Abstract
Purpose: :
We examined 3 variants of multifocal pupillographic perimetry in glaucoma to investigate the effects of stimulus colour, and delivery rate upon diagnostic power and test-retest variability.
Methods: :
We tested both visual fields of 73 normal and 49 glaucoma age matched subjects. All subjects gave informed written consent. Testing was repeated about 2 weeks apart yielding 488 visual fields. All eyes were examined with Stratus OCT, HFA achromatic and Matrix 24-2 perimetry. Visual fields were classified by HFA mean defects: moderate: 6 to 12 dB, severe: >12 dB. Multifocal stimuli having 44 test regions/eye, extending to 30 deg eccentricity, were presented concurrently to both eyes using a prototype of the FDA cleared TrueField Analyser. Recording duration was 6 minutes, divided into 9 segments of 40 s. Data from blinks and fixation losses was automatically removed and a 40 s segment only needed to be repeated if more that 15% of the data from that segment was lost. The 3 stimulus protocols examined differed in terms of mean presentation intervals (MPI) of 1 or 4 s per region. The stimuli were either yellow on a dim yellow background, or green presented on a dim red background. The peak test luminances were 150 cd/m2, backgrounds 10 cd/m2. The 3 stimulus types employed a luminance balancing strategy designed to make pupil response amplitudes more even across the visual field [ARVO 2009, E-Abstract 5281]. Diagnostic performance was assessed by areas under ROC curves (AUCs). Test-retest variability was examined with correlation coefficients derived from Bland-Altman plots.
Results: :
The yellow MPI 4 s stimulus performed best diagnostically providing an AUC of 0.981 for severe fields. The median signal to noise ratio of the peak pupil response expressed as a t-statistic for that protocol and normal subjects was 2.88 ± 0.36 (mean ± SD), 2.96 ± 0.47 for patients. In glaucoma patients the correlation coefficient for the two repeats was 0.72.
Conclusions: :
Good diagnostic power, reproducibility and signal to noise ratios were obtained with a test duration of 3 min/eye. Also separate information on afferent and efferent defects is obtained at every point in the field, as is response delay.
Keywords: perimetry • pupil • visual fields