Monitoring glaucoma progression

Introduction

When dealing with patients affected by chronic open-angle glaucoma, it is important to remember that an accurate and systematic follow-up is just as important as an early diagnosis of the disease. It is rather pointless to spend vast amounts of resources and time in the diagnostic phases, unless proper treatment and careful follow-ups are not continued throughout this chronic disease. It is also of utmost importance to assess the rate of disease progression in each patient in order to properly determine how aggressive treatment should be, which ought to be based on damage

Corresponding author. Tel.: ++432-552747; Fax: ++432-552741; E-mail: brusini@libero.it

severity, rate of ganglion cell loss, and patient life expectancy. An adequate and sufficient treatment can be defined if it is able to stop, or at least significantly slow down, the rate of glaucomatous damage progression, thus limiting severe visual impairment.

To fully assess glaucomatous progression, both structural and the functional damage need to be considered. In early diagnosis of the disease, morphological alterations of the optic disc (OD) or retinal nerve fiber layer (RNFL) can precede visual field (VF) defects, or, even if quite rarely, vice versa. The same holds true when monitoring progression, in that structural damage worsening can often be seen before corresponding VF defect progression; however, the opposite can also occur at times (Chauhan et al., 2001; Hudson et al.,

60

2007). In addition, disease management can properly be assessed by considering whether or not the various types of information regarding structure and function are in accordance.

Monitoring structural damage progression

Structural damage evaluation and its progression can be assessed with both lowand high-tech testing methods (Jonas et al., 1999; Hoffmann et al., 2007). The simplest technique consists of OD clinical assessment using slit lamp biomicroscopy with a 78-diopter (or similar) Volk lens. When properly performed, the ophthalmologist can gather precious information on both the presence of structural loss and the progressive worsening of previously existing morphologic defects. In assessing progression, it is mandatory to use standardized methods in which observed structural alterations can be recorded and compared over time. A good starting point is a hand-drawn representation; a picture is worth a thousand words, and can add considerably to a simple written description. This method is of course subjective and it is highly dependent on the observer’s experience and drawing capabilities. To overcome these limitations, objective and standardized method have been used over the years, and several classification systems have been proposed to subdivide structural damage in different stages of severity (Spaeth et al., 2006).

The disc damage likelihood scale (DDLS) was recently designed by Spaeth et al. (2003); it takes both OD size and radial neuroretinal rim width (measured at its thinnest point) into consideration, and uses ten stages to classify damage. Radial rim width is compared to OD diameter at the axis in which the rim is thinnest. In cases that show no remaining rim, the circumferential extent of rim loss is measured in degrees. DDLS is a very detailed and accurate method, however, it tends to be time-consuming and not user friendly on a day- to-day basis, especially for non-experts.

A new classification method, known as optic disc damage staging system (ODDSS), was recently designed by Brusini and is currently under evaluation. ODDSS (Fig. 1) provides a clinical

classification of the OD (using three digits) based on OD size (small, medium, large), severity of neural rim loss (six stages), and localization of the neural rim loss (four types). Preliminary results using this method seem quite promising, and show that ODDSS offers a high sensitivity and specificity compared to HRT II results (personal communication).

OD color photographs are useful and objective in reporting morphologic appearance; however, assessing progression strictly on photos can become a difficult, debatable, and tedious task. Stereophotography offers a better representation of the OD, and is still taken as the gold standard in most international clinical studies. There are, however, several drawbacks in stereophotography, which include: the need of relatively expensive equipment; image quality being dependent on operator expertise and good patient collaboration; a method that is not in widespread use; a subjective interpretation based on qualitative data; high interobserver variability; and, the need of stereo-visors to assess images.

High-tech instruments offer objective, standardized, and reproducible ways to assess and continuously monitor glaucomatous structural damage.

Heidelberg retina tomograph (HRT) is a confocal scanning laser ophthalmoscope that takes a series of OD optical scans along the Z-axis to generate a topographic image. Numerous studies have shown that HRT measurements are reliable and reproducible (Chauhan et al., 1994; Miglior et al., 2002; Owen et al., 2006), and that the instrument is clinically useful in detecting OD structural changes over time. Built-in statistical software programs have currently been designed to specifically assess OD damage progression. Change probability maps, like the one proposed by Chauhan et al. (2000), can be useful for indicating areas of change. In brief, a baseline examination (mean of three scans) is formed based on a reduced number of points (pixels), grouped in clusters of 16 pixels (superpixels). This approach allows measurement variability and confidence limits to be minimized, thus providing a robust statistical analysis of variation. Areas having statistically significant OD changes (more than 20 superpixels) compared to baseline and confirmed

on two subsequent examinations are highlighted in red and green, indicating a reduction or increase in retinal height, respectively (Fig. 2). The topographic change analysis (TCA) is slightly different, in that a color gradient map is used to represent the magnitude of change (Chauhan, 2005).

Progression with HRT can also be assessed with a trend analysis (Trend Report) of several HRT stereometric parameters over time (Fig. 3).

The parameter analysis can be done for the following sectors: superior; inferior; superior temporal; inferior temporal; entire upper; and, entire lower disc. Although calculations are not based on regression analysis, progression can empirically be defined as a downward sloping trend in at least three consecutive examinations. Cluster defect trend analysis, in which area and volume of a selected cluster are plotted over time, is now available in the most recent version of HRT 3 (TCA Overview).

The analysis of RNFL changes over time has proven to be another interesting and promising

Fig. 2. HRT 2 change probability map. (See Color Plate 6.2 in color plate section.)

method in assessing glaucomatous progression. High-tech instruments have been of great advantage, considering that red free RNFL photographs are difficult to perform and clinical assessments is highly subjective and variable. The first instrument that offered clinically useful in vivo RNFL analysis was the nerve fiber analyzer (NFA); a scanning laser polarimeter. The principle used in this technology is based on the assumption that the RNFL is birefringent and can cause a change in polarization (called retardation) of an illuminating laser beam. The retardation can be quantified and

has shown to be linearly related to RNFL thickness (Weinreb et al., 1990; Dreher and Reiter, 1992). The NFA, along with the second and third generation versions (NFA II and GDx), have shown to provide reproducible RNFL measurements. Studies have shown, however, that this technology is heavily influenced by cornea polarization, which is not individually accounted for by the built-in fixed compensator. The clinical usefulness of these past versions is limited, considering that a general corneal compensation is assumed for all subjects as opposed to an individually

determined compensation. The most recent GDx version has been equipped with a variable corneal compensator (GDx VCC), which has greatly improved the diagnostic capabilities of this technology (Greenfield et al., 2000; Zhou and Weinreb,

2002; Reus and Lemij, 2004). Corneal compensation with this technology is determined on an individual basis, and no longer assumed to be the same fixed compensation for all. Moreover, recent studies have shown that the new enhanced corneal

compensator (GDx ECC) software seems to reduce the incidence of the atypical birefringence patterns (ABP), which can have a confounding effect in interpreting results in up to 50% of

glaucoma patients (Bagga et al., 2005; Toth and Hollo, 2005).

With regards to the progression analysis, GDx VCC currently provides an advanced serial

analysis, which includes a trend analysis of the main parameters and a series of maps in which changes over time (compared to baseline) are color-coded (Fig. 4). This type of progression analysis promises to be clinically useful, however, further long-term prospective multicenter studies are needed to confirm this.

New statistical software called GDx VCC guided progression analysis (GPA) is currently available, which uses the following three different approaches for depicting progression (Fig. 5): (1)

Image Progression Map, in which focal RNFL progressing defects are topographically shown in red (Fig. 5A); (2) TSNIT Progression Graph, which considers change in the 64 sectors in a ring area around the OD and highlights significant change (at least three adjacent segments) compared to the baseline in red (Fig. 5B); and, (3) Parameters Progression Chart that plots the rate of progression for global parameters (TSNIT average, superior average, and inferior average) using regression graphs to show diffuse defect