Page 29 - CJO_F17_GLAUCOMA_SUPPLEMENT

MANAGING OPEN ANGLE GLAUCOMA

It is also important to remember that glaucoma is a clinical, not a statistical diagnosis, despite the analysis that
accompanies each scan. Reference databases are helpful, but have limitations. As an example, the Cirrus RNFL
220
reference database comprises only 284 individuals aged 19 to 84: just 31 were older than 70, 43% were Caucasian,
and none had any associated ocular or systemic disease. Applying these data to a 49-year old individual, RNFL
thickness can decrease by 30% (from 107 to 75μm) yet still remain in the ‘normal’ range. Reference databases
and segmentation algorithms are also instrument-specific, meaning that comparing measures from two different
instruments is all but impossible. 221

Objective imaging can complement a clinical examination based upon the ‘Five Rs’ paradigm.
1. Use the scleral Ring to determine the size of the optic nerve head:

OCT quantifies disc size by delineating Bruch’s membrane opening (BMO), the true anatomic bottle-
neck through which all RGC axons must pass. If nothing else, this has demonstrated that clinicians
222
consistently over-estimate disc size with ophthalmoscopy, which in turn leads to an over-estimation of
neuroretinal Rim (NRR) thickness. While OCT may be of particular value with anomalous discs where
223
subjective identification of the scleral Ring is difficult, atypical anatomy also makes comparison to refer-
ence databases of questionable value.
220
2. Identify the width of the neuroretinal Rim
Confocal scanning laser ophthalmoscopy (CSLO: Heidelberg Retinal Tomography, HRT) and OCT can
accurately identify glaucomatous NRR thinning: a rim area <1mm or a statistically abnormal vertical
2
cup-to-disc ratio should be considered suspicious. Progressive NRR thinning is also predictive of fu-
224
ture VF loss: in the Ocular Hypertension Treatment Study, HRT was able to identify structural change up
to 8 years before VF assessment detected functional change. 225,226
Some of the currently available OCTs have the ability to detect and quantify neuroretinal Rim defects
while others do not. The Heidelberg Spectralis utilizes the detection of Bruch’s membrane opening to
quantify a parameter known as Bruch’s membrane opening-minimum rim width (BMO-MRW). This
value represents the shortest distance between BMO and the internal limiting membrane that forms the
anterior border of the ONH. BMO-MRW has proven to be as sensitive in detecting glaucomatous dam-
227
age and its progression as RNFL thickness, and may be particularly valuable in highly myopic patients
where clinical disc margins may be difficult to delineate and RNFL bundles may be shifted tempo-
rally. 228,229 However, objective imaging does not replace careful clinical evaluation: reference databases
can be confounded by anomalous ONHs and OCT cannot detect the rim pallor that often accompanies
non-glaucomatous optic neuropathy. 178,230

3. Examine the Retinal nerve fiber layer

The Retinal nerve fiber layer is where objective imaging with OCT initially made its mark, and still
shines. Subtle RNFL thinning often precedes VF loss but can be difficult to appreciate clinically; how-
ever, both corneal compensated scanning laser polarimetry (SLP: GDx) and OCT are able to detect this
thinning much earlier than clinical examination. 231,232 When utilizing spectral domain OCT, the param-
eter with the best diagnostic accuracy tends to be average RNFL thickness, followed by inferior and
superior quadrant thicknesses. A follow-up study comparing RNFL, ONH, and ganglion cell complex
233
(GCC) values confirmed the RNFL assessment software as the best at detecting glaucomatous damage.
234
Although the use of multiple parameters could increase false-positive results, structural damage may be
present in one parameter and not another, meaning that it is helpful to have information from the ONH,
RNFL and macula in glaucoma diagnosis. 213,235,236

It is important to always subjectively assess the RNFL deviation map (the OCT equivalent of a red-free
photograph) and not simply ‘trust the numbers’: over-diagnosis (the ‘red disease’ of anomalous ONHs
and high myopia) or under-diagnosis (incorrectly assuming that ‘green is always good’) are definite
risks. 48,219,237-239 Figure 10 exemplifies a case of ‘green disease’ in a patient with POAG. Relying entirely on
summary parameters or reference database comparisons (in fact, on any single structural or functional
test result in isolation) is simply not good enough. As a general rule, however, an average RNFL thick-
240

CANADIAN JOURNAL of OPTOMETRY | REVUE CANADIENNE D’OPTOMÉTRIE VOL. 79 SUPPLEMENT 1, 2017 29