18_08_14

1. CORRELATION OF FUNDUS
AUTOFLUORESCENCE WITH FOVEAL
MICROSTRUCTURES AND VISION IN
BRANCH RETINAL VEIN OCCLUSION
BYEONGJUN PARK, MD, JAEYOUNG KIM, MD, HYEWON CHUNG, MD, PHD,
HYUNG CHAN KIM, MD, PHD
Purpose: To investigate the correlation of fundus autofluorescence (FAF) with the
findings of spectral domain optical coherence tomography and visual acuity in patients with
branch retinal vein occlusion and to determine the visual prognostic factors.
Methods: Retrospectively, an evaluation of FAF, spectral domain optical coherence
tomography images, and visual acuity before and after intravitreal injection of bevacizumab
(pre- and post-IVB) was obtained in 42 patients with branch retinal vein occlusion who
underwent IVB as their first treatment. The FAF of fovea was graded on a scale of 1 to 4.
Results: The visual acuity post-IVB was associated with the visual acuity pre-IVB.
Preservation of external limiting membrane and photoreceptor inner and outer segment
junction pre- and post-IVB were associated with better visual acuity post-IVB. Eyes with
less FAF pre-IVB were closely associated with better visual acuity post-IVB.
Conclusion: The shorter length of photoreceptor inner and outer segment junction and
external limiting membrane defect and less FAF pre-IVB showed a significant association
with better visual acuity post-IVB. These associations could help to predict potential
restoration of photoreceptor integrity and visual recovery in patients with branch retinal vein
occlusion, in whom photoreceptor integrity before treatment could not be adequately
evaluated, even with spectral domain optical coherence tomography.
RETINA 34:531–538, 2014
Retinal vein occlusion is the most common retinal
vascular disease after diabetic retinopathy.1
Loss
of visual function because of branch retinal vein occlu-
sion (BRVO) is mainly caused by macular edema
(ME).1
Current treatments for ME include grid laser
treatment,2
intravitreal injection of steroids,3
surgical
procedures,4
and intravitreal antivascular endothelial
growth factor (VEGF) agents.5,6
There is increasing
evidence that intravitreal injection of bevacizumab
(IVB) can lead to significant reduction of ME. How-
ever, bevacizumab therapy is not effective in some
patients, despite multiple intravitreal injections. It is
still unclear whether there are prognostic factors that
can predict the success of bevacizumab therapy.
Compared with conventional time domain optical
coherence tomography (TD OCT), spectral domain
optical coherence tomography (SD OCT) yields a high
degree of axial resolution and provides detailed views
of the intraretinal microstructures, including the integ-
rity of the photoreceptor inner and outer segment
junction (IS/OS) and the external limiting membrane
(ELM).7,8
Recent SD OCT studies have proposed that
the integrity of the foveal photoreceptor layer is sig-
nificantly associated with visual function. The restora-
tion of the structure of the foveal photoreceptor is
needed to achieve good visual recovery. In general,
the presence of a visible IS/OS and ELM is believed
to be indicative of the integrity of the foveal photore-
ceptor layer, which is a visual prognostic factor in
patients of BRVO.9
However, the detailed microstruc-
tures of the photoreceptor layer in patients with severe
From the Department of Ophthalmology, Konkuk University Med-
ical Center, Konkuk University School of Medicine, Seoul, Korea.
Supported by Basic Science Research Program through the
National Research Foundation of Korea (NRF) funded by the Min-
istry of Education, Science and Technology (2012R1A1A11012171).
None of the authors have any conflicting interests to disclose.
Reprint requests: Hyung Chan Kim, MD, PhD, Department of
Ophthalmology, Konkuk University Medical Center, Konkuk Uni-
versity School of Medicine, 120-1 Neungdon-ro, Gwangjin-gu, Seoul
143-729, Korea; e-mail: eyekim@kuh.ac.kr
531

2. ME are difficult to visualize clearly, even using SD
OCT.
Fundus autofluorescence (FAF) is important in
diagnosing various diseases, such as age-related
macular degeneration or retinal dystrophy.10–12
How-
ever, interpretation of the FAF images is not fully
understood yet. Fundus autofluorescence derives from
lipofuscin in the retinal pigment epithelium (RPE),
which is a mixture of fluorophores that represent the
digested residues of the retinal outer segments. Since
the lipofuscin is a by-product of phagocytosis of pho-
toreceptor by RPE, FAF is useful to describe the func-
tion and morphology of the retina and the RPE.
Vujosevic et al13
proposed that an increased level of
FAF was associated with a functional and structural
impairment of the macula in patients with diabetic
ME. They reported a significant association between
the presence of areas with increased FAF and
decreased retinal sensitivity. The increased level of
FAF was expected to decrease or disappear after the
reduction or resolution of ME with treatment. How-
ever, we found that not all patients with ME had
increased FAF. In contrast, some patients with severe
ME had relatively low FAF, and some patients were
found to have persistently increased FAF after the
resolution of ME in patients with diabetes mellitus.14
We studied the association between various foveal
microstructures, which are thought to be hallmarks of
the integrity of the foveal photoreceptor layer, and
logarithm of the minimum angle of resolution (log-
MAR) visual acuity to better characterize and predict
the visual outcome in patients with BRVO.
Methods
From May 1, 2010 to October 31, 2011, 81 patients
underwent IVB for the treatment of BRVO associated
with ME at the Department of Ophthalmology,
Konkuk University Medical Center. We performed
retrospective chart review of 42 of these patients for
this study. Twenty-three patients were excluded from
the analysis because of history of grid laser, intra-
vitreal steroid injection, pars plana vitrectomy, and
massive preretinal hemorrhage. Sixteen patients were
excluded because of the follow-up loss.
All patients underwent 3 consecutive monthly IVB
(1.25 mg per 0.05 mL). Each first injection was made
within 1 week of the initial visit for the treatment of
ME. We did not encounter any serious side effects
related to IVB. All patients underwent a complete
ophthalmic examination including Snellen best-cor-
rected visual acuity, slit-lamp biomicroscopy, fundus
photography, fluorescein angiography, SD OCT, and
FAF imaging at before IVB (pre-IVB) and 1 month
after the third last IVB (post-IVB) and 6 months after
the third last IVB. Follow-up for all patients was more
than 6 months. Visual acuity was expressed as
logMAR.
Spectral Domain Optical Coherence
Tomography Measurement
Spectral domain OCT examination was performed
at the initial visit and 1 month after every IVB during
the follow-up to evaluate ME and foveal micro-
structures. A 9 · 6-mm area of the macular region
centered on the fovea was examined using SD OCT
(Spectralis HRA + OCT; Heidelberg Engineering,
Heidelberg, Germany). Volume scans of 25 sections
(each with a distance of 240 mm) were centered on the
fovea, and 9 replicates of the B-scan image of each
section were averaged. The eye-tracking system was
used to assure that the scanning was performed in the
correct position. The parameters were recorded by
a well-trained technician who was unaware of the other
information about the eyes. Only images with a quality
score over 16 dB were selected as high-quality images.
An SD OCT image from the horizontal scans of the
midline was used for the analysis of foveal micro-
structures and was evaluated in a masked fashion
by two retina specialists. The integrity of the foveal
photoreceptor layer was measured using gray-scale
SD OCT images of the ELM and the IS/OS. A dis-
rupted length of the ELM and IS/OS was measured at
pre- and post-IVB using a caliper built into the soft-
ware of the SD OCT machine. We recorded the dim
ELM line and IS/OS line as “visible” if it was at all
detectable. A disruption in the ELM and IS/OS were
defined as the loss of the hyperreflective line. Central
macular thickness (CMT) was measured automati-
cally as an average retinal thickness within a circle
having a 500-mm radius centered on the fovea. The
thickness of the outer nuclear layer (ONL), as defined
by the distance between the outer border of internal
limiting membrane and the inner border of ELM, was
measured manually at the very center of the fovea
using a computer-based caliper measurement tool in
the SD OCT system (Figure 1).
Fundus Autofluorescence Measurement
and Quantification
Since the fluorescence intensity emitted by the RPE at
approximately 500 nm is extremely low, an SLO (HRA2;
Heidelberg Engineering, Heidelberg, Germany) was used
with both excitation illumination (488-nm laser) and
barrier filters in place. The degree of FAF was graded
subjectively (Grade 1–4) and independently by the 2
532 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2014 VOLUME 34 NUMBER 3

3. retina specialists. The grading of FAF was determined as
follows; samples with no or scarcely visible faint FAF
within 500 mm in diameter centered on the fovea as
Grade 1; samples with areas of increased FAF being less
than 1/2 of the total area of 500-mm circle as Grade 2;
samples with areas of increased FAF being between 1/2
and 3/4 of the total area of 500-mm circle as Grade 3;
samples with areas of increased FAF being more than 3/4
of the total area of 500-mm circle as Grade 4.
To confirm the grading of FAF, a semiautomatic
program was used to validate the subjective grading.
After the manual delineation of the one or more areas of
increased FAF by another ophthalmologist who was
unaware of the FAF grade and patient information, the
area and the average brightness of those segments were
determined using an appropriate image analysis program.
The area was measured by counting the pixels in the
delineated areas and multiplying this number by a scale
(micrometers per pixel) as determined by Spectralis
HRA + OCT. The average brightness was averaged
within the delineated area and was in the range from
0 (black) to 255 (white). Finally, an adjustment for back-
ground brightness was performed. The resulting value
represents the “flux” of the delineated areas, defined as
the average brightness (in lux units: flux per square
micrometer) multiplied by the area (square micrometers).
Statistical Analysis
Statistical analysis was performed using SPSS v.17.0
for Windows software (SPSS V17.0K, SPSS, Chicago,
IL). All values are presented as the mean ± standard
deviation. The data were analyzed by the Pearson cor-
relation test and Spearman correlation test to evaluate
associations between variables. Linear regression anal-
ysis was used to investigate the various parameters at
pre-IVB to determine whether they were significant
visual prognostic factors. Significant factors in the linear
regression analysis were included in a stepwise selection
of multivariate analysis, so that the independent effect of
each factor could be assessed. In multivariate analysis,
the variance inflation factor (VIF) was used as an indi-
cator of multicollinearity, which is an undesirable sta-
tistical phenomenon in which the variables in a multiple
regression model are highly correlated, meaning that
one can be linearly predicted from the others. Multi-
collinearity misleadingly inflates the standard errors of
the coefficients. Thus, it makes some variables statisti-
cally insignificant while they should be otherwise
significant. The VIF quantifies the severity of multicol-
linearity in a multivariate regression analysis. It provides
an index that measures how much the variance of an
estimated regression coefficient is increased because of
multicollinearity. If two variables are not correlated,
then all the VIFs will be one. A commonly given rule
is that VIFs $10 may be the reason for excessive multi-
collinearity. The diagnostic predictive power of all
examined parameters was finally assessed on the basis
of the receiver operating characteristic (ROC) curves
using categories of mean change from baseline in log-
MAR visual acuity. Effective cutoff values of the vari-
ables were optimized on the basis of the ROC analysis.
P # 0.05 was considered statistically significant.
Results
A total of 42 eyes of 42 patients (12 men and 30
women) were analyzed. The mean age of the subjects
was 55.67 ± 9.40 years. The mean disrupted ELM and
IS/OS length, CMT, and ONL thickness were decreased
when compared with the pre- and post-IVB. The differ-
ence was statistically significant by paired t-test. The
mean flux pre-IVB and post-IVB was 7.32 ± 5.64 and
2.26 ± 3.52, respectively (Table 1). The mean logMAR
visual acuities pre-IVB and post-IVB were 0.44 and
0.16, respectively. The difference was statistically sig-
nificant (P , 0.001, paired t-test).
Correlation Between Spectral Domain Optical
Coherence Tomography Parameters and
Visual Acuity
Better visual acuity pre-IVB and post-IVB were
closely associated with preserved photoreceptor integ-
rity pre-IVB and post-IVB, respectively. Longer dis-
rupted length of the ELM pre-IVB and IS/OS pre-IVB
correlated with larger logMAR visual acuity pre-IVB
(ELM; r = 0.709, P , 0.001, R2 = 0.5024 and IS/OS;
r = 0.405, P = 0.008, R2 = 0.1639, respectively). Sim-
ilarly, in accord with an increase in the disrupted length
of ELM and IS/OS post-IVB, the logMAR visual acuity
post-IVB increased (ELM; r = 0.842, P , 0.001, R2 =
0.7084 and IS/OS; r = 0.687, P , 0.001, R2 = 0.4721,
respectively). Visual acuity at both the pre- and
Fig. 1. Disrupted length of the ELM and IS/OS, and the ONL thickness
were measured manually at the fovea using a computer-based caliper
measurement tool in the SD OCT system.
CORRELATION OF FAF WITH VISION IN BRVO PARK ET AL 533

4. post-IVB was more closely correlated with the integrity
of the ELM than that of the IS/OS. Moreover, a signif-
icant positive correlation was observed between the
disrupted length of the ELM pre-IVB and the logMAR
visual acuity post-IVB, and between the disrupted
length of the IS/OS pre-IVB and the logMAR visual
acuity post-IVB (ELM; r = 0.634, P , 0.001, R2 =
0.4020 and IS/OS; r = 0.615, P , 0.001, R2 =
0.3783, respectively) (Table 2). Central macular thick-
ness pre-IVB was correlated with the logMAR visual
acuity post-IVB (r = 0.436, P = 0.004, R2 = 0.1899,
Pearson correlation test and linear regression test).
However, the ONL thickness pre-IVB was not corre-
lated with the logMAR visual acuity post-IVB.
Correlation Between Fundus Autofluorescence and
Spectral Domain Optical Coherence Tomography
Parameters and Visual Acuity
The subjectively graded measurement of FAF corre-
lated well with the flux (r = 0.847, P , 0.001, Spear-
man correlation test). We performed an analysis to
discern which of the SD OCT parameters were associ-
ated with flux in pre-IVB patients with treatment-naive
BRVO using Pearson correlation test and linear regres-
sion test. The larger logMAR visual acuity pre-IVB and
thicker CMT, longer length of IS/OS, and ELM defect
showed a significant association with flux pre-IVB.
However, the ONL at pre-IVB was not associated with
the flux pre-IVB (Table 3).
Linear regression analysis was used to investigate
various parameters, such as age, sex, presence of
diabetes mellitus and hypertension, logMAR visual
acuity, disrupted length of the ELM and IS/OS, CMT,
ONL thickness, and flux at pre-IVB to determine
whether they were predictive factors for the posttreat-
ment visual acuity. Among these parameters, logMAR
visual acuity, disrupted length of the ELM and IS/OS,
CMT, and flux pre-IVB were prognostic factors
correlated with the visual acuity post-IVB. Also, these
factors significantly correlated with the visual acuity at
6 months after the third IVB (Table 4). After identifying
potentially predictive factors by applying a stepwise
selection procedure, multivariate analysis was per-
formed. The disrupted length of IS/OS and ELM, and
CMT in SD OCT parameters were correlated with each
other, so these three factors were not considered simul-
taneously in multivariate analysis. To avoid the problem
of multicollinearity, multiple regression analysis
included only one of the SD OCT parameters, disrupted
IS/OS length pre-IVB, as this parameter had the largest
R2 value in the linear regression analysis. As a result,
the logMAR visual acuity and flux pre-IVB were the
significant factors predicting visual acuity post-IVB, and
the logMAR visual acuity and disrupted IS/OS length
pre-IVB were strongly associated with the visual acuity
Table 1. Mean Value in SD OCT Parameters and Flux at Pre- and Post-IVB (n = 42)
Pre-IVB Value Post-IVB Value P
SD OCT parameters
Disrupted ELM length (mm) 1,071.95 ± 817.77 210.45 ± 363.36 ,0.001*
Disrupted IS/OS length (mm) 536.67 ± 329.97 173.55 ± 373.45 ,0.001*
ONL thickness (mm) 134.00 ± 32.50 86.60 ± 28.23 ,0.001*
CMT (mm) 534.67 ± 167.82 307.86 ± 76.83 ,0.001*
FAF parameter
Mean flux 7.32 ± 5.64 2.26 ± 3.52 ,0.001*
Paired t-tests.
*Statistically significant difference.
Table 2. Correlation and Linear Regression Between SD OCT Parameters and VA (n = 42)
LogMAR VA Pre-IVB LogMAR VA Post-IVB
Disrupted ELM Length
Disrupted IS/OS
Length Disrupted ELM Length Disrupted IS/OS Length
Pre-IVB Post-IVB Pre-IVB Post-IVB Pre-IVB Post-IVB Pre-IVB Post-IVB
Regression coefficient 0.0003 0.0005 0.0004 0.0004 0.0002 0.0005 0.0004 0.0004
r 0.709 0.594 0.405 0.524 0.634 0.842 0.615 0.687
P ,0.001* ,0.001* 0.008* ,0.001* ,0.001* ,0.001* ,0.001* ,0.001*
R2 0.5024 0.3531 0.1639 0.2750 0.4020 0.7084 0.3783 0.4721
Pearson correlation test and simple linear regression analysis.
*Statistically significant difference.
r, Pearson correlation coefficient; VA, visual acuity.
534 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2014 VOLUME 34 NUMBER 3

5. at 6 months after the third IVB. This was supported by
low VIF values indicating that the multicollinearity
assumption was not violated. Indeed, the low VIF val-
ues are considered to be indicative that collinearity con-
cerns are minor (Table. 5).
Receiver operating characteristic analysis was per-
formed to assess the diagnostic potential of the investi-
gated factors in relation to visual acuity post-IVB. We
classified visual acuity into categories for ROC analysis,
which were based on the mean change of logMAR visual
acuity (−0.28) from pre-IVB to post-IVB. The analysis
showed that the flux pre-IVB and disrupted ELM length
pre-IVB had significant discrimination potential for
visual acuity improvement that was less than the mean
change of logMAR visual acuity. All the other factors did
not reach recognizable area-under-the-curve values, sen-
sitivity, and specificity. In other words, the ROC analysis
proved the significant diagnostic power of initial flux and
disrupted ELM length (Table 6). The representative cases
are shown in Figures 2 and 3.
Discussion
Generally, the visual prognosis for patients with
BRVO is relatively good, but exact prognosis remains
difficult. After the resolution of ME associated with
RVO, it is generally believed that a decrease in the
foveal thickness leads to an improvement in visual
acuity. In a clinical setting, however, some patients
achieve only limited visual improvement despite the
complete resolution of ME.9,15,16
Therefore, it may not
be sufficient that a decrease in the foveal thickness
alone can predict good visual recovery. The reason
for the various visual outcomes is not completely
understood, but the status of the foveal photoreceptor
layer and lower degree of increase in FAF might
explain the difference in visual outcome after the res-
olution of ME.
This study was able to demonstrate that the visual
acuity of patients with BRVO is significantly associ-
ated with the status of the photoreceptor layer as
assessed by ELM and IS/OS integrity using SD OCT
and the amount of FAF at the fovea and/or foveola.
The SD OCT findings in our study agree with past
studies. Wakabayashi et al17
reported that the ELM
may account for the visual recovery and the postoper-
ative preservation of the ELM may predict the sub-
sequent recovery of the foveal photoreceptor layer.
A good initial ELM status correlated with the restora-
tion of the IS/OS. In addition, good ELM status re-
sulted in a better visual outcome. In this study,
Table 3. Correlation and Linear Regression Between SD OCT Parameters, LogMAR VA and Flux at pre-IVB (n = 42)
Pre-IVB
Flux Pre-IVB
Regression Coefficient r P R2
LogMAR VA 0.0714 0.385 0.012* 0.1486
SD OCT parameters
Disrupted ELM length 0.0001 0.487 0.001* 0.2370
Disrupted ISOS length 0.0001 0.417 0.006* 0.1754
ONL thickness 0.0005 0.280 0.072 0.0785
CMT 0.0001 0.403 0.008* 0.1621
Pearson correlation test and simple linear regression analysis.
*Statistically significant difference.
r, Pearson correlation coefficient; VA, visual acuity.
Table 4. Association of LogMAR VA Post-IVB/LogMAR VA at 6 Months After Third IVB With Flux and SD OCT Parameters
Pre-IVB, Respectively (n = 42)
Pre-IVB
LogMAR VA Post-IVB LogMAR VA at 6 Months After Third IVB
Regression Coefficient P R2 Regression Coefficient P R2
LogMAR VA 0.4984 ,0.001* 0.4480 0.4295 ,0.001* 0.3868
SD OCT parameters
Disrupted ELM length 0.0002 ,0.001* 0.4020 0.0001 0.001* 0.2274
Disrupted IS/OS length 0.0004 ,0.001* 0.3783 0.0004 ,0.001* 0.4001
CMT 0.0006 0.004* 0.1899 0.0005 0.014* 0.1409
FAF parameter
Flux 1.8903 0.002* 0.2208 1.5887 0.005* 0.1813
Simple linear regression analysis.
*Statistically significant correlation.
VA, visual acuity.
CORRELATION OF FAF WITH VISION IN BRVO PARK ET AL 535

6. a disrupted ELM and/or IS/OS in foveola was associ-
ated with the visual outcome. This finding indicates
that the disrupted length of the ELM and the IS/OS is
representative of the severity of retinal damage and
the degree of retinal recovery after ME is reduced.
In accord with a previous study,18
a good initial
visual acuity was significantly associated with a good
visual outcome. Thus, visual prognosis and potential
visual acuity appear be determined before treatment.
Sekiryu et al19
reported that autofluorescence gen-
erated by the RPE cells travels through the sensory
retina, and the intensity of the FAF changes according
to the status of the overlying retina. Pece et al20
reported that increased FAF during CME in patients
with diabetes was not abnormal FAF, but instead, was
the normal fluorescence of the RPE as observed
through a defect in the xanthophyll pigment. However,
they also reported that the visual acuity of patients
with diabetic macular edema with increased FAF
was worse than that of patients without increased
FAF. We speculate that the FAF of patients with
ME would be high in cases in which the compensatory
function of retinal/glial cells does not effectively block
autofluorescence of the RPE and in cases in which
there are more activated microglia at the fovea. Thus,
FAF in patients with BRVO and ME is functionally
correlated with the status of retinal/glial cells in the
macula and could predict functional outcomes. More-
over, damaged retinal/glial cells initially associated
with increased FAF would be the reason for lower
initial visual acuity and lower chances of visual recov-
ery. Increased FAF at the foveola in patients with
BRVO and ME is of special significance in our study
because it was associated with a large value of CMT,
a longer length of the IS/OS and ELM defect, and
a large logMAR visual acuity.
Fundus autofluorescence at the foveola is very weak
or almost absent in normal eyes because lutein and
zeaxanthin are dense in the axons of the cone photo-
receptors at the foveola. Thus, increased FAF at the
foveola could indicate significant damage to cone
photoreceptor cells, thereby resulting in poor visual
acuity.
The main limitations of this study are the retrospec-
tive design and the relatively few patients. Although
SD OCT produces a better visualization of the retina
Table 5. Multivariate Analysis of the Associations With the LogMAR VA Post-IVB and LogMAR VA at
6 Months After Third IVB
Predictive Factor
LogMAR VA Post-IVB
Regressions Coefficient P VIF
LogMAR VA pre-IVB 0.4270 ,0.001* 1.1745
Flux pre-IVB 1.0013 0.049* 1.1745
Predictive Factor
LogMAR VA at 6 Months After Third IVB
Regressions Coefficient P VIF
LogMAR VA pre-IVB 0.3022 ,0.001* 1.1961
Disrupted IS/OS length pre-IVB 0.0003 ,0.001* 1.1961
Multivariate regression analysis.
*Statistically significant correlation.
VA, visual acuity.
Table 6. Evaluation of Diagnostic Potential of Prognostic Pre-IVB Factors in ROC Analysis
Pre-IVB
Improved VA Less Than Change of Mean VA at Post-IVB
AUC 95% CI P Sensitivity (%) Specificity (%)
LogMAR 0.5770 0.411–0.731 0.409 17.65 100
SD OCT parameters
Disrupted ELM length 0.7300 0.567–0.858 0.006* 64.71 91.3
Disrupted IS/OS length 0.6450 0.478–0.789 0.108 41.18 100
ONL thickness 0.6070 0.441–0.758 0.242 52.94 73.91
CMT 0.6140 0.447–0.763 0.214 52.94 73.91
FAF parameter
Flux 0.7140 0.549–0.845 0.012* 82.35 47.83
ROC analysis.
*Statistically significant correlation.
AUC, area under the curve; CI, confidence interval; VA, visual acuity.
536 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2014 VOLUME 34 NUMBER 3

7. layer than the time domain OCT, it is still not clear
whether these structures are truly absent when they are
invisible on SD OCT or whether they are not apparent
because of the insufficient resolution of the SD OCT
device or the presence of exudates, or hemorrhage, or
any other intraretinal abnormalities. In addition, Ho
et al21
reported that artifacts could be detected using
SD OCT, although these artifacts occurred at a lower
frequency than when time domain OCT was used. In
this study, the flux pre-IVB was correlated with the
Fig. 2. Case of a 53-year-old
man with ME treated with IVB.
The initial visual acuity was 20/
25. A. The disrupted length of
ELM and IS/OS was 41 mm and
753 mm, respectively, and CMT
was 442 mm. B. The FAF grade
was 1 and flux was 0.015 at
initial visit. After 3 times of
IVB, the visual acuity was
improved to 20/20. C. There
was no disruption of ELM and
IS/OS, and CMT was 304 mm.
D. The FAF grade was 1, and
flux was 0.006.
Fig. 3. Case of a 66-year-old
woman with ME treated with
IVB. The initial visual acuity
was 20/40. A. The disrupted
length of ELM and IS/OS was
832 mm and 540 mm, respec-
tively, and CMT was 410 mm.
B. The FAF grade was 4 and
flux was 0.204 at initial visit.
After 3 times of IVB, the visual
acuity was not improved (20/
100). C. The disrupted lengths
of ELM and IS/OS were 527
mm and 310 mm, respectively,
and CMT was 227 mm. D. The
FAF grade was 2 and flux was
0.031.
CORRELATION OF FAF WITH VISION IN BRVO PARK ET AL 537

8. visual acuity post-IVB, not with the visual acuity at
6 months after the third IVB. These results were
thought to be because of the short follow-up period.
Therefore, a further study with a longer follow-up
period is needed to confirm the association between
them. Lastly, because of the lack of standardization in
the related equipment and the insufficiency of stan-
dardized normative database for FAF, currently, FAF
is not a standard examination method for retinal and
RPE diseases. Therefore, interpretation of the FAF
image is not yet common in everyday practice.
In summary, along with the initial visual acuity and
the integrity of the ELM and IS/OS, the FAF image
could be a visual prognostic factor in patients with
BRVO. The analysis of the initial FAF in patients with
BRVO seems to be an important prognostic factor for
distinguishing patients in whom visual function may
improve after therapy from those patients in whom
retinal damage has already progressed, which limits
a positive functional outcome. Functional improve-
ment after the treatment of ME can be quantified by
FAF and correlates morphologically with traits exam-
ined by SD OCT. In this study, multiple regression
analysis showed that the flux pre-IVB was the
significant predictive factor associated with the visual
acuity post-IVB (Regression coefficient = 1.0013, P =
0.049). Both linear and multivariate analyses proved
flux pre-IVB to be a significant predictor for the visual
acuity post-IVB. In addition, the ROC analysis con-
firmed the flux pre-IVB as the significant predictor of
visual outcome. Therefore, we can conclude that the
flux pre-IVB influences the visual acuity post-IVB in
patients with BRVO.
Finally, the correlation between the FAF and the
visual acuity post-IVB in patients with BRVO could
aid in the prediction of visual recovery, especially in
patients whose foveal microstructure could not be
adequately evaluated.
Key words: bevacizumab, branch retinal vein occlu-
sion, external limiting membrane, fundus autofluores-
cence, prognostic factor, SD OCT.
References
1. Demir M, Oba E, Gulkilik G, et al. Intravitreal bevacizumab
for macular edema due to branch retinal vein occlusion:
12-month results. Clin Ophthalmol 2011;5:745–749.
2. Olk RJ. Modified grid argon (blue-green) laser photocoagula-
tion for diffuse diabetic macular edema. Ophthalmology 1986;
93:938–950.
3. Martidis A, Duker JS, Greenberg PB, et al. Intravitreal triamcin-
olone for refractory diabetic macular edema. Ophthalmology
2002;109:920–927.
4. Lewis H, Abrams GW, Blumenkranz MS, et al. Vitrectomy for
diabetic macular traction and edema associated with posterior
hyaloidal traction. Ophthalmology 1992;99:753–759.
5. Haritoglou C, Kook D, Neubauer A, et al. Intravitreal bevaci-
zumab (Avastin) therapy for persistent diffuse diabetic macular
edema. Retina 2006;26:999–1005.
6. Arevalo JF, Fromow-Guerra J, Quiroz-Mercado H, et al. Pan-
American Collaborative Retina Study Group. Primary intravi-
treal bevacizumab (Avastin) for diabetic macular edema: results
from the Pan-American Collaborative Retina Study Group at
6-month follow-up. Ophthalmology 2007;114:743–750.
7. Chen TC, Cense B, Pierce MC, et al. Spectral-domain optical
coherence tomography: ultra-high speed, ultra-high resolution
ophthalmic imaging. Arch Ophthalmol 2005;123:1715–1720.
8. de Boer JF, Cense B, Park BH, et al. Improved signal-to-noise
ratio in spectral-domain compared with time domain optical
coherence tomography. Opt Lett 2003;28:2067–2069.
9. Murakami T, Tsujikawa A, Ohta M, et al. Photoreceptor status
after resolved macular edema in branch retinal vein occlusion
treated with tissue plasminogen activator. Am J Ophthalmol
2007;143:171–173.
10. von Rückmann A, Fitzke FW, Bird AC. Distribution of fundus
autofluorescence with a scanning laser ophthalmoscope. Br J
Ophthalmol 1995;79:407–412.
11. Schmitz-Valckenberg S, Fleckenstein M, Göbel AP, et al.
Evaluation of autofluorescence imaging with the scanning laser
ophthalmoscope and the fundus camera in age-related geo-
graphic atrophy. Am J Ophthalmol 2008;146:183–192.
12. Spaide RF, Noble K, Morgan A, et al. Vitelliform macular
dystrophy. Ophthalmology 2006;113:1392–1400.
13. Vujosevic S, Casciano M, Pilotto E, et al. Diabetic macular
edema: fundus autofluorescence and functional correlations.
Invest Ophthalmol Vis Sci 2011;52:442–448.
14. Chung H, Park B, Shin HJ, et al. Correlation of fundus auto-
fluorescence with spectral domain optical coherence tomogra-
phy and vision in diabetic macular edema. Ophthalmology
2012;119:1356–1365.
15. Ota M, Tsujikawa A, Murakami T, et al. Association between
integrity of foveal photoreceptor layer and visual acuity in
branch retinal vein occlusion. Br J Ophthalmol 2007;91:
1644–1649.
16. Ota M, Tsujikawa A, Murakami T, et al. Foveal photoreceptor
layer in eyes with persistent cystoids macular edema associated
with branch retinal vein occlusion. Am J Ophthalmol 2008;
145:273–280.
17. Wakabayashi T, Oshima Y, Fujimoto H, et al. Foveal micro-
structure and visual acuity after retinal detachment repair:
imaging analysis by Fourier-domain optical coherence tomog-
raphy. Ophthalmology 2009;116:519–528.
18. Kondo M, Kondo N, Ito Y, et al. Intravitreal injection of bev-
acizumab for macular edema secondary to branch retinal vein
occlusion: results after 12 months and multiple regression anal-
ysis. Retina 2009;29:1242–1248.
19. Sekiryu T, Lida T, Maruko I, et al. Clinical application of
autofluorescence densitometry with a scanning laser ophthal-
moscope. Invest Ophthalomol Vis Sci 2009;50:2994–3002.
20. Pece A, Isola V, Holz F, et al. Autofluorescence imaging of
cystoid macular edema in diabetic retinopathy. Ophthalmologica
2010;224:230–235.
21. Ho J, Sull AC, Vuong LN, et al. Assessment of artifacts
and reproducibility across spectral- and time-domain optical coher-
ence tomography devices. Ophthalmology 2009;116:1960–1970.
538 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2014 VOLUME 34 NUMBER 3