Comparison of Methods for Quantification of Kinetic Visual Field Areas

Abstract

Purpose : To compare three methods of digitally quantifying kinetic visual field (VF) areas. Kinetic perimetry is commonly used to monitor disease course and treatment efficacy, which necessitates accurate quantitative analysis of the VF area.

Methods : Octopus 900 semiautomated kinetic VFs (OCVF) from 7 patients with proliferative diabetic retinopathy were evaluated at 0 months (0M) before treatment with pan retinal photocoagulation and at 12M post treatment. OCVFs were mapped for targets V4e, III4e and I4e, and areas (deg²) were measured using the Octopus EyeSuite program. Scotoma areas were subtracted to obtain the resulting seeing VF area for each target. Areas of scanned OCVFs were measured using Photoshop (PVF, deg²) and with FieldDigitize (FD) software V4.20 (Johns Hopkins Technology Ventures, Baltimore, USA). FD corrects for distortions associated with planar charts and converts planar areas into retinal areas. FD area outputs are: retinal (FDR mm²), bowl: FDB deg² and FDB steradians (str), and chart (CHT deg² and in mm²). Six measures for each target were examined: OCVF, PVF, FDR, FDB, FDBstr and CHT. Spearman correlations (r) were used to test for linear associations among the 6 measures and paired t-tests to compare between visits. ANOVA was performed to compare planar areas OCVF, PVF, and CHT and to evaluate % change in measures from 0M to 12M.

Results : 7 eyes of 7 patients, mean age 46.7±SD 11.2 years were evaluated for the 6 measures. All measures were significantly correlated with one another for each target (r=0.99, P<0.0001) at both visits. Means were significantly lower at 12M for I4E (P<0.008) and for III4e (P<0.05) but not for V4e (P=0.45). Average planar areas OCVF, PVF, and CHT showed no significant difference (p>0.05) for any of the targets. The FDR area was significantly lower than CHT area for I4e (P<0.02), and also for III4e and V4e (P<0.0001). All measures showed a similar amount of change between visits for each target (mean: I4e: -1.7%± 0.7%; III4e: -12.5%± 1.7%; V4e: -2.9%± 0.9%). FDR % change was lower than the mean percent change of the chart areas. Coefficient of variation was comparable for all measures for each target.

Conclusions : OCVF, PVF and FD measures had similar variability. The strong agreement in chart areas by the 3 methods substantiates compatibility. The lower FDR areas compared to CHT areas advocate for the utility of retinal areas to correct for chart area distortions.

This is a 2020 ARVO Annual Meeting abstract.