Abstract
Purpose:
To create a novel system to detect the progression of glaucomatous visual field (GVF) damage by carrying out the binomial test on the result of point-wise analysis on visual field (VF).
Methods:
Ninety-nine eyes of 59 open angle glaucoma patients (53±12 years, mean±sd) who underwent more than 15 VFs (Humphrey Field Analyzer, SITA standard, 30-2), were retrospectively collected. 15 VFs (6.0±1.5 years), excluding the first VF, were used for the analysis. Using the total deviation (td) of each point on the 2nd to 16th VF (VF2-16), linear regression analysis was carried out. The numbers of the test points with p<0.025, 0.05, 0.075 and 0.1 were calculated to perform binomial test (one-side) and then the median value of the p values was calculated. A VF was defined as “progressive” if the median value was <0.025 (“point-wise method”). Similarly, the progression analysis was carried out using VF2-6 to VF2-15. Then these probabilities were calculated; both of VF2-16 and prior VFs were “progressive”: sensitivity, both of VF2-16 and prior VFs were “non-progressive”: sensitivity, and VF2-16 was “non-progressive” however prior VFs were “progressive”: false positive. Same probabilities were calculated using the conventional MD slope method.
Results:
The specificity and false positive rates with the point-wise method (42±24, 8±10%) outperformed MD slope (38±25, 18±13%, P<0.05), while the specificities were not significantly different (96±2 and 96±3% p=0.95).
Conclusions:
It was suggested that the point-wise method may be able to detect the progression of GVF earlier than the MD slope method.
Keywords: 758 visual fields