Agreement for Detecting Change Using HRT Topographic Change Analysis (TCA), Standard Automated Perimetry Glaucoma Progression Analysis (GPA) and Stereo-Photograph Assessment

Abstract

Purpose:: Recent advances in HRT software (version 3.0) include improved image registration that likely reduces baseline variability used to define change in the HRT TCA. This reduction in variability should increase sensitivity to detect significant change in follow-up visits. The purpose of this study was to assess the agreement between change detected using HRT TCA, standard automated perimetry (SAP) GPA and assessment of stereo-photographs.

Methods:: 163 eyes with ≥ 4 HRT examinations, ≥ 2 years HRT follow-up with SAP testing and stereo-photography within one year of first HRT exam were selected from the UCSD Hamilton Glaucoma Center Diagnostic Innovations in Glaucoma Study (DIGS) database. TCA analysis was conducted using all available images of acceptable quality from each eye. Significant change was defined as a ≥ 20 mega-pixel and/or ≥ 2% of disc area contiguous, significant, reduction in retinal height within the optic disc margin on ≥ 3 consecutive exams. GPA was performed on all eyes using the 2 visual fields (VFs) closest to the first HRT exam as baseline, and all available reliable subsequent VFs as follow-up. Significant change was defined based on Early Manifest Glaucoma Trial criterion of ≥ 3 changed (p<5%) test points repeatable on ≥ 3 consecutive tests. Stereo-photograph change was based on differences detected when comparing the photograph closest to the first HRT exam with the most recent photograph in a masked fashion (2 observers, 1 adjudicator).

Results:: HRT TCA detected change in 21/31 (68%) eyes that showed change by either SAP GPA (11/16), stereo-photograph assessment (20/23) or both (5/8). In addition, HRT detected change in 55/132 (42%) eyes that showed no change with either other technique. In most cases of agreed change between TCA and GPA, location of change (superior versus inferior hemi-retina/hemi-field) corresponded (10/11).

Conclusions:: HRT TCA detects more change over time than GPA or stereo-photograph assessment. These results might suggest increased sensitivity compared to SAP GPA and stereo-photograph assessment, however, increased follow-up with VFs and stereo-photography is necessary for confirmation. Poor agreement between TCA change detection and change detection using VFs and/or stereo-photography (overall agreement in 98/163 eyes, Kappa = 0.17) suggests that these techniques likely cannot be used interchangeably for glaucoma monitoring.

Keywords: imaging/image analysis: clinical • optic disc • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)