IMAGING IN GLAUCOMA- investigation based approach

IMPORTANCE OF IMAGING IN GLAUCOMA
• Functional loss in glaucoma being insiduous in
nature..asymptomatic in early stages
• Fallacies of current diagnostic techniques – no
direct way to measure neuronal loss in glaucoma
• Surrogate measures – structural and functional
assessment
• With the advent of automated imaging-
structural changes can be reliably assessed
before functional damage

• Sommer et al showed that 50–85% of patients
who developed visual field defects had RNFL
defects prior to visual field loss, as much as 6
years.
• Early diagnosis is a must to safe guard the
ganglion cells at risk
Sommer A, et al. Clinically detectable nerve fiber atrophy precedes the onset of glaucomatous field loss. Arch Ophthalmol 1991; 109(1):77–83.

ONH and RNFL analysis
• Ophthalmoscopy
Direct or S/L using 90 D or 78D
• Fundus camera with red-free filter
• Retinal contour analysis -HRT
(Scanning laser ophthalmoscope)
• OCT
• Scanning laser polarimetry –GDX VCC

Fundus photography
• RNFL is invisible to red light but easily seen
with short wavelength light..red free light
• Good visualisation is facilitated by clear media,
high contrast (dark pigmented fundus
background)
• RNFL photography-described by Airaksinen &
Nieminen

Fundus photography
• Healthy RNFL appears slightly opaque with visible
radially oriented striations emanating from the optic
nerve head
• Glaucomatous defects-
slit / groove / wedge /diffuse
Quigley detected RNFL damage in 84% of eyes with
visual field loss, in 3% of normal eyes, and in 13% of
eyes with suspicious glaucoma and no visual field loss
Arch Ophthalmol 1980; 98:1564–1571

Fundus photography
• Sommer et al demonstrated RNFL photography
to have a sensitivity of about 80% with a
specificity of 94%.
• Disadvantage – no quantification of RNFL damage
Sommer A, Quigley HA, Robin AL, et al. Evaluation of nerve fiber layer assessment. Arch Ophthalmol 1984; 102:1766– 1771.

IMAGING IN GLAUCOMA- investigation based approach

Pearls
Appreciation of even the subtle
qualitative changes
&
Accurate quantification of such changes
is fundamental

OCT
• Optical coherence tomography (OCT) is based
upon the principle of low-coherence
interferometry
• OCT differs from the confocal scanning laser
tomography - it provides histological
representation of the tissue and is also not
subjected to anterior segment polarization
issues as seen with scanning laser polarimetry

Stratus OCT 3/Time Domain
• OCT 3 has several built-
in protocols for
scanning the retina and
the optic nerve head
and for analysing the
images obtained.
• The protocols are
represented by
descriptive icons in the
software

Fast scan protocols
• The "fast" scan protocols of the OCT 3 reduce the
time needed for multiple scans from 3+ seconds
to about 2 seconds.
• The scan time reduction is intended to minimize
the error created by patient movement.
• The downside of the fast scan is that fewer A-
scans are grabbed in the 6 mm length of the scan.
• The normal 6 mm scan contains 512 A-scans,
whereas the fast 6 mm scan contains only 128
A-scans

OCT in GLAUCOMA
3 STRATEGIES
• RNFL ANALYSIS AROUND OPTIC DISC
• ONH ANALYSIS
• MACULAR THICKNESS ANALYSIS

Limitations of time domain OCT
3 major sources of error-
• Motion artifacts
• Failure to identify the upper and lower
borders of the RNFL on low signal to noise
ratio
• Lack of registration

RNFL Normative Data Display
RNFL Normative Distribution
Of the normal population:
• 5% WHITE
• 90% GREEN
• 4% YELLOW
• 1% RED
• 5% fall within the white band
• 95% fall within or below the green band
• 90% fall within the green band
• 5% fall within or below the yellow band
• 1% fall within the red band; (considered
outside the normal limit)

ONH ANALYSIS
• Radial line scan
through optic disc
• 6 consecutive line
scans in a star
pattern on the disc

ONH Analysis
RPE Markers
• The end of the RPE is indicated
Shown blue on the scan image,red on
the composite diagram.
• Disc Diameter
• Illustrated and measured on a
straight line between the 2 RPE
markers. Blue line on scan image.
• Cup Diameter
• Illustrated and measured on a
straight line parallel to, and 150 μm
anterior to the disc diameter line.
Red on image,green on diag, this line
is adjustable.
• Rim Area
Indicated with red shading onscan
image, this area is bounded by the
cup diameter line and a line from the
RPE marker to the anterior surface of
the disc,at a 90-degree angle to the
cup diameter line.

Glaucoma analysis protocols on cirrus OCT
• RNFL and ONH: Optic disc cube 200X200
• Guided progression analysis
• 3 D visualisation
• GPA by manual selection

CIRRUS – RNFL and ONH analysis report

Cirrus GPA report – optional printout

Newer/Spectral domain OCT
OCT TIME
DOMAIN/STRATUS
SPECTRAL
DOMAIN/CIRRUS
Scan protocol Fast RNFL Optic disc scan
Scan 256 scans x 3 circles 512 scans x 1
Laser source
wavelength
820nm 840nm
A-Scan speed 400/sec 27,000 / sec
Axial resolution 8-10µ 5µ
Transverse
resolution
20µ 15µ

Heidelberg Retina Tomograph
• The Heidelberg Retina Tomograph is a confocal laser
scanning system designed for acquisition and analysis of three
dimensional images of the posterior segment.
• It enables the quantitative assessment of the topography of
ocular structures and the precise follow-up of topographic
changes thereby detecting and monitoring very subtle and
early glaucomatous nerve head changes

Mode of action
• Based on Confocal scanning laser
technology
• Confocal optics (including a 670 nm
diode laser) are used to obtain
multiple measures of retinal height,
each with a very shallow depth of
field
• Multiple con-focal ‘slices’ (XY
measurements in multiple Z
planes), obtained at consecutive
focal planes are integrated to
provide a three-dimensional
reconstruction from lamina
cribrosa to the anterior surface of
the retina.

Mode of action
• For the current version of the
Heidelberg Retina Tomograph (HRT
II/HRT 3), the scan is centered on
the optic disc and a variable
number of slices are obtained at a
rate of 16/mm of scan depth.
• The depth of scanning ranges from
0.5 mm to 4 mm and image field of
view is 15°, yielding 384x384
picture elements.
• Three consecutive scans are
obtained within 2 seconds and
automatically combined to
improve reproducibility.

Analysis, printouts and interpretation
• Stereometric parameters are
provided to describe the retinal
topography of each image
• Stereometric parameters are
obtained within the optic disc
margin, which is defined
manually by placing a contour
line along the inner edge of the
scleral ring at the baseline
examination
• The contour line is automati-
cally transferred to all follow-up
examinations

Analysis
• Stereometric parameters are calculated
based on a standard reference plane set
50µm posterior to the average contour
line height
• In glaucoma the nerve fibers at the
papillo-macular bundle remain intact
longest and the nerve fiber layer
thickness at that location is
approximately 50 microns
• All structures located below the
reference plane are considered to be
cup, all structures located above the
reference plane and within the contour
line are considered to be rim

Description of HRT stereometric parameters

Moorfields Regression Analysis (MRA)
• Moorfields' Regression Analysis
makes comparisons between
neuroretinal rim area, disc area, to
a normative database to make an
initial analysis and classification.
• In addition to the numeric tables, it
also provides a graphical analysis
of optic nerve and rim area
topography by dividing the nerve
into six sectors and reporting the
status of each sector.
• A check mark means "within
normal limits," an exclamation
mark means "borderline," and an X
suggests that the topography is
"outside normal limits."

Topography standard deviation (TSD)" value
• Each image is
accompanied by a
"topography standard
deviation (TSD)“/MPHSD
(mean pixel height std
deviation) value
• Image quality is
statistically graded by the
HRT so that clinicians can
delete images that fall
below acceptable criteria

• In the contour graph the white
line represents the reference
plane at which there is a height
of zero.
• The red line represents the
height of the reference line
between the cup and disk.
• The green line is the retinal
height of the subject eye at the
contour line showing the typical
double hump feature at the
superior and inferior poles.

Interpretation
The printed report shows
• The topographic (top) and
reflectivity image (bottom)
illustrates the ONH.
• The topographic image is shown
with the cup represented in red,
the sloping neural tissue in blue,
and the rim in green.
• Reflectivity image

Progression analysis/GPS (Glaucoma probability
score)
• Change over time can be
measured in several ways
1. Linear regression analysis can be
obtained that describes the
change in stereometric
parameters from the baseline
examination to follow-up
examination..TREND ANALYSIS
2. With the ‘Follow-up Report,’ the
Topographic Change Analysis
(TCA) analysis, MRA analysis,
and difference in stereometric
parameters are displayed

Topographic change analysis
3. A more sophisticated analysis based
on local change in retinal height
compared to baseline is available.
• TCA compares the change in retinal
height between baseline and follow-
up images to the change in retinal
height between the three images
that compose the baseline image,
using analysis of variance on a
super-pixel by super pixel basis
• Super pixels that change
significantly compared to base-line
are flagged as red (decrease in
retinal height) or green (increase in
retinal height).

• Clusters of red contiguous pixels
in areas most associated with
glaucomatous change
(e.g.STQ&ITQ) likely indicate
glaucomatous progression. The
volume and area of these clusters
can be charted over time using
the ‘Cluster Change Analysis’
printout
• TCA flags of progression when
change exceeds measurement
variability and cluster of 20 or
more significantly depressed
super pixels are noticed.

Strengths and limitations
Strengths
• Newest generation of HRT
(HRT III) has race-specific
normative data-base
• Real-time quality control
during image acquisition,
sophisticated analysis
software for glaucoma
detection and progression
Limitations
• Small normative data with no
racial representation
• Many measurements rely on a
reference plane based on the
placement of a user-defined
contour line
• Another possible limitation of
HRT is that stereometric
measurements can be
influenced by moderate
changes in IOP

Scanning laser polarimetry
• Only RNFL analyser to be specifically developed for
glaucoma
• PRINCIPLE: Birefringent property of the axons
(attributed to microtubules within the axons) causes
the polarized light to undergo a phase shift and the
degree of phase shift in the light that returns is
measured by a detector.
• The amount of phase shift, also known as
retardation, is directly proportional to the amount
of nerve fiber layer

GDx VCC
• The total retardation signal detected by scanning
laser polarimetry consists of contributions from the
anterior segment (Cornea and lens) and the RNFL.
• GDx-VCC, is equipped with a variable corneal
compensation (VCC) that allows eye-specific
compensation of the birefringent effects of the
anterior segment
• Anterior segment birefringence is assessed with the
method described by Zhou and Weinreb

Mechanism
• A near-infrared laser (wavelength,
785 nm) scans the fundus at a
40°x20° (HXV) scanning angle
• The interaction between the
birefringence of the radially oriented
axons (Henle’s fiber layer) in the
macula and the birefringence of the
anterior segment results in a bow tie
shaped pattern in the retardation
image
• An algorithm determines the anterior
segment birefringence and the
software automatically negates the
anterior segment birefringence

Mechanism
• A second scan is then obtained
and the patient’s RNFL thickness is
then estimated with eye-specific
corneal compensation in a 20°x20°
field of view at a resolution of 128
pixelsx128 pixels.
• The software positions circles of
3.2mm outer diameter and
2.4mm inner diameter centered
on the disc. Based on the
retardation values beneath this
band (0.4mm wide), the software
calculates six parameters.

Gdx pro/ Gdx ECC
• In subset of eyes Gdx VCC shows atypical
birefringence patterns,
(such that the brightest areas of the retardation maps
are not consistent with the thickest portion of
peripapillary RNFL)
• Enhanced corneal compensation algorithm improves
the signal to noise ratio and eliminates artifacts
associated with atypical birefringence

Procedure
• Test is completely objective and makes the
actual measurements in less than one second.
• It works best on undilated pupils and can
work quite well through cataracts up to 20/60.
• A normative database compares the patient's
measurements to those of normal patients of
the same age, sex, and race.
• Reported sensitivity of 96% and specificity of
93%.

Interpretation GDx VCC
• The first thing to consider in the interpretation of a scan
is whether it is of high quality. Every image has Q score
representing the overall quality of the scan.
• The Q score (displayed above the Fundus Image) ranges
from 1-10, with values 8-10 representing acceptable
quality.
This score is based on a number of factors including
 Focusing
 Illumination
 Centration of the optic disc
 Ellipse placement.

Quantitative RNFL
evaluation is provided
through four key elements
of the printout: the
• Thickness Map,
• Deviation Map,
• TSNIT graph,
• Parameter Table

Interpretation
• The fundus image is useful
to check for image quality
• Fundus image in a high
quality GDx VCC image is
properly focused, evenly
illuminated, with well
centered optic disc and
distinct blood vessels.

Interpretation-Thickness map
• The thickness map shows
the RNFL thickness in a
color-coded format from
blue to red.
• Thick RNFL values are
colored yellow, orange, and
red while thin RNFL values
are colored dark blue,light
blue, and green.

Deviation map
• The deviation map reveals the
location and magnitude of
RNFL defects over the entire
thickness map
• The Deviation Map uses a
grayscale fundus image of the
eye as a background, and
displays abnormal grid values
as colored squares over this
image
• This helps the user determine
precisely the location of the
abnormalities

Deviation map
• For each scan, the RNFL thickness at
each super pixel is compared to the
age-matched normative database,
and the super pixels that fall below
the normal range are flagged by
colored squares based on the
probability of normality.
• Dark blue squares
represent areas where the
RNFL thickness is below the
5th percentile of the
normative database.
• Light blue squares represent
deviation below the 2% level,
yellow represents deviation
below 1%, and red represents
deviation below .05%.

TSNIT Graph
• The TSNIT map is displayed
at the bottom of the
printout.
• In a normal eye the TSNIT
plot follows the typical
‘double hump’ pattern, with
thick RNFL measures
superiorly and inferiorly and
thin RNFL values nasally and
temporally
The TSNIT graph for a healthy eye lies in
the normal range (95%) for a given age as
shown in the shaded area.

TSNIT Graph
• TSNIT graphs for both eyes
are displayed together at
the bottom of the printout
• In a healthy eye there is
good symmetry between
the TSNIT graphs of the two
eyes and the two curves will
overlap. A dip in the curve
of one eye relative to
another is indicative of
RNFL loss.

Parameter table
• The Parameter Table for a
healthy patient
• The values of each
parameter (except the NFI)
are displayed in green if they
fall within the normal range
• Abnormal values are
colorcoded based on their
probability value, similar to
the super-pixels in the
Deviation Map
• The NFI values are not color-
coded even if abnormal

Parameters
The five TSNIT parameters
are:
• TSNIT Average
• Superior Average
• Inferior Average
• TSNIT Standard
Deviation (TSNIT SD),
• Inter-eye Symmetry
• TSNIT Average: The
average RNFL thickness
around the entire
calculation circle.
• Superior Average: The
in the superior 120°
region
• Inferior Average: The
in the inferior 120°

TSNIT SD
• TSNIT SD: This measure captures
the modulation (peak to trough
difference) of the double-hump
pattern.
• A normal eye will have high
modulation in the double-hump
RNFL pattern,
• while a glaucoma eye will typically
have low modulation in the
double-hump pattern.
• As a result, high modulation will
have a high TSNIT SD value while
low modulation will have low
TSNIT SD value

Inter –eye symmetry
• Inter-eye Symmetry: Measures the
degree of symmetry between the right
and left eyes by correlating the TSNIT
functions of the two eyes.
• Values range from –1 to 1, where
values near one represent good
symmetry.
• Normal eyes have good symmetry with
values around 0.9.
• This measure is the Pearson Product
correlation coefficient (r-value) from
the correlation of the TSNIT curves of
the two eyes
• The parameter is very useful because in
glaucoma, one eye is often more
advanced than the fellow eye.

NFI (Nerve fiber indicator)
• The NFI is a global measure based on the entire RNFL
thickness map. It is calculated using an advanced form of
neural network, called a Support Vector Machine (SVM).
• NFI is a single value that ranges from 1-100 and indicates the
overall integrity of the RNFL.
• The NFI is not color-coded based on probability, but rather it
is based on an absolute scale.
• Classification based on the ranges:
 1-30 -> normal
 31-50 -> borderline
 51+ -> abnormal

NFI
• Clinical research has shown that the NFI is the
best parameter for discriminating normal from
glaucoma
• The sensitivity and specificity of the NFI has
been reported to be extremely high, with values
of 89% and 98% respectively

Split Bundles vs Wedge Defects
• A split bundle is when the
RNFL bundle is divided
into two more-or-less
nearly symmetrical
segments. The split
bundle occurs where the
RNFL is thickest, in the
superior and inferior
regions
• The resulting RNFL
profile will have 3 or 4
humps rather than 2

The three main clues to distinguish a split bundle defect
form a wedge defect are
1) Location: split bundle defects are most often directly
superior or inferior to the disc while wedge defects are
usually more temporal
2) Extent: split bundle defects do not extend all the way to
the disc margin while wedge defects do extend to the disc
margin, and
3) Visibility: split bundles are not easily visible on the
Thickness Map while wedge defects frequently are visible

Detecting RNFL Change Over Time: Serial
Analysis
• The GDx VCC monitors RNFL change over time
and displays this analysis in a printout called
Serial Analysis. It has five key elements that
should be considered when assessing RNFL
change over time
• These elements include the Thickness Maps,
Deviation Maps, Deviation from Reference
Maps, Parameters Tables, and TSNIT Graph.

The Deviation from Reference Map
displays the RNFL difference, pixel by
pixel, of the followup exam compared
to the baseline exam.
If the difference exceeds 20 microns at
any pixel, the pixel is color coded
according to the legend
RNFL change is color coded in 20 micron
increments,where the first 20 micron
change is coded dark green, a 40
micron change is coded light
blue, 60 is dark blue, etc.

• Serial Analysis can compare up to four exams
• The first exam is the baseline or reference exam,
and all follow-up exams are compared to this
baseline exam. A colored rectangle to the left of
the Thickness Map contains the date and quality
score of each exam.
• The same color is used in the TSNIT graph to
indicate which TSNIT curve corresponds to which
exam (i.e., yellow rectangle for first exam
corresponds to yellow curve on theTSNIT graph.)

Advantages
• GDx larger normative data
• Measures the “actual thickness”
• More RNFL specific
• Higher resolution 5microns
• Shorter test time
• Undilated pupil with media opacities

Comparison
• Correlation of structural and functional measures using
Heidelberg Retinal Tomography and Spectralis spectral
domain optical coherence tomography at different levels of
glaucoma severity demonstrated Spectral domain optical
coherence tomography retinal nerve fibre layer
measurements demonstrated closer correlations to visual
field threshold reductions
• The sensitivity of RNFL damage detection using HRT-III was
lower compared with OCT-3, especially in early glaucoma.
• Leaney J, Healey PR, Lee M, Graham SL - Clin. Experiment. Ophthalmol. - Nov 2012; 40(8); 802-12
• Moreno-Montañés J, Antón A, García N, Olmo N, Morilla A, Fallon M - J. Glaucoma - Sep 2009; 18(7); 528-34

Comparison
• A study by Federico et al demonstrated, diagnostic performance of
qualitative evaluation of stereoscopic optic disk photographs by glaucoma
specialists and contemporary versions of three quantitative imaging
techniques (StratusOCT GDx VCC, and HRT III) in patients with early to
moderate perimetric glaucoma.
• StratusOCT, GDx VCC, and HRT III performed as well as, but not better than,
qualitative evaluation of optic disk stereophotographs for detection of early
perimetric glaucoma.
• For StratusOCT, they found average RNFL thickness and thickness in the
inferior quadrant to have the largest AUC and the greatest sensitivities at
80% and at 95% specificities
American Journal of Ophthalmology, 2007-11-01, Volume 144, Issue 5, Pages 724-732

• On HRT II Rim area and
cup shape
measurement were
found to have the
largest AUCs
• Among GDx VCC
parameters, NFI and
superior average RNFL
thickness had the
largest AUCs and the
highest sensitivities at
80% and 95%
specificities.
Graph showing a comparison of the areas under receiver operator characteristic
curves (± standard error [SE]) of the best parameters from StratusOCT; average
retinal nerve fiber layer thickness [RNFL] , 0.96 ± 0.02), GDx VCC; nerve fiber
indicator, 0.92 ± 0.03), Heidelberg Retina Tomograph (HRT) III; (Frederick S.
Mikelberg [FSM] discriminant function, 0.91 ± 0.03), and the cumulative score of
the three observers for disk photograph evaluation (0.97 ± 0.02).

The Visual Fields for this patient show a superior arcuate scotoma
in both eyes.
PSD PLOT PSD PLOT
This example is a 30 year old Asian male with
POAG OU. He is myopic with a sphere of –5.25
in both eyes and tilted optic discs. Treated IOP is
18 mm Hg in both eyes. His visual fields
show a superior arcuate scotoma in both eyes,
worse in the right eye. These
global indices are indicative of early
glaucomatous field loss.

The Visual Fields for this patient show a clear nasal step inferiorly for both
eyes.

• This example is from a 63 year old male.
Intraocular pressure for this patient is 22 mm
Hg for the right eye and 20 mm Hg for the left
eye.
• The visual field for this patient reveals an
inferior nasal step in both eyes. The MD and
PSD for the right eye are –4.96 and 7.4
respectively and –4.14 and 6.31 for the left
eye. The superior visual field for both eyes is
still normal.

The Visual Fields for this patient show superior loss in the right eye
and both superior and inferior loss in the left eye.

• This example is a 69 year old male with
advanced visual field loss OU.
• The MD and PSD for the right eye is –
11.29 and 14.91 respectively, and –18.81
and 11.73 for the left eye.

IMAGING IN GLAUCOMA- investigation based approach

More Related Content

Similar to IMAGING IN GLAUCOMA- investigation based approach (20)

Recently uploaded (20)

IMAGING IN GLAUCOMA- investigation based approach

Editor's Notes