The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ.)

The first application of VR perimetry is the diagnosis of
mild and moderate glaucoma. A recent case-control study in-
vestigated a new VR head-mounted device visual perimetry
in which a full threshold strategy was deployed to test a range
of 24 degree of visual field.39 The eye trackers and the sys-
tem ensure that the test to be conducted under central gaze
fixation. Researchers used this VR perimetry and HFA on 25
healthy participants and 26 mild or moderate glaucoma pa-
tients to assess the sensitivity of their visual fields. In both
subject groups the 2 tests produced significantly correlated re-
sults, including the quadrants and global mean sensitivity. Us-
ing the retinal sensitivity to discriminate between glaucoma
and normal subjects, the VR method had an excellent diagnos-
tic performance, with a receiver operating characteristic curve
(0.98) greater than that of HFA (0.93). In contrast to the HFA,
during the VR test, patients could be in different positions and
move their head, reducing discomfort and fatigue. VR perime-
try is user-friendly and applicable in telemedicine. Limitations
includes a long finishing time, lack of follow-up analysis, and
aberrations due to flat screens.39 A case series showed that
the same headset could complete a HFA 24-2 equivalent test,
suggesting at-home monitoring and management as potential
applications.31
Apart from diagnostic purposes, 2 studies have proposed
using VR perimetry to screen for glaucoma and found the VR
perimetry to have a better sensitivity in screening moderate
or severe glaucoma, but lower sensitivity in mild disease.28 , 37
The first published in 2000 compared the effectiveness of a
VR scanning laser entoptic perimetry with Humphrey perime-
try. In the VR test, subjects’ report on persistently abnor-
mal areas would be analyzed to determine visual field ab-
normalities. For moderate or severe glaucoma patients, the
test achieved relatively high sensitivity (0.71–0.9) and per-
fect specificity.37 In a study published in 2020, a VR head-
mounted visual field-testing device presented stimuli in HFA
24-2 SITA testing positions and used a suprathreshold algo-
rithm. Although it was not as reliable as the HFA in iden-
tifying visual field defects, it could identify glaucoma pa-
tients, especially those with moderate to advanced glau-
coma. Most patients (93%) preferred the VR device to HFA
because the VR device was more comfortable and easier to
use.28
Recently a study on glaucoma patients suggested a VR
perimetry as a probable home-based telemedicine screening
tool of glaucoma. In that study, stimuli from an LCD screen
were adjusted to appear on the retina as if they were pro-
jected on a classic bowl perimeter. Subjects under central fix-
ation responded to the stimuli by clicking. The perimetry im-
plemented suprathreshold and ophthalmokinetic algorithms
to evaluate the central 24 degree of visual field. During the test,
 s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0 1521

Table 2 – VR and AR tests in Binocular vision disorders.

Year Proposed Main Objective Subjects Investigated Comparison Key findings Reference
Application measure technique

2021 Evaluation of To investigate the 38 strabismus VR-based Conventional • ICC of ocular deviation Yeh et al56
ocular accuracy of an patients APCT APCT angle: 0.897 (overall),
deviation eye-tracking virtual 0.962 (esotropia group),
angle in reality (VR)-based ocular 0.862 (exotropia group)
strabismus. deviation • Esotropia:
measurement system in overestimation (4.65 PD)
strabismus patients • Exotropia:
underestimation
(-3.01PD)

2020 Identification To develop and 5 orthotropia, VR-based Conventional Mean difference Miao
and investigate an 12 exotropia APCT & ACT PCT by between the 2 VR-based et al29
assessment automated VR subjects doctors tests and conventional
of ocular technique for APCT was less than 0.7
deviation measurement of ocular degree (insignificant
deviation difference) for all
patients.
2017 Evaluation To compare the 3 patients VR headset- Traditional Deviation patterns
and performance of a with ocular based Lees screen obtained from the Nesaratnam
automated VR-based test of ocular misalignment deviation test VR-based test were in et al32
diagnosis of misalignment to that of pattern test agreement with Lees
ocular the traditional Lees test.
misalignment screen, and establish VR headset could be
the feasibility of using used to assess
virtual reality misalignment and
technology in conduct dissociative
ophthalmic test in strabismus
settings in the future.
2009 Measurement To investigate a 40 VR display Major Strong correlation (P < Handa
of binocular hemispherical visual healthy subje- system-based amblyoscope 0.0001) between the et al16
vision display system which binocular results obtained from
function tests binocular visual cts- visual the VR testing system
function intervention function test and the Major
Using dichoptic stimuli. and control Amblyoscope in the
following binocular
vision perimeters:
- Subjective angle of
strabismus
- Motor fusion
amplitude (convergence
and divergence)

ACT = alternate cover tests; APCT = alternate prism cover test; AR = augmented reality; ICC = Intraclass correlation coefficient; PD = prism
diopter; PCT = Prism cover test; VR = virtual reality; year = year of publication.

the system also monitored fixation losses. The same test can
alternatively be implemented on VR glasses using a smart-
phone, and the authors expected comparable results given
similar display technology. To overcome the subjective error
and unreliability in telemedicine visual field tests conducted
by patients, the system monitored ambient luminosity by a
web camera, rejected improperly conducted tests by software
validation, and combined results from multiple attempts to
improve statistical accuracy. Using HFA perimetry as a ref-
erence, the receiver operating characteristic/area under the
curve coefficient of the VR test ranged from 0.762 to 0.837 (P <
0.001). Apart from the accuracy, the test was short, easy, and
simple, and did not require any specialized equipment. There-
fore, the test was proposed to be used in glaucoma screen-
ing.50
VR perimetry has also been applied to visual field test-
ing and at-home monitoring of glaucoma. In a cross-sectional
study, a VR perimeter, consisting of a VR display and eye-
tracker simulated full threshold HFA 24-2 tests in 2 modes.
The first mode resembled standard perimetry, and the sec-
ond mode utilized the eye movement induced by stimula-
tions and thus required no manual input. The 2 modes of
VR tests produced similar results, but the retinal sensitivities
measured by the VR tests were systematically shifted (around
-5dB) compared to the HFA results. The VR test was better re-
ceived than HFA by patients and was better able to monitor
1522 s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0

Table 3 – VR test for amblyopia.

Year Proposed Main objective Subjects Investigated Comparison Key findings Reference
application measure technique

2012 Quantitative To describe an N/A VR amblyopia N/A This balance point Black et al11
measure- amblyopic suppression technique was
ment of suppression test using a psychophysically validated
suppression measurement “balance in other studies in
in amblyopia technique using point” control23 , 57 and patient
a portable VR technique populations10, 24 , 27
device
2012 Quantitative To propose a 51 normal VR amblyopia N/A Studies on normal subjects Panachakel35
grading of method which subjects; 6 suppression - Proved that differences
suppression grade amblyopic amblyopic test using between the accuracies of
in amblyopia suppression by children asymmetric the left and right eyes can
using VR glasses suppression be used to measure
and dichoptic during asymmetry.
image dichoptic - Equivalence of accuracy of
recognition task. task the 2 eyes are statistically
significant (P = 0.03)
Amblyopic subjects
- Difference between the
mean of accuracies of
normal and amblyopic eye
is significant (P < 0.04)

Table 4 – VR and AR test for choroidal, macular, retinal diseases.

Year proposed main objective subjects investigated Comparison Key findings Reference
application measure technique

19997 , 17 , Examination of To develop and 1 Model eye AR fundus N/A A robust way for Berger
20012 fundus in investigate an AR and 1 microscopy registration, tracking and
choroidal, fundus patient7 ; 5 and overlaying of cowork-
Retinal and biomicroscopy subjects- images ers7 , 8 , 17
macular which overlays intervention8 - Accurate tracking
disease (including photos or with updates at 3–5
angiography) images Hz
onto real-time - Insignificant
slitlamp fundus view overlay inaccuracy,
in biomicroscopy undetectable by
observers
2020 Determination To describe the 1 patient with VR-OCT N/A VR-OCTA aided Gruber
of neo- clinical presentation an acute tiny angiography - Confirmation of the et al15
vascularization and novel floater and (as an retinal origin of the
origin in CSCR anatomical features chronic adjunct neovascularization.
of a patient with bilateral method) - Detailed
chronic central CSCR investigations of
serous neovascular
chorioretinopathy complex and its
(CSCR) complicated interaction with the
by retinal surroundings.
neovascularization
(RNV).

AMD = age-related macular degeneration; AR = augmented reality.

 s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0
gaze fixation.54 In a case study during COVID-19, VR ocular-
kinetic perimetry (OKP) test strategy was remotely installed
and used for at-home visual field monitoring. The strategy
showed good correspondence with HFA 24-2, suggesting that
VR can make perimetry at home or remote visual field moni-
toring possible.14 Another study implemented a VR visual field
test on glaucoma patients. Smartphone and VR headset con-
ducted a staircase algorithm to evaluate the central 24 degree
of visual field. The results showed a high correlation coeffi-
cient between this VR visual field test and the HFA perime-
try, suggesting the test to be potentially useful for visual field
analysis.49
Visually evoked potentials (VEPs), considered as “objective
perimetry”,3 , 4 may be parameters suitable for detection of
early visual field loss. More than a decade ago, Arvind and
coworkers found that early glaucoma diagnosis and glauco-
matous visual field loss detection are possible using their in-
vestigated multifocal VEPs (mfVEPs) under dichoptic VR stim-
ulation.3 , 4 After a proof of concept in the first study,3 they
tested the method on 28 glaucoma patients and 30 healthy
subjects in the subsequent study. The VR goggles presented
black-and-white check stimulus randomly on segments of a
dartboard. While each eye received stimulation at a different
time, the system measured dichoptic mfVEP. Then, the partic-
ipants underwent a monocular mfVEPs test under a similar
stimulus. The dichoptic VR mfVEP method successfully de-
tected perimetric losses on HFA. More importantly, it identi-
fied early defects undetected in HFA and monocular mfVEP,
suggesting that the dichoptic VR method had good sensi-
tivity in detecting early glaucoma.2 In the test, VR goggles
help achieve dichoptic stimulation. This saved time and im-
proved analysis by providing identical testing conditions to
both eyes, thus overcoming the weaknesses of mfVEP monoc-
ular recording.3 , 4 Another study, which is discussed in the
final sub-section, proposed a new VEP sensor for VR head-
set and suggested the sensor as a visual field testing tool as
well.51
3.2. Binocular vision disorders
Binocular vision is important for stereopsis, perception, and
visuomotor function.52 Binocular vision disorders compro-
mise visual function and reduce stereoacuity. The mainstay of
detection and assessment in current practice includes cover
and uncover tests, prisms, and major ambyloscopes.52 Four
studies have investigated the use of VR to screen and diagnose
binocular vision disorders. Two papers validated VR headset-
based cover tests while the rest discussed other ways to eval-
uate binocular function and disorder Table 2. provides a sum-
mary of the quantitative findings and important characteris-
tics of these 4 studies.
Two recent studies have supported the use of VR-based
cover tests in strabismus patients. VR cover tests can over-
come the shortcomings of the conventional cover tests. For
example, with less human judgement and input, VR tests are
more objective and less prone to operative errors. Moreover,
the eye trackers in the VR headsets precisely quantify ocu-
lar movement, including small movements unobservable by
naked eyes. Since users only require a VR headset and a com-
puter to conduct the test and acquire measurements, the VR
1523
cover test is potentially suitable for automated strabismus
assessment. Also, potential synergy between this technique
and telemedicine may break geographical barriers between
doctors and patients.29 , 56 In both studies, researchers pro-
grammed commercially available VR headsets to conduct the
test according to the principles of alternate prism cover test
(APCT) or alternate cover test (ACT). Despite a similar prin-
ciple, the VR-based test utilized programmes and built-in eye
trackers instead of human to present stimuli and record ocular
movement. Both studies validated the accuracy of their own
VR-based APCT, by confirming that the ocular deviation angles
measured by the VR APCT and standard cover tests were well
correlated.29 , 56 In addition, 1 study implemented a VR ACT,
which provided results consistent with an examiner’s cover
test, but was less stable than the VR APCT.29 Given good initial
results and the extensive use of cover tests, we believe these
VR systems can be potentially helpful for diagnosing and as-
sessing strabismus and are thus worthy of investigations with
a larger sample size.
Apart from cover tests, a pilot study showed that a VR dis-
sociative test may diagnoses ocular misalignment in patients
with normal retinal correspondence. Conventionally, ocular
misalignment in normal retinal correspondence patients is
diagnosed by dissociative tests of strabismus, e.g., Lancaster
red-green test, Hees screen. In the VR test, VR headsets pre-
sented visual signals of a dissociative test to patients. At the
same time, examiner could receive the same signals. The test
measured ocular deviation (both vertical and horizontal) and
torsion. In various ocular motility disorders, the VR-based test
and traditional Lees screening test produced similar patterns
of deviation. Moreover, the relatively low cost of VR headsets
means that the VR-based test could become an automated di-
agnostic tool in non-specialist settings.32
Besides VR headsets, a cross-sectional study published in
2009 used large polarized screens and polarized glasses to
conduct a VR-based binocular function test on normal indi-
viduals.16 This VR test used patients’ response to programmed
dichoptic stimuli to measure 3 binocular vision parameters.
For the 2 parameters, the angle of strabismus and motor fu-
sion amplitude (convergence and divergence), the measure-
ment of the VR system was in excellent agreement with that
of major amblyoscope. Although the results of the last param-
eter stereopsis were not provided, the authors claimed the VR
system can present images at different binocular disparities
to offer a simple quantitative test. Comparing with major am-
blyoscope, the VR test can provide a more natural simulation,
evaluate a wider visual field, and provide game-like enjoyable
tests.16
3.3. Amblyopia
Two studies used VR headsets dichoptic tests to quantify am-
blyopia. Although the degree of amblyopia is an important
predictor of treatment efficacy, most traditional clinical tests,
e.g., Worth 4-dot test and Bagolini striated lens tests evaluate
amblyopia qualitatively, but not quantitatively.11 , 35 Therefore,
these VR quantitative tests could become useful Table 3. pro-
vides a summary of the quantitative findings and important
characteristics of these 2 studies.
1524 s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0
The earlier VR test utilized a staircase algorithm to quan-
tify suppression of the input from amblyopic eye.11 During
the test, VR goggles and laptop presented different patterns
to each eye. While “signal dots” for the amblyopic eye had a
unidirectional movement and constant full contrast, “noise
dots” for the control eye moved randomly and had a staircase
increase of contrast from 0% until a balance point at which
neither eye was dominant. This “balance point” was a quanti-
tative measure of suppression. The authors stated that multi-
ple studies have demonstrated the success of this technique
in control23 , 57 and patient populations;10 , 24 , 27 however, since
the pathology and prior treatment affects the degree of sup-
pression, it is difficult to define a “normal” reference range,
but a more severe suppression generally indicates a worse vi-
sual acuity. A limitation is that patients with strabismus and
intense suppression may have difficulty following the test.11
The other recent study used another VR system to quantify
amblyopia.35 In the test taken by healthy subjects after being
shown a target human face, the VR headsets presented differ-
ent faces randomly, and the 2 eyes received different images.
Then, the subjects responded if the target appeared. Since di-
choptic stimulation caused interocular suppression, the ac-
curacy in identifying the target should be compromised. In
healthy subjects, the accuracies of both eyes during dichoptic
stimulation were similar. Six amblyopia patients with differ-
ent underlying causes underwent a similar test and analysis,
the only difference being that a cartoon character was used
instead of human face. The mean difference between the ac-
curacy of the amblyopic eye and that of the normal eye was
significant, and the underlying cause was thought to be the
asymmetric suppression of amblyopic eyes. The sign of the
difference of accuracy between the 2 eyes indicates the am-
blyopic eye while the magnitude of that difference measures
the extent of asymmetry. Therefore, this tool may be able to
grade amblyopia accurately and predict therapy response.35
3.4. Choroidal or retinal diseases
Four publications used concepts of VR or AR to diagnose
choroidal and retinal diseases. Three articles were published
by the same team on an AR fundus biomicroscopy. The other
paper used VR to diagnose neovascularization under chori-
oretinopathy Table 4. presents the quantitative findings and
important characteristics of these 4 articles.
In 3 papers published about 2 decades ago, Berger and
coworkers proposed an AR slit lamp fundus biomicroscopy
system to examine the fundus and diagnose relevant patholo-
gies.7 , 8 , 17 The final system, which comprises computers, slit
lamp biomicroscope, camera, and other hardware, applied the
AR concept by overlaying previously captured images (includ-
ing custom photos and angiograms) onto the slit lamp fundus
view of the biomicroscope in real time. A preliminary study
and a later case series suggested that their system can reg-
ister, track, and overlay images robustly and accurately.7 , 8 , 17
The AR biomicroscopy used 1 eyepiece to display the original
fundus and the other to display the fundus with overlaid im-
ages. Without referring to a separate display apparatus, doc-
tors can compare and analyze images and fundi. Moreover, the
real-time tracking feature is even more important when the
inspected eyes are moving. Apart from assisting the doctor in
front of the patient, the system can share the local eyepiece
view to a remote collaborator.7 , 8 , 17
In 2020, a case report used VR OCT angiography (OCTA) to
investigate and determine the origin of a neovascular com-
plex. The patient had a history of chronic bilateral central
serous chorioretinopathy and developed neovascularization.
VR OCTA, 3D-OCTA, and a new rendering method provided
a 3D reconstruction VR, which was in turn used in conjunc-
tion with swept source OCTA. This combination investigation
method successfully determined the retinal origin of the neo-
vascularization.15
3.5. Other diseases or conditions
Several papers discussed the applications of VR and AR on di-
agnosing or screening other conditions, e.g., eye injury, age-
related macular degeneration. This section also discusses 2
studies that focused on hardware testing.26 , 51 Table 5 summa-
rized the quantitative findings and important characteristics
of articles discussed in this section.
Three articles have discussed the use of VR and AR to mea-
sure visual function parameters. A recent pilot study devel-
oped and evaluated an optokinetic nystagmus (OKN)-based
VR tool for detecting contrast sensitivity. The tool utilized VR
head-mounted displays to show a virtual rotating OKN drum
and an algorithm to detect OKN occurrence. From the OKN oc-
currence, contrast sensitivity was inferred. When this VR tech-
nique was implemented on healthy eyes and eyes under low
vision simulation, some correct trends were shown; however,
the test could only estimate the contrast sensitivity well in
the low vision simulation eyes, and the results from VR test
were inconsistent with that from the manual test. The VR test
was unsuccessful in healthy eyes due to inadequate contrast
resolution of VR headsets. Despite the results, due to its ease
of use, short test duration, and automated analysis, the pro-
posed system might promote the use of contrast sensitivity
testing in clinical practice or even become a household screen-
ing tool.45
Other than contrast sensitivity, current evidence has
shown good results on the use of VR and AR to measure dy-
namic visual acuity, but not on static visual acuity.21 , 26 For dy-
namic visual acuity, a cross sectional study utilized VR and
AR head-mounted display to implement the King-Devick (KD)
test and evaluated these VR and AR KD tests in 30 healthy sub-
jects by comparing these VR and AR KD tests with hard copy
KD tests. The results showed that the AR KD test correlated
with the hard copy KD well, but took a longer time. Compared
to the AR KD test, VR KD had a worse correlation with hard
copy KD and took an even longer time. The larger discrepancy
between VR and hard copy test was proposed to be stemming
from the fact that VR was further from the actual reality. In
terms of user experience, the VR and AR tests caused slightly
more subjective symptoms (e.g., dizziness, ocular fatigue).21
For the purpose of assessing static visual acuity, a recent hard-
ware validation research found VR head-mounted devices in
the market technically unsuitable, because the minimum an-
gular resolution of these devices could not reach the resolu-
tion in normal acuity (log MAR = 0.0, equivalent to 20 of 20).26
The same hardware validation study also assessed limita-
tions of those VR devices for their potential to be applied in
 s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0 1525

Table 5 – Other VR and AR tests.

Year Proposed Main objective Subjects Investigated Comparison Key findings Reference
application measure technique

2020 Measurement To develop a fast 10 healthy OKN-based contrast Accuracy of OKN Tatiyosyan
or at-home OKN-based virtual subjects, then VR contrast sensitivity prediction: 88% et al45
screening diagnosis tool to repetition measure- assessment Reproduction of correct
tool of estimate contrast under low ment based on trends, including difference
Contrast sensitivity vision tool Manual in contrast sensitivity
sensitivity automatically simulation report between healthy and low
without active vision conditions, and
cooperation of the lower contrast sensitivity at
patient and the lower special frequency
practitioner Successful estimation of
contrast sensitivity mainly
in low vision simulation
2019 Measurement To evaluate the 30 human KD test Hard copy KD Subjective symptom: all Kim et al21
of dynamic effectiveness of a VR subjects implemented test were mild, but compared to
visual acuity KD test and AR KD without eye by a VR head- hard copy test, there was
test to measure or mental mounted significantly more dizziness
dynamic visual disease display and and ocular fatigue in VR
acuity an AR head- and AR test respectively.
mounted Correlation analysis:
display - AR KD correlated with
hard copy KD in all 3
components of KD test.
- VR KD correlated with
hard copy KD only in 1 out
of 3 components
2020 Evaluation of To determine 2 standalone N/A N/A Visual acuity: Minimum Lynn et al26
visual acuity, feasibility of VR VR devices, 4 angular resolution (in
convergence headsets for visual smartphone logMAR) or smallest
ability, and care applications by VR headsets displayable letter ranged
visual field measuring the from 0.28–0.79.
physical or optical Convergence ability:
characteristics of the binocular convergence
headset, including demand was adequate for 4
angular resolution headsets.
and field of view. Visual field: although a
manufacturers’ reports
overestimate
field of view, the real field of
view is sufficient for some
visual field tests (e.g.,
conducting HFA 24–2, 30–2
by moving fixation point)
2019 Diagnosis of To assess the Case studies A visual N/A Amblyopia test: amblyopia Versek
amblyopia, applicability of a (Human evoked eye had a smaller P100 et al51
AMD, and visual evoked subjects) potentials amplitude and contrast
early potentials and fields and fields sensitivity
glaucoma sensor in vision and sensor, with AMD-related test: delays in
neurological smartphone dark adaption time
disorders VR that hypothesized to be useful.
conduct Visual field test in
condition- glaucoma patient: diseased
specific eye showed higher contrast
tests sensitivity, and suppressed
P100 peak amplitude and
high later peak in transient
Visual Evoked Potentials

(continued on next page)

1526 s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0

Table 5 (continued)

Year Proposed Main objective Subjects Investigated Comparison Key findings Reference
application measure technique

2021 Diagnosis of To propose the 89 patients VR-based Traditional Superior performance in Li25
anterior ultrasound with closed ultrasound B-ultrasound ultrasound biomicroscopy
chamber biomicroscope eye trauma biomicro- bioassays and better
Trauma diagnosis and (56- scope diagnostic efficiency.
analysis of traumatic intervention;
anterior chamber 33- control
injury based on VR.
2019 Measurement To use augmented 97 normal Augmented N/A Binocular imbalance was Xu et al55
of binocular VR platform to subjects VR platform found in 65.9% and 62.89%
imbalance assess the presence (stimuli with of participants under by
of binocular variable low and high temporal
imbalance in spatial and frequency respectively.
individuals with temporal Severe binocular imbalance
good ocular health. frequency) was found in 16.5% and
7.21% of subjects under low
and high temporal
frequency stimulation
respectively.

AMD = age-related macular degeneration; AR = augmented reality; KD = King Devick; OKN = Optokinetic nystagmus; VR = virtual reality.

visual field and convergence ability tests, suggesting favorable
results. For visual field testing, after moving fixation point, the
VR headsets could provide a sufficient field of view for tests in-
cluding threshold field tests HFA 24-2 and 30-2. For near con-
vergence testing, to challenge the fusional convergence break
point, VR headsets need to generate a convergence demand.
The results showed that most VR headsets fulfilled that re-
quirement.26
A hardware study suggested that a portable sensor sys-
tem as a companion to VR headsets can improve the diag-
nosis and clinical examination of glaucoma, age-related mac-
ular degeneration (AMD), and amblyopia. This system mea-
sured electric field encephalography and conventional elec-
troencephalography. The authors claimed that it may be use-
ful in diagnosing glaucoma, amblyopia, and age-related mac-
ular degeneration. For each disease, they proposed custom
VR stimuli to provoke visual evoked potentials and other re-
sponses. In amblyopia, using the dichoptic ability of the sen-
sor to record the response of each eye, the pathological eyes
showed abnormality of contrast sensitivity and transient vi-
sual evoked potentials (tVEP). In glaucoma, eyes showed an-
other form of abnormality in tVEP. These pathological pat-
terns can be useful in diagnosis. For AMD the transient visual
evoked potentials and fields during recovery from dark adap-
tation was measured to evaluate the recovery response, which
was expected to be delayed in AMD. Several breakthroughs
were claimed, including its portability, usability in many po-
sitions and settings, convenience, synergy with telemedicine,
and quantitative measurement. Despite the breakthroughs,
the system has several limitations, including signal errors,
interference due to eye closure and movement, and training
effort.51
A recent study used VR-technology-based ultrasound
biomicroscope to diagnose anterior chamber injury. Eighty-
nine patients with closed trauma were recruited and ran-
domly assigned to VR-technology-based ultrasound biomicro-
scope or traditional B-scan ultrasound for diagnosis. Through
statistical analysis of the ultrasound biomicroscopy biomet-
ric values, the VR-technology-based ultrasound biomicro-
scope performed better than traditional B-scan ultrasound.25
This study was not explicit in describing and explaining
the mechanism of VR implementation in the biomicro-
scope, possibly because of its emphasis on patient care and
outcome.
Another study used an augmented VR platform to detect
binocular imbalance in normal subjects. In the test, stimuli
were presented under different spatial and temporal frequen-
cies. Based on psychophysical feedback, any induced binocu-
lar imbalance was evaluated. When the visual stimuli have a
lower spatial or temporal frequency, the detection of binocu-
lar imbalance is more sensitive. Although the authors believed
that imbalance is physiological, transient, and generally not
causing amblyopia, they argued that their findings may help
the diagnosis and analysis of binocular abnormalities.55 As
this information was only verified on normal subjects, further
validation on patients is required.
4. Discussion
Although the American Academy of Ophthalmology has advo-
cated eye disease screening at age 40, underdiagnosis around
the world is still a serious concern. For example, in 2020 an
estimated 43.78 million primary open-angle glaucoma cases
were undetected.42 In countries with high human develop-
ment index, more than 70% of cases were undetected; in coun-
tries with low human development index, the undetected pro-
portion is nearly 95%.42 The low detection ratio is likely due to
 s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0
a lack of widely accepted and cost-effective screening strategy,
and a limited and uneven access to eye care services.42 In the
face of the underdiagnosis issues, as well as the diagnostic er-
ror and ophthalmologist shortage discussed in the introduc-
tion, this work provides a comprehensive update on current
VR and AR test. The following discussion will suggest why they
may be able to improve and complement current diagnosis or
screening services.
These VR and AR tests may have advantages over tradi-
tional tests in the following 5 aspects. First, VR and AR tests
could be standardized and objective because they can use ob-
jective machinery measurement methods, and their proce-
dures are typically programmed and standardized. For exam-
ple, in the VR cover test, eye trackers measured eye move-
ment accurately and subjectively, and programmed approach
can avoid inter-examiner discrepancy (e.g., position of patient)
and provide a standard background.29 Second, in VR and AR
tests, some measurement techniques are less reliant on pa-
tient’s response or cooperation. For instance, the VR OKN-
based contrast sensitivity test collected data by an embed-
ded eye tracker, so the test required minimal patient cooper-
ation.45 A VEPs and fields sensor for VR, which required min-
imal patient cooperation, could reliably diagnose unrespon-
sive or malingering patients.51 This may also be advantageous
in amblyopia diagnosis or measurement, because while chil-
dren often cannot cooperate and thus make clinical examina-
tion challenging, early diagnosis and treatment are essential
to modify the disease course.
Third, VR and AR paves the way for automation. This
can be done by programming VR and AR screen to present
stimuli, and devices like eye trackers and even electrodes
to measure responses like eye movement and visual evoked
potentials. Therefore, measurement can be taken and re-
ported automatically, saving time, and allowing ophthalmol-
ogists to see more patients. Automation can encourage non-
ophthalmologist settings to deploy these VR and AR tests, thus
promoting detection of undiagnosed cases and earlier refer-
ral to specialist. Fourthly, compared to other digital methods
(e.g., LCD screen), VR and AR simulate reality better. For ex-
ample, in a study on a binocular function VR test, VR con-
cept provided a visual experience that is more realistic than
phase-difference haploscope.16 Finally, but importantly, pa-
tients seem to prefer some VR and AR tests over traditional
tests. Several studies have stressed that their VR/AR tests
can present engaging target stimuli29 , 35 , or even game-like
tests.11 , 16 Some survey on patients’ experience and preference
regarding VR tests yielded favorable result, e.g., participants
preferred a VR perimeter and found it more usable compared
to the HFA.28
Apart from strengths of VR and AR tests, these tests can
expand the coverage of current clinical service due to its
lower resource requirement and synergy with telemedicine.
Being compact, VR and AR apparatus can save space, therefore
potentially making more test stations available. Automation
means less manpower requirement, and with adoption of af-
fordable headsets, costs of operations may be further reduced.
Moreover, the digital nature of VR and AR test makes them
convenient for telemedicine, helping health care providers
to serve rural areas and facilitating at-home testing. Exam-
ples include a VR cover test,56 and an at-home glaucoma
1527
screening method.50 The lower resource use and synergy with
telemedicine mean that even if some VR or AR tests are less
accurate than traditional tests, they may be useful. One ap-
plication is improvement of screening to tackle the under-
diagnosis issue. Considering the Wilson and Jungner’s princi-
ples of screening,53 the advantages of VR and AR tests (e.g., low
resource use, portable set-up, acceptance by patients) granted
VR and AR a potential to become cost-effective screening al-
gorithms. Another potential application is remote examina-
tion and consultation, which provides convenience to patients
and improves follow-up compliances. Under COVID-19, in-
fection prevention and pressure on the health care system
disrupted clinical services, and at-home monitoring, e.g., at-
home VR visual field perimetry,14 may become important. ad-
vantages and potential applications, many barriers still ex-
ist between VR and AR technology development and their
application in clinical settings for diagnosis or screening of
eye diseases. First and foremost, the accuracy and validation
of such tools need to be improved. From Tables 1–5, VR and
AR test are often inferior in accuracy compared to standard
tests, and most studies had small sample size. Studies aim-
ing for the same purpose (e.g., visual field testing, cover test)
involved different VR systems and algorithms, making it dif-
ficult to conduct meta-analysis and draw a reliable conclu-
sion. While tests with good preliminary results needs more
validations, for less accurate tests, investigators should sug-
gest improvements or possible applications with evidence.
Second, instances of VR and AR fatigue and cybersickness
may deter some patients from using the tests. Designers of
VR tests shall consider and minimize factors that increase
VR sickness e.g., gaming content, younger age, long exposure
time.41
Third, current hardware is often inadequate to conduct VR
or AR tests. A hardware testing study suggested their deficits
and limitations and problems of current specifications con-
ventions, e.g., insufficient resolution for visual acuity testing,
and incorrect specifications of field of view.26 Hardware error
sources are often considered 1 of the future improvement di-
rections for VR or AR tests. For instance, in a study on a VR
contrast sensitivity estimation test, pixel resolution and eye-
tracking accuracy limited the accuracy, so a hardware devel-
opment phase preceding clinical validation was advocated.45
With the large potential market of VR and AR in health care,A
and the underserved areas and manpower deficit in ophthal-
mology, it is feasible to design VR and AR hardware specifi-
cally for ophthalmology. In fact, regarding this matter, some
companies have already launched products like VR perime-
ters. Fourth, commercial involvement in validation may raise
doubts on the subjectivity and reliability of results. Despite
the necessity of commercial involvement in the development
phase, third-party and academic validation would bring more
confidence to health practitioners and patients. Fifth, the
use of vague or poorly-defined terms in studies increases
difficulty in interpretation and comparison, thus impeding
potential users from deploying the VR or AR tests. There-
fore, in the future, terms shall be either standard or clearly
defined.
Most of the studies discussed implemented VR tests in-
stead of AR tests. This phenomenon exists possibly because
a standardized environment, which only VR can provide, is
1528 s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0
required for diagnostic tests to minimize distractions or in-
terferences from the real world. Moreover, some investiga-
tors replicated standard test with a very standardized envi-
ronment, e.g., HFA, and in visually evoked potential measure-
ment, so they may have preferred VR to have a closer repli-
cation. AR also has several potential advantages. AR is closer
to the real world and that could bring about translational ad-
vantages. In a study about VR and AR King-Devick test, com-
pared to VR King-Devick test, the measurement of AR King-
Devick test was close to that of real King-Devick test, poten-
tially because the AR environment was closer to the reality.21
Another possible advantage of AR lies on user experience. In
AR, patients may be less prone to subjective symptoms than
VR systems.B Besides, AR devices are generally more compact
and affordable than VR devices. As VR and AR both have some
strengths over each other, mixed reality, which incorporates
the best functions of VR and AR, and displays both the real
world and virtual objects simultaneously,13 may be a future
direction.
5. Conclusions
We systematically introduced and discussed how VR and AR
are used as a diagnostic and screening tool in ophthalmology.
We found 26 relevant articles that implemented a variety of
tests, such as VR cover tests and VR perimetry, showing the
wide potential clinical application of VR and AR. While VR and
AR tools can become standardized, automated, cost-effective
tools with good user experience, these tools were often less
accurate, weakly validated, and affected by hardware limita-
tions. Finally, we suggested that, while most studies used VR
instead of AR because VR provides a more controlled and stan-
dardized environment, AR may be closer to reality and pro-
duce less subjective symptoms. Therefore, both VR and AR
may contribute to translational research in the future.
Author contributions
The first author named is lead author and the last author
named is corresponding author. All other authors are listed
according to their amount of contribution. We describe contri-
butions of the authors to the paper using CRediT taxonomy12 :
Conceptualization: Y.K.C.∗ , M.K.I.M., C.S., S.H.L.P., R.S.W.Y.,
K.C.S.; Methodology: Y.K.C., M.K.I.M., C.S. S.H.L.P., R.S.W.Y.,
K.C.S.; Investigation: M.K.I.M.∗ , C.S.∗ , Y.K.C., S.H.L.P., R.S.W.Y.,
K.C.S.; Writing- Original draft: M.K.I.M.; Writing- Review&
Editing: equal contribution among all authors; Supervision:
Y.K.C.; Project Administration: Y.K.C., M.K.I.M.
An asterisk is used to denote a ‘lead’ degree of contribution
in a specific role. Otherwise, it can be assumed that all authors
listed contribute to that role equally. Unlisted roles are not ap-
plicable in our study.
Disclosure
Acknowledgement of Grant: Teaching and Development Grant
(Project number: 823), the University of Hong Kong.
Conflict of interest
The authors have no conflict of interest to disclose.
Method of literature search
The screening process is based on a list with 766 records gen-
erated by the systematic and final search on PubMed on 16th
July, 2021. The keywords “((Virtual reality) OR (Augmented re-
ality) OR (Mixed reality)) AND ((ophthalmology) OR (eye) OR
(ophthalmic) OR (ocular) OR (orbit)) AND ((Diagnosis) OR (Diag-
nosing) OR (Diagnostic) OR (Detection) OR (Detect) OR (Screen-
ing) OR (screen) OR (Examination) OR (examine) OR (Sever-
ity) OR (assessment) OR (assess) OR (test))” and filter “English”
were applied. The authors imposed no limit on publication
date.
The same inclusion and exclusion criteria have been used
consistently throughout the process. Articles included shall
involve original findings regarding the use of VR or AR in di-
agnosis and screening in ophthalmology, discuss application
on any ophthalmic diseases or conditions, and be written in
English. Articles without original findings regarding our topic
of interest, or studies purely discussing disease disability and
rehabilitation would be excluded.
Key references
The highlighted references in this review are as follows. Dis-
cussed in the introduction, a review discussed how vari-
ous new digital technologies, such as artificial intelligence,
telecommunication, can improve ophthalmology service in
the past and in the future.22 Other key references come
from the included references in the result section. For glau-
coma, a study on VR perimetry had a relatively large sam-
ple size of over 150 people, and did a comprehensive study
on both accuracy and patient’s experience.28 A study on an-
other VR perimetry suggested that the diagnostic perfor-
mance of the investigated VR method was at least simi-
lar to that of Humphrey Field Analyzer.39 For binocular vi-
sion disorders, a research group validated VR cover test with
a moderate sample size, showing decent initial results.56
For amblyopia, despite a small sample size, a study sug-
gested that VR methods may be useful quantifying amblyopia
suppression.35
Other cited materials
A. BCC Publishing Staff. Virtual and Augmented Real-
ity: Technologies and Global Markets., BCC Research,
2018.
B. Hughes C, Fidopiastis CM, Stanney KM, et al. The Psy-
chometrics of Cybersickness in Augmented Reality: Frontiers
in Virtual Reality, 2020
 s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0
references
1. Ahn JM, Kim S, Ahn K-S, et al. Accuracy of machine learning
for differentiation between optic neuropathies and
pseudopapilledema. BMC Ophthalmol. 2019;19:178.
2. Anitha SP, Prabha Susy M. Impact of virtual reality in
healthcare: a review.Giuliana G, editor. Virtual and
augmented reality in mental health treatment. Hershey, PA,
USA: IGI Global; 2019. p. 17–31.
3. Arvind H, Klistorner A, Graham SL, Grigg JR. Multifocal visual
evoked responses to dichoptic stimulation using virtual
reality goggles: multifocal VER to dichoptic stimulation. Doc
Ophthalmol. 2006;112:189–99.
4. Arvind H, Klistorner A, Graham S, et al. Dichoptic stimulation
improves detection of glaucoma with multifocal visual
evoked potentials. Invest Ophthalmol Vis Sci. 2007;48:4590–6.
5. Aydındoğan G, Kavaklı K, Şahin A, et al. Applications of
augmented reality in ophthalmology [Invited]. Biomed Opt
Express. 2020;12:511–38.
6. Bakshi SK, Lin SR, Ting DSW, et al. The era of artificial
intelligence and virtual reality: transforming surgical
education in ophthalmology. Br J Ophthalmol. 2020.
7. Berger JW, Shin DS. Computer-vision-enabled augmented
reality fundus biomicroscopy. Ophthalmology.
1999;106:1935–41.
8. Berger JW, Madjarov B. Augmented reality fundus
biomicroscopy: a working clinical prototype. Arch
Ophthalmol. 2001;119:1815–18.
9. Biousse V, Newman NJ, Najjar RP, et al. Optic disc
classification by deep learning versus expert
neuro-ophthalmologists. Ann Neurol. 2020;88:785–95.
10. Black JM, Thompson B, Maehara G, Hess RF. A compact
clinical instrument for quantifying suppression. Optom Vis
Sci. 2011;88:E334–43.
11. Black JM, Hess RF, Cooperstock JR, et al. The measurement
and treatment of suppression in amblyopia. J Vis Exp.
2012(70):e3927.
12. Brand A, Allen L, Altman M, et al. Beyond authorship:
attribution, contribution, collaboration, and credit. Learn
Publ. 2015;28:151–5.
13. Brigham TJ. Reality check: basics of augmented, virtual, and
mixed reality. Med Ref Serv Q. 2017;36:171–8.
14. Deiner MS, Damato BE, Ou Y. Implementing and monitoring
at-home virtual reality oculo-kinetic perimetry during
COVID-19. Ophthalmology. 2020;127:1258.
15. Gruber M, Wolf J, Stahl A, et al. Novel insights into retinal
neovascularization secondary to central serous
chorioretinopathy using 3D optical coherence tomography
angiography. Am J Ophthalmol Case Rep. 2020;18:100609.
16. Handa T, Ishikawa H, Shimizu K, et al. A novel apparatus for
testing binocular function using the ’CyberDome’
three-dimensional hemispherical visual display system. Eye
(Lond). 2009;23:2094–8.
17. Hariprasad R, Shin DS, Berger JW. An intelligent, interactive
platform for ophthalmic teaching, telemedicine, and
telecollaboration: design considerations and prototype
construction. Stud Health Technol Inform. 1999;62:124–9.
18. Hussain A, Oestreicher J. Clinical decision-making: heuristics
and cognitive biases for the ophthalmologist. Surv
Ophthalmol. 2018;63:119–24.
19. Iskander M, Ogunsola T, Ramachandran R, et al. Virtual reality
and augmented reality in ophthalmology: a contemporary
prospective. Asia Pac J Ophthalmol (Phila). 2021;10:244–52.
20. Khalifa YM, Bogorad D, Gibson V, et al. Virtual reality in
ophthalmology training. Surv Ophthalmol. 2006;51:259–73.
21. Kim J-H, Son H-J, Lee S-H, Kwon S-C. VR/AR head-mounted
1529
display system-based measurement and evaluation of
dynamic visual acuity. J Eye Mov Res. 2019;12:1.
22. Li J-PO, Liu H, Ting DSJ, et al. Digital technology, tele-medicine
and artificial intelligence in ophthalmology: a global
perspective. Prog Retin Eye Res. 2021;82:100900.
23. Li J, Lam CS, Yu M, et al. Quantifying sensory eye dominance
in the normal visual system: a new technique and insights
into variation across traditional tests. Invest Ophthalmol Vis
Sci. 2010;51:6875–81.
24. Li J, Thompson B, Lam CS, et al. The role of suppression in
amblyopia. Invest Ophthalmol Vis Sci. 2011;52:4169–76.
25. Li X. Ultrasound biomicroscopy diagnosis analysis and fine
care of anterior segment injury of traumatic anterior
chamber based on intelligent virtual reality technology. J
Healthcare Eng. 2021;2021:9930686.
26. Lynn MH, Luo G, Tomasi M, et al. Measuring virtual reality
headset resolution and field of view: implications for vision
care applications. Optom Vis Sci. 2020;97:573–82.
27. Mansouri B, Thompson B, Hess RF. Measurement of
suprathreshold binocular interactions in amblyopia. Vision
Res. 2008;48:2775–84.
28. Mees L, Upadhyaya S, Kumar P, et al. Validation of a
head-mounted virtual reality visual field screening device. J
Glaucoma. 2020;29:86–91.
29. Miao Y, Jeon JY, Park G, et al. Virtual reality-based
measurement of ocular deviation in strabismus. Comput
Meth Prog Biomed. 2020;185:105132.
30. Milea D, Najjar RP, Jiang Z, et al. Artificial intelligence to
detect papilledema from ocular fundus photographs. N Engl J
Med. 2020;382:1687–95.
31. Montelongo M, Gonzalez A, Morgenstern F, et al. A virtual
reality-based automated perimeter, device, and pilot study.
Transl Vis Sci Technol. 2021;10:20.
32. Nesaratnam N, Thomas P, Vivian A. Stepping into the virtual
unknown: feasibility study of a virtual reality-based test of
ocular misalignment. Eye (Lond). 2017;31:1503–6.
33. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020
statement: an updated guideline for reporting systematic
reviews. BMJ. 2021;372:n71.
34. Page MJ, Moher D, Bossuyt PM, et al. PRISMA 2020 explanation
and elaboration: updated guidance and exemplars for
reporting systematic reviews. BMJ. 2021;372:n160.
35. Panachakel JT, Ramakrishnan AG, Manjunath KP. VR glasses
based measurement of responses to dichoptic stimuli: a
potential tool for quantifying amblyopia? Ann Int Conf IEEE
Eng Med Biol Soc. 2020;2020:5106–10.
36. Patrik P, Bostjan Z. Augmented Reality in Healthcare.Maria
Manuela C-C, Isabel Maria M, Ricardo M, Rui R (eds.).
Encyclopedia of e-health and telemedicine. Hershey, PA, USA:
IGI Global; 2016. p. 423–38.
37. Plummer DJ, Lopez A, Azen SP, et al. Correlation between
static automated and scanning laser entoptic perimetry in
normal subjects and glaucoma patients. Ophthalmology.
2000;107:1693–701.
38. Rachapelle S, Legood R, Alavi Y, et al. The cost-utility of
telemedicine to screen for diabetic retinopathy in India.
Ophthalmology. 2013;120:566–73.
39. Razeghinejad R, Gonzalez-Garcia A, Myers JS, Katz LJ.
Preliminary report on a novel virtual reality perimeter
compared with standard automated perimetry. J Glaucoma.
2021;30:17–23.
40. Resnikoff S, Felch W, Gauthier TM, Spivey B. The number of
ophthalmologists in practice and training worldwide: a
growing gap despite more than 200,000 practitioners. Br J
Ophthalmol. 2012;96:783–7.
41. Saredakis D, Szpak A, Birckhead B, et al. Factors associated
with virtual reality sickness in head-mounted displays: a
1530 s u rv e y o f o p h t h a l m o l o g y 6 7 ( 2 0 2 2 ) 1 5 1 6 – 1 5 3 0
systematic review and meta-analysis. Front Hum Neurosci.
2020;14:96.
42. Soh Z, Yu M, Betzler BK, et al. The global extent of undetected
glaucoma in adults: a systematic review and meta-analysis.
Ophthalmology. 2021;128:1393–404.
43. Sommer A, Taylor HR, Ravilla TD, et al. Challenges of
ophthalmic care in the developing world. JAMA Ophthalmol.
2014;132:640–4.
44. Stunkel L, Newman NJ. Biousse V. Diagnostic error and
neuro-ophthalmology. Curr Opin Neurol. 2019;32:62–7.
45. Tatiyosyan SA, Rifai K, Wahl S. Standalone cooperation-free
OKN-based low vision contrast sensitivity estimation in VR -
a pilot study. Restor Neurol Neurosci. 2020;38:119–29.
46. Taylor RH, Levin A. Key topics in ophthalmology. Oxford: CRC
Press; 2001.
47. Thompson AC, Parikh PD, Lad EM. Review of ophthalmology
medical professional liability claims in the united states from
2006 through 2015. Ophthalmology. 2018;125(5):631–41.
48. Thomsen AS, Subhi Y, Kiilgaard JF, et al. Update on
simulation-based surgical training and assessment in
ophthalmology: a systematic review. Ophthalmology.
2015;122:1111–30 e1.
49. Tsapakis S, Papaconstantinou D, Diagourtas A, et al. Visual
field examination method using virtual reality glasses
compared with the Humphrey perimeter. Clin Ophthalmol.
2017;11:1431–43.
50. Tsapakis S, Papaconstantinou D, Diagourtas A,
et al. Home-based visual field test for glaucoma screening
comparison with Humphrey perimeter. Clin Ophthalmol.
2018;12:2597–606.
51. Versek C, Rissmiller A, Tran A, et al. Portable system for
neuro-optical diagnostics using virtual reality display. Mil
Med. 2019;184:584–92.
52. Von Noorden GK, Campos EC. Binocular vision and ocular
motility: theory and management of strabismus. St. Louis:
Mo., Mosby; 2002.
53. Wilson JMG, Jungner GWorld Health O. Principles and practice
of screening for disease. Geneva: World Health Organization;
1968.
54. Wroblewski D, Francis BA, Sadun A, et al. Testing of visual
field with virtual reality goggles in manual and visual grasp
modes. Biomed Res Int. 2014;2014:206082.
55. Xu L, Huang M, Lan J, et al. Assessment of binocular
imbalance with an augmented virtual reality platform in a
normal population. Cyberpsychol Behav Soc Netw.
2019;22:127–31.
56. Yeh PH, Liu CH, Sun MH, et al. To measure the amount of
ocular deviation in strabismus patients with an eye-tracking
virtual reality headset. BMC Ophthalmol. 2021;21:246.
57. Zhang P, Bobier W, Thompson B, Hess RF. Binocular balance
in normal vision and its modulation by mean luminance.
Optom Vis Sci. 2011;88:1072–9.