Identifying Progression With Visual Fields: Past and Present
The latest advance in perimetry is designed to improve accuracy and efficiency.
By Donald L. Budenz, M.D., M.P.H.
Today, we have available to us several tools for following glaucoma progression. We are able to detect both structural and functional changes, which helps us with our treatment decision-making. It is interesting to note that our first glaucoma progression tool was the Humphrey Field Analyzer (HFA). I will review past approaches to following glaucoma progression with visual fields and explain why the newest approach, Glaucoma Progression Analysis (GPA) with the HFA II-i, is more advanced and more efficient.
Monitoring Progression With Visual Fields: Yesterday
In 1986, shortly after the HFA was released, we began using mean deviation to detect glaucoma disease progression. We looked for decline in mean deviation, which was the global index showing severity of visual field loss. Mean deviation was plotted over time, and if the P value of the slope was less than 5%, it indicated that the visual field was deteriorating at a statistically significant rate. This, of course, did not tell us why the visual field loss was occurring. It could have been due to cataract or other factors.
At that time, we may have considered decline in the pattern or corrected pattern standard deviation. This would indicate localized progression, which is more likely to be glaucoma than cataract.
About 3 years later, the software called Glaucoma Change Probability became available. This gave us plots for observing progression. The plots took into consideration individual points over time rather than the entire field. This was not a time-dependent analysis, but what we call an event-based analysis. Two baseline fields were merged for comparison with sub-sequent fields. Colored-in black triangles on the plots represented potential problem points and were considered significant if confirmed on a follow-up test. If two or more adjacent points within or adjacent to an existing scotoma were worsening at a P value less than 5%, and a second or perhaps third visual field confirmed that, we would diagnose progression.
Figure 1. The Glaucoma Progression Analysis software for the HFA II/ HFA II-i plots visual field defects using the symbols shown here.
Monitoring Progression With Visual Fields: Today
Not much changed in our approach to monitoring glaucoma progression until 2005, when Glaucoma Progression Analysis (GPA) software for the HFA was introduced. Since that time, we have been using GPA in all of the established glaucoma patients in our practice as a tool for detecting disease progression.
GPA is similar to Glaucoma Change Probability, but it is much "smarter." Key features of GPA include:
■ It adjusts for cataract, a common problem among glaucoma patients, in progressively worsening visual fields.
■ It works with baseline full threshold fields, which many of us have from before we began using the HFA II, as well as baseline SITA fields. All follow-up fields, however, must be SITA.
■ It uses the criteria and statistical analysis from the Early Manifest Glaucoma Trial (EMGT) to identify progression at individual points, which makes it an event-based analysis, not a trend analysis. In the EMGT, a patient was considered to have progressed if three or more test points in the same location showed deterioration at the 5% level (P<0.05) on three consecutive field tests.
| Creating an Accurate GPA From Baseline |
|---|
| All SITA Standard and SITA Fast tests stored in your HFA II or HFA II-i perimeter database are eligible for use with the Humphrey Glaucoma Progression Analysis (GPA) software. Tests completed using the Full Threshold strategy may be used as baseline, but not as follow-up. A simple graphical interface guides you or your staff through the selection of baseline tests. The instrument remembers which tests are designated as baseline, making it unnecessary to repeat the process at each visit. The next version will further simplify the exam selection and will default to a one-page summary. Establishing a Baseline While GPA provides default selections to automate the baseline selection process, it is important that the tests chosen for baseline be reliable and representative of the patient's test-taking ability. The GPA software selects baseline tests by identifying the earliest two tests that are either full threshold or one of the SITA strategies. If at least three SITA tests have been taken — either all SITA Standard or all SITA Fast — the GPA software will choose the earliest two SITA tests to ensure the most accurate analysis. However, you always have the option to overrule the automated choices if necessary. For example, patients who are taking their first threshold perimetry test frequently produce unreliable results. If the field appears to be unreliable, or if the first field looks significantly worse than the second, you should designate the second and third examinations as baseline. Sometimes, even experienced patients have sessions that produce clearly unreliable results. These examinations should not be included, and you can manually remove them from the analysis. Choosing to 'Re-baseline' a Patient From time to time, you may want to review your baseline exam choices. If your baseline exams are full threshold, you should switch to SITA when three exams are available. You also may want to change your baseline choices when any of the following occur:
|
Figure 2. An illustration of Glaucoma Progression Analysis (GPA) with the HFA II. As shown above, progression analysis following a series of five visual field tests is displayed in the fourth column (enlarged on the left). Baseline fields were averaged. Several open white triangles appear, representing individual points that changed from baseline to one follow-up test. Repeat testing shows several half-darkened triangles, indicating possible progression. A second confirmatory test shows more than three points of change in the same location, indicating likely progression. As shown at right, it is also possible to generate a single-page summary printout. It displays the current field test results and an insert box containing the current GPA, the "possible progression" or "likely progression" message and the dates of baseline and previous follow-up tests.
In the development of GPA, several hundred normal subjects and glaucoma patients were tested with the full threshold SITA Standard and SITA Fast programs four times in 1 month. The amount of fluctuation detected at each individual point was determined and defined as noise. The result is a more precise statistical analysis of the probability of change over time.
Interpreting GPA Results
As we apply GPA to each patient in clinical practice, it compares current and previous perimetry results and uses a series of triangle symbols to plot the results of the comparison (Figure 1). When an individual point has deteriorated from baseline to one follow-up test at P<0.05, an open white triangle appears. If that point is confirmed on a second follow-up test, a half-darkened triangle appears. If that point is confirmed a third time, it is represented by a fully darkened triangle. When three or more points at the same location change in that manner on two consecutive tests, GPA reports "possible progression." When three or more points at the same location change in that manner on three consecutive tests, GPA reports "likely progression." (See Figure 2.)
It is also possible to generate a single-page summary printout. It displays the current field test results and an insert box containing the current GPA, the "possible progression" or "likely progression" message and the dates of baseline and previous follow-up tests.
Figure 3. By the end of this year, the Glaucoma Progression Analysis (GPA) software will generate the new graphic analysis shown here. It plots the visual field index, an improved metric of visual field loss, against the patient's age. It also provides a projection of the amount of additional field loss that would occur in 5 years if therapy is not changed and the patient's disease continues to progress. The GPA software and its new enhancements are based on algorithms pioneered and developed by Boel Bengtsson, Ph.D., publication pending, and in collaboration with Anders Heijl, M.D., Ph.D.
New GPA Enhancements
A new single-page GPA printout will be available this year, which is more of a trend-based analysis. It includes the two baseline visual field tests, the current field test and the current GPA plot. It also includes two additional pieces of information, a visual field index and its accompanying trend analysis graphic (Figure 3). The GPA software and its new enhancements are based on algorithms pioneered and developed by Boel Bengtsson, Ph.D., publication pending, and in collaboration with Anders Heijl, M.D., Ph.D.
The visual field index is an improved metric of visual field loss. It is a number between 0 and 100% with 100% being a perfect visual field. Central points in the field are weighted more heavily than those in the periphery. The points closer to fixation are weighted more heavily because they are the points we most want to preserve. The index calculation also reduces the contribution of cataract to the measurement of visual field loss.
The trend analysis graphic plots the visual field index on the Y axis against the patient's age on the X axis. It also projects the amount of additional field loss that would occur in 5 years if therapy is not changed and the patient's disease continues to progress. This is a convenient, easily understandable way for the physician to convey to the patient the status of his or her glaucoma.
Convenience: an Added Bonus
Overall, GPA for the HFA II/HFA II-i provides us with a practical tool for using visual field testing to evaluate glaucoma progression. The analysis is based on sound statistical principles, a robust database and a large clinical trial.
Furthermore, GPA saves time by simplifying results. It does not require that we circle numbers on the threshold printout as we have done in the past. With one click, it generates and displays the information we are looking for quickly and conveniently, making it easy to apply in every case of glaucoma that we follow. Some user input is sometimes required, for instance, if the doctor wants to identify or change baseline exams or eliminate certain exams because of artifacts or other problems.
Dr. Budenz is professor of ophthalmology, epidemiology and public health at Bascom Palmer Eye Institute, University of Miami Miller School of Medicine, Miami, Fla.
![]({"src":"/media/uoqfp2td/62007.jpg","focalPoint":null,"crops":[{"alias":"thumbnail","width":1110,"height":700,"coordinates":null}]})