Cluster Location, Size and Volume Could Predict Repeatable VF Defects

Visual field (VF) testing in suspected and early glaucoma is often characterized by non-repeatable defects, making it difficult to distinguish test variability from true functional loss. A common definition used for a glaucomatous VF defect is the identification of clusters of contiguous test points exhibiting statistically significant sensitivity loss—typically defined as one or more adjacent VF locations with a probability level of 5% or less on pattern deviation plots.  Knowledge of a likely non-repeatable cluster defect is useful as it may guide the clinician to extend the subsequent VF test interval. Conversely, a baseline VF that displays a likely repeatable cluster defect may benefit from repetition and closer follow-up, as a treatment modification may be warranted. These images from the study show the frequency of defects (%) by location on the pattern deviation probability (PDP) grid of the initial visual field test (A) vs. findings in defects that were repeatable over the three consecutive VF tests (B) in 197 eyes. There was a greater frequency of initial and repeatable defects in the superior and inferior arcuate distribution, the researchers reported. Photo: Tan JCK, et al. Br J Ophthalmol. July 14, 2024. Click image to enlarge. Cluster defects have demonstrated varying degrees of specificity in diagnosing glaucoma. Researchers at the Centre for Eye Health in Sydney wondered whether certain characteristics of the cluster, such as the eccentricity of locations on the grid involved and the size and depth of the defect, can be used to predict a subsequent repeatable defect. If so, these features may help differentiate between variability and true sensitivity loss. This distinction is especially relevant in glaucoma suspects or early glaucoma, since VFs in these patients are more likely to display cluster defects or non-repeatable changes.In their recent study, published in British Journal of Ophthalmology, the team determined that cluster size and volume were predictive of repeatable defects, especially when defining cluster defects by location of involvement.“Distinguishing between patients with a higher risk of repeatable defects may be helpful in risk stratification and resource allocation, especially given the lack of capacity in hospital eye services for patients with glaucoma,” the study authors wrote in their paper.This retrospective cohort study included 197 eyes of 103 patients with healthy (20.8% of eyes), suspect (27.4%) or early glaucoma (51.8%). The mean baseline mean deviation was -1.03, and the mean age was 62.9. Using the initial VF pattern deviation probability grid, they defined the number of clusters (≥one location of p<5%) and associated size (number of adjoining defect locations) and volume (sum of corresponding total deviation values) for each cluster stratified by the four probability levels (i.e., p<5%, p<2%, p<1% and p<0.5%).Only 26.9% of locations with a probability of 5% or below on the initial VF test were repeatable over two subsequent tests. Repeatable defects tended to fall within the superior arcuate or inferior arcuate distribution. Of note, central locations were not more repeatable than non-central locations. “This either indicated that center-involving clusters needed to be larger to achieve similar or better sensitivity and specificity compared with peripheral rim clusters or could simply reflect an inherent tendency for center-involving clusters to display larger size and volumes than peripheral rim clusters,” the researchers noted.The optimal thresholds for predicting a repeatable location within each cluster at 95% specificity based on initial cluster size were >six locations at p<5%, >four locations at p<2%, >three locations at p<1% and >two locations at p<0.5%. Defining cluster defects by involvement of central or peripheral rim locations improved the predictive value compared with the entire 24-2 grid.Examining cluster characteristics on the pointwise pattern deviation probability grid may be an objective method to help identify eyes that are more likely to display a subsequent repeatable defect. This can assist in the interpretation of VF outputs, which is often subjective. However, the researchers surmised that implementing such a system will likely require automation through a computer application, given the tedious nature of calculating these characteristics manually, which may be prone to human error or bias.Click here for the journal source. Tan JCK, Phu J, Bell K, et al. Prediction of repeatable glaucomatous visual field defects based on cluster characteristics. Br J Ophthalmol. July 14, 2024. [Epub ahead of print].