Elsevier

Ophthalmology

Volume 121, Issue 10, October 2014, Pages 2023-2027
Ophthalmology

Original article
Visual Field Progression in Glaucoma: What Is the Specificity of the Guided Progression Analysis?

Presented at: the Association for Research in Vision and Ophthalmology meeting, May 1–5, 2011, Fort Lauderdale, Florida.
https://doi.org/10.1016/j.ophtha.2014.04.015Get rights and content

Purpose

To estimate the specificity of the Guided Progression Analysis (GPA) (Carl Zeiss Meditec, Dublin, CA) in individual patients with glaucoma.

Design

Observational cohort study.

Participants

Thirty patients with open-angle glaucoma.

Methods

In 30 patients with open-angle glaucoma, 1 eye (median mean deviation [MD], −2.5 decibels [dB]; interquartile range, −4.4 to −1.3 dB) was tested 12 times over 3 months (Humphrey Field Analyzer, Carl Zeiss Meditec; SITA Standard, 24-2). “Possible progression” and “likely progression” were determined with the GPA. These analyses were repeated after the order of the tests had been randomly rearranged (1000 unique permutations).

Main Outcome Measures

Rate of false-positive alerts of “possible progression” and “likely progression” with the GPA.

Results

On average, the specificity of the GPA “likely progression” alert was high—for the entire sample, the mean rate of false-positive alerts after 10 follow-up tests was 2.6%. With “possible progression,” the specificity was considerably lower (false-positive rate, 18.5%). Most important, the cumulative rate of false-positive alerts varied substantially among patients, from <1% to 80% with “possible progression” and from <0.1% to 20% with “likely progression.” Factors associated with false-positive alerts were visual field variability (standard deviation of MD, Spearman's rho = 0.41, P<0.001) and the reliability indices (proportion of false-positive and false-negative responses, fixation losses, rho>0.31, P≤0.10).

Conclusions

On average, progression criteria currently used in the GPA have high specificity, but some patients are more likely to show false-positive alerts than others. This is a natural consequence of population-based change criteria and may not matter in clinical trials and studies in which large groups of patients are compared. However, it must be considered when the GPA is used in clinical practice where specificity needs to be controlled for individual patients.

Section snippets

Patients

Thirty patients were recruited from the glaucoma clinics at the Queen Elizabeth Health Sciences Centre in Halifax, Nova Scotia. Inclusion criteria were a clinical diagnosis of open-angle glaucoma, a mean deviation (MD) >−15.0 decibels (dB) in at least 1 eye, the absence of ocular or systemic pathology known to reduce visual field sensitivity, and the ability and willingness to participate for 12 consecutive weekly sessions. All patients were experienced with static automated perimetry and had

Results

The median age of the patients was 69.1 years (interquartile range, 64.4–70.7 years). Patients had early to moderate visual field damage (median MD, −2.5 dB; interquartile range, −4.4 to −1.3 dB) as illustrated in Figure 1 (available at www.aaojournal.org). All patients were experienced test-takers, and there were no clinically important learning or practice effects—the mean MD of the 30 patients changed by <0.1 dB between the first and last tests (Fig 2, available at www.aaojournal.org).

Discussion

The aim of our study was to investigate the specificity of the Glaucoma Progression Analysis, that is, the likelihood of encountering a “possible progression” or “likely progression” alert in a series of visual fields in which no meaningful change has taken place. Stable series were established by testing patients frequently over a short period during which disease progression was unlikely, such that any GPA progression alert could be regarded as a false-positive event. Under the assumption

Acknowledgments

The authors thank Gary Lee and Mary Durbin of Carl Zeiss Meditec for generating the GPA output for the permuted visual fields and Ivan Marin Franch of the Open Perimetry Initiative for critical comments and incorporating the dataset in the R visualFields package.

References (27)

  • J. Katz

    A comparison of the pattern- and total deviation-based Glaucoma Change Probability programs

    Invest Ophthalmol Vis Sci

    (2000)
  • A. Heijl et al.

    Measuring visual field progression in the Early Manifest Glaucoma Trial

    Acta Ophthalmol Scand

    (2003)
  • P.H. Artes et al.

    Visual field progression in glaucoma: total versus pattern deviation analyses

    Invest Ophthalmol Vis Sci

    (2005)
  • Cited by (52)

    • Estimating Global Visual Field Indices in Glaucoma by Combining Macula and Optic Disc OCT Scans Using 3-Dimensional Convolutional Neural Networks

      2021, Ophthalmology Glaucoma
      Citation Excerpt :

      The median values for MAC, ONH, and MAC + CONH were 1.1, 0.7, and 0.8, respectively, for VFI, and 0.42 dB, 0.29 dB, and 0.30 dB, respectively, for MD. The variability was smaller than the variability of visual field tests.28 Like accuracy, estimation variability was also dependent on the severity of the disease (Fig 8).

    View all citing articles on Scopus

    Supplemental material is available at www.aaojournal.org.

    Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article.

    Supported by the Glaucoma Research Foundation, United States (P.H.A. and D.P.C.). Grant Number MOP-11357, Canadian Institutes of Health Research (B.C.C.).

    View full text