spotlight on technology & technique
Nailing Down Progression
An upgrade for the Humphrey Field Analyzer separates developing glaucoma from variability and false positives.
By Christopher Kent, Senior Associate Editor

Analyzing visual fields has always been a tough job. You have to compare multiple exam results and attempt to decide whether a complex, shifting field of information contains signs of glaucoma progression. Unfortunately, test results can be affected by numerous variables, including the individual's test-taking experience, the test strategy used (i.e. SITA [Swedish Interactive Thresholding Algorithm] or full threshold), eccentricity in the field, and factors such as developing cataracts. Which changes are meaningful and which are just background noise? And how much change constitutes true progression?

A new upgrade for the popular Humphrey Field Analyzer II and II-i instruments can help. Humphrey GPA (Glaucoma Progression Analysis) software from Carl Zeiss Meditec uses statistical analysis to identify or rule out the presence of visual field loss in a series of visual fields. Instead of having to rely solely on comparing test results by hand or interpreting symbols to detect progression, the GPA software (which was developed by Drs. Anders Heijl and Boel Bengtsson, who conducted the Early Manifest Glaucoma Trial) provides a "plain language" interpretation of the results.

How It Works

The GPA software uses a minimum of three exams (two baseline exams and at least one follow-up) to determine whether visual field artifacts represent progression. (Baseline exams can be either full threshold or SITA; all follow-up exams must use SITA or SITA fast.) The software averages the thresholds of the two baseline exams and then compares the current (follow-up) exam to that data. Once five or more exams have been done, the current exam is compared to the baseline data and the two most recent follow-up exams. The software also automatically identifies tests with questionable reliability indices as a result of false positives and/or fixation loss, and warns the doctor before including them in the baseline.

The GPA program defines progression as repeatable change. When statistically significant degradation appears in three or more points on two consecutive follow- up exams, the software alerts you that "Possible Progression" has occurred. If the change is seen over three exams, the software reports "Likely Progression."

The GPA software also estimates rate of progression by analyzing the mean deviation over time. (Other more sophisticated means of rate analysis are currently under development.) The slope of the change in mean deviation from exam to exam is plotted on the baseline exam printout.

	A Single Field Analysis includes GPA results in the box on the right. Note the warning of "Likely Progression," the list of exams included, and the key to understanding GPA symbols.

Speaking in Triangles

The device uses triangle symbols to indicate progression at a 95% significance level for a given point in the field. In other words, it indicates that that point has worsened by an amount exceeding the variability expected in 95% of glaucoma patients with similar visual field status -- patients with the same mean threshold deviation from normal. (For an example, see the sample printout on page 64.)

► clear triangle indicates that this has happened once at a particular point in the field.

► half-darkened triangle indicates that the change from baseline at that point has occurred in two consecutive exams.

► fully darkened triangle indicates that change has occurred in three consecutive exams.

Three half-darkened triangles automatically trigger a warning of "Possible Progression;" three fully darkened triangles trigger a warning of "Likely Progression."

Quantifying Variability

A key factor in identifying progression is being able to tell which changes represent normal testing variability or factors such as cataract development. The GPA uses data from a 16-center glaucoma trial to quantify expected variability in visual field test-taking.

Hundreds of glaucoma patients at all stages of the disease were tested, worldwide. Each patient was examined four times in the space of a month, undergoing three tests (SITA, SITA Fast and full threshold) at each visit. Variations between visits were recorded; new mathematical models were developed to calculate expected variability, and these were incorporated into the GPA software. (The resulting significance limits have been validated manually.)

To differentiate between glaucomatous damage and defects that might occur from either pupil factors or developing cataracts, GPA uses pattern deviation plots from SITA standard and SITA Fast exams.

Features and Benefits

Like the Humphrey Field Ana-lyzer, the new software is designed to be easy to use:

► When you first use the GPA software, it lets you compare recent SITA exams with baseline full threshold exams previously stored in the Field Analyzer, so your existing database can be used.

► Using GPA requires no special training or instruction.

► Setup is automated. A simple graphic interface guides the user through the process of selecting baseline tests and choosing data presentation format. The instrument remembers settings and data for each patient, automatically adding new data to the appropriate file.

► Baseline test results are presented in the familiar STATPAC format. The overview display/printout after a follow-up exam shows the following statistics for the current exam and two most recent exams:

• the patient's visual field data in four formats: gray scale, pattern deviation, numerical (showing deviation from baseline in dB), and using the GPA triangle symbols to indicate the status of each problematic point
• reliability indices for the exam
• a "GPA Alert" (a plain language interpretation of the results -- either "Possible Progression" or "Likely Progression") if progression is detected.

Feedback from the Field

Howard Barnebey, M.D., a glaucoma specialist practicing in Seattle and clinical assistant professor at the University of Washington, has been using the GPA software for more than 6 months. "I think it's a significant step forward," he says. "It certainly makes looking for change a lot easier, and it's been reliable and consistent in my hands. It won't replace the doctor interpreting the field, but it does help me focus on changes that are consistent from one field to the next. It tells me where to look."

Dr. Barnebey notes that there are some limitations to what the software can do. "If a patient has advanced field loss, statistics won't always tell you whether there's been further progression. Nevertheless, it's difficult for a person to do what statistical analysis programs are able to do in the majority of cases.

"I'd guess that it will help an ophthalmologist do a better job nine out of ten times. And unlike some interpretive strategies that might be hard for a general ophthalmologist, the GPA is something that anybody with a Humphrey machine can incorporate into his or her practice.

"If somebody asked me whether it's a good investment, I'd say absolutely," he adds. "We need to take advantage of all the information we can get."

For More Information

To find out more about upgrading your HFA II or HFA II-i, call (800) 342-9821, or send an e-mail to info@meditec.zeiss.com to receive an information packet.