Two important factors in the clinical management of glaucoma patients are determining that progression is occurring and calculating the rate of glaucomatous progression. The functional (visual fields) and structural (optic disc and retinal nerve fiber layer [RNFL]) assessment of eyes with glaucoma and those with risk factors for the development of the disease is critical. These evaluations help the practitioner to determine whether treatment is appropriate or if more aggressive procedures should be initiated. Currently, every multicenter clinical trial in glaucoma has used different criteria for defining visual field progression1,2; the situation is similar for structural properties, and methods to determine the rate of progression are just now emerging.3,4
To date, the methods that have been used to determine progression have been clinical judgment,5 classification or glaucoma disease staging systems,1,6 event analysis,1 and trend analysis.1 Each has its advantages and disadvantages. These procedures demonstrate large differences in sensitivity, specificity, and time to detect progression, and they only agree with each other about 50% to 60% of the time.2 Developing a standardized method of determining glaucomatous progression and the rate of progression remains a challenge, and achieving a consensus among glaucoma practitioners on such a procedure remains hopeful. This article briefly reviews the commercially available methods currently used for determining progression and its rate for the structural and functional evaluation of glaucoma.
VISUAL FIELD PROGRESSION SOFTWARE
At present, two primary analysis procedures are used
for the determination of visual field progression: event
analysis and trend analysis. The Glaucoma Change
Probability program for the Humphrey Field Analyzer
(Carl Zeiss Meditec, Inc., Dublin, CA) compares follow-up
visual fields to the average sensitivity values for two baseline
visual fields. It then determines if pointwise changes
in visual field sensitivity are within or outside the 95%
confidence limits of variability for glaucoma patients
(based on four tests performed within a 1-month period
in a group of glaucoma patients).7 In this manner, the
user can determine whether individual test locations
have a sensitivity that has remained stable or undergone
progression.
Because more than 50 locations are tested during a visual field examination, several locations (5%) are likely to show differences in sensitivity that are outside of the expected variability from one examination to another. The Glaucoma Progression Analysis software, based on results obtained from the Early Manifest Glaucoma Trial (EMGT), required the confirmation of three or more locations outside of the 95% confidence limits that were confirmed on two (possible progression) or three (likely progression) successive follow-up visual field examinations.8 The system also employs the pattern deviation probability plots rather than total deviation plots to minimize the influence of cataract and other nonglaucomatous conditions that can produce a wide-spread or diffuse loss of sensitivity. These modifications improved the performance of this analysis, but it only indicated a change of sensitivity from baseline for one to three follow-up visual fields and did not take into account information from other interim tests that had been performed. Thus, change could be determined, but the magnitude and rate of change were not directly evident from this analysis.
Two procedures that provide this information are the Progressor analysis program9 and the visual field index (VFI).4 By performing a least squares linear regression of visual field sensitivity for individual visual field test locations over time (successive perimetric examinations), Progressor determines the intercept (baseline sensitivity estimate) and slope (rate of change in sensitivity). A color-coded graphical display indicates for each visit and visual field location whether the sensitivity has remained stable (slope not statistically significantly different from zero) or has a low-to-moderate suspicion of progression or improvement (slope that is statistically significantly different from zero at P < .05) or a high suspicion of progression or improvement (slope that is statistically significantly different from zero at P < .01). A test location is considered to have undergone significant progression if the rate of change is greater than -1.0 dB per year and the slope of sensitivity over time is negative and statistically significant at P < .01. The implementation of additional analytical techniques and spatial filtering procedures has also improved the performance of this analytical method.10-12 Progressor is most useful for determining localized glaucomatous visual field changes over time.
The VFI evaluates the rate of visual field progression over time.4 Unlike Progressor, the VFI is designed to indicate the overall general status of the glaucomatous visual field, ranging from normal (100%) to end-stage glaucoma (0%). The procedure requires a minimum of five follow-up visual fields, and it conducts a linear regression of the glaucoma progression index that is based on pattern deviation values (to minimize the influence of cataract progression). The VFI provides greater weight for central points than peripheral test locations, and it predicts visual field status over the next 5 years. Figures 1 and 2 demonstrate the value of this approach.
Figure 1 shows early glaucomatous visual field damage in a patient's right eye. Visual fields were obtained every 6 months for 6 years. There is a suggestion of progression when comparing two successive visual fields (eg, from visual field No. 2 to visual field No. 3), but the lower graphs indicate that the VFI remains essentially unchanged over the 6-year follow-up. Figure 2 presents the right eye of a patient with more advanced glaucomatous visual field loss over a 6-year period (testing every 6 months). When comparing two successive visual field examinations, it is difficult to determine whether the visual field loss has progressed, but the lower graphs indicate a steady loss of visual field sensitivity over time. Thus, the VFI can be a useful clinical tool for the management of glaucoma patients and for following the visual field over time.
OPTIC DISC PROGRESSION SOFTWARE
Similar to progression techniques for glaucomatous
visual field loss, the use of imaging devices to measure
the topography of the optic nerve and RNFL provide greater standardization, reliability, and consistency
than subjective clinical assessments of ophthalmoscopic
viewing or evaluation of optic disc and RNFL
photographs. Additionally, event analysis and trend
analysis are the techniques that are most commonly
used for evaluating the optic disc and RNFL.
Scanning Laser Tomography
The Heidelberg Retina Tomograph (Heidelberg
Engineering GmbH, Heildelberg, Germany) is a commonly
used confocal scanning laser tomography system
that provides a three-dimensional reconstruction
of the optic disc and RNFL. The Topographic Change
Analysis uses superpixels (4 X 4 arrays of 16 pixels) to
compare follow-up images with baseline measures to
determine whether they are within the normal 95%
confidence limits of variability or if they are statistically
significantly better (green color-coding) or worse
(red color-coding).13-15 The system also performs a
trend analysis (linear regression) of changes in key
parameters (rim area, rim volume, cup volume, cup
shape, mean RNFL thickness, mean height of contour,
mean contour elevation, contour line modulation
temporal, mean cup depth, mean height inside contour
line, and a combination of these parameters).13-15
Scanning Laser Polarimetry
This technique presents a scanning polarized laser
beam to the optic nerve and RNFL to measure the
retardation of reflected light produced by the birefringent
properties of the retinal nerve fibers. The primary
analytical procedure is Guided Progression Analysis,
which is an event analysis technique that compares the
retardation patterns of follow-up images to a baseline
image. Locations that are outside normal limits of variability
due to RNFL thinning are indicated as possible
progression (P < .05) and likely progression (P < .01).
This is done not only for the entire RNFL region but
also for the temporal, superior, nasal, inferior, and temporal
(TSNIT) nerve fiber bundle regions.16
Optical Coherence Tomography
Optical coherence tomography uses low-coherence
interferometry to produce reflective images that provide
differential signals from structures in the retina
and optic nerve. The use of image segmentation procedures
and feature analysis techniques permits optical
coherence tomography to identify different layers
of the retina, fluid, and other structural elements.
Linear regression techniques and event analysis
(Guided Progression Analysis) are used to measure
changes in the RNFL and optic disc over time.17
CONCLUSION
The procedures described in this article offer glaucoma
practitioners methods that are standardized, consistent,
easy to interpret, and less variable than individual
clinical judgments. These tools are of great value in the
clinical management of patients with glaucoma and
glaucoma suspects. Future research is needed to determine
the best method of following glaucomatous progression
and to achieve a consensus among glaucoma
practitioners.
Chris A. Johnson, PhD, is a professor in the Department of Ophthalmology and Visual Sciences at the University of Iowa Hospitals and Clinics in Iowa City. He acknowledged no financial interest in the products or companies mentioned herein. Dr. Johnson may be reached at (319) 356- 0384; chris-a-johnson@uiowa.edu.
