Progression detection is complex clinical challenge

Las Vegas-The detection of visual field progression is a critically important aspect of the management of glaucoma, yet it remains a major challenge for clinicians, said Angelo P. Tanna, MD, speaking here during a Glaucoma Subspecialty Day program at the annual meeting of the American Academy of Ophthalmology.

One of the key confounding factors in the diagnosis of visual field progression is fluctuation in the measurement of threshold sensitivity on repeat testing in eyes with stable glaucoma and even in normal eyes, said Dr. Tanna, director, glaucoma service, Feinberg School of Medicine, Northwestern University, Chicago. Factors that influence the magnitude of fluctuation in threshold sensitivity, he said, include the baseline severity of damage at the particular point in the visual field being retested, the location of that point in the visual field-the more peripheral the point, the more variable the reading-and the global severity of visual field loss.

"Because it is difficult for a clinician to know whether a given magnitude of deterioration in threshold sensitivity that is observed at a particular location in a follow-up visual field exceeds that which is expected to occur due to normal fluctuation, it may be helpful to use some type of software designed to help identify clinically significant deterioration in conjunction with clinical judgment," Dr. Tanna said. "It is important to realize, however, that there still is no gold-standard algorithm for the detection of visual field deterioration.

Visual field progression must be verified by repeat testing because clinical trials consistently have shown that what initially may appear to be clinically significant progression often is not verified on repeat testing, Dr. Tanna said.

"The variability and fluctuation that occurs in testing can lead to the erroneous diagnosis of progression when, in fact, true progression has not occurred," he added.

Three ways to evaluate

Dr. Tanna said that serial visual field tests can be evaluated for progression in one of three ways: subjective evaluation, trend-type analysis, and event-type analysis.

In subjective analysis, he said, a clinician uses purely subjective judgment, often relying on the gray-scale maps of serial visual fields to attempt to detect progression.

"Unfortunately, this may be the most commonly used method and also the least accurate," Dr. Tanna added.

Trend-type analysis, he said, involves the use of software that employs linear regression analysis to determine whether some global index value, such as the mean deviation or the pattern standard deviation, has statistically deteriorated over time.

"Pointwise linear regression analysis software called Progressor, which was developed at Moorfields Eye Hospital in London, is a powerful tool that can be used to determine whether threshold sensitivity has significantly deteriorated over time at each individual point in the visual field," Dr. Tanna said.

Event-type analysis involves the use of a predetermined amount of deterioration that must occur in some visual field measurement before a particular point in the visual field can be declared as having progressed, he said. Glaucoma change probability for full-threshold Hum-phrey visual fields and Glau- coma Progression Analysis (GPA) software for visual fields obtained with the Swedish interactive threshold algorithm (SITA) are commercially available examples, he added.

GPA software

GPA software, developed for use with a particular perimeter (Humphrey Field Analyzer, Carl Zeiss Meditec), uses a database that was developed using 768 patients with glaucoma who were experienced with perimetry and were enrolled in the Early Manifest Glaucoma Trial (EMGT).

"Each subject underwent visual field testing four times in 1 month," Dr. Tanna said, explaining that the variability in threshold sensitivity observed in these subjects constitutes the GPA database.

"GPA uses the average of first two visual field tests obtained on a patient as a baseline for comparison against follow-up tests," he said. "Although for use with SITA, GPA does allow the use of baseline visual field tests obtained using the full threshold algorithm."