Predicting Goldmann III test size sensitivity values using standard Goldmann V results: implications in testing patients with glaucoma

Michael Kalloniatis


Purpose: We predicted Goldmann III (GIII) dB values using GV results and compare predicted GIII to actual GIII dB results to assess if the more reliable GV can be a surrogate for GIII. As smaller Goldmann test sizes may show greater threshold elevation in glaucoma (Redmond et al IOVS 2010; Wall et al IOVS 2013), we tested patients with glaucoma to determine the accuracy of the GV to predict GIII values. Methods: We tested 60 normal subjects and an initial cohort of 6 glaucoma patients using the full threshold paradigm for GI-V using the Humphrey Visual Field Analyzer (HFA). We previously determined spatial summation characteristics – critical area (Ac) and the slope of partial summation (n2) – at all 30-2 test locations. As Ac is smaller than GIII at all test locations within the 30-2 grid, we calculated GIII dB values from GV values using n2. Calculated GIII values were compared to actual GIII values using Bland-Altman plots for normal and glaucoma patients. Results: In normal subjects, GIII predicted values were accurate (mean difference = 0.19 dB, SD = 0.86 dB) with limits of agreement within 3 dB at nearly all locations with highest variability in more peripheral test locations. Initial analysis of patients with glaucoma showed that unaffected locations displayed agreement similar to that of normal subjects. Locations with loss using GIII displayed greater differences from predicted GIII values. This disparity appeared to be eccentricity dependent with greater disparities in peripheral locations where Ac is close to the size of GIII. Conclusions: GV and GIII values are closely linked, confirming that the location specific n2 values can accurately predict threshold values within the partial summation region. However, the use of GV in early glaucoma led to marked disparities in predicted GIII loss, always under predicting true GIII loss. This is likely due to the increase in Ac in glaucoma, leading to a greater loss when using GIII compared to GV.


Year: 2016

Program Number: 160048

Author Affiliation: Centre for Eye Health & SOVS, UNSW

Co-Authors: Jack Phu, Barbara Zangerl, Sieu Khuu

Co-Author Affiliation: Centre for Eye Health & SOVS, UNSW, Centre for Eye Health & SOVS, UNSW, SOVS, UNSW

Room: 210 A-D