Abstract
Purpose: :
To assess the ability of Topographic Change Analysis (TCA) of HRT-III to detect progression in glaucoma.
Methods: :
Prospective study. Seventy six eyes (36 glaucomas and 40 normal eyes) were followed during a mínimum of 36 months. All were examined with HRT3 (yearly) and Humphrey visual field (every 6-12 months). A minimum of 4 good quality HRT images were required for inclusion in the study. The results of TCA were compared to expert evaluation of fields and the results of Glaucoma Progression Analysis I (GPA I) to calculate sensitivity and specificity. The 90 and 95 percentiles of minimum repeatable cluster area (CAREA) and minimum repeatable cluster volume (CVOL) observed in normal eyes were considered as definition of progression in TCA. A cluster was defined as a group of 20 continuous superpixels with significant change compared to baseline.
Results: :
A cluster of a minimum of 20 superpixels with significant change was present and repeatable in 25% of normal eyes and 47.22% of Glaucoma eyes. Table I shows sensitivity and specificity of the two defined parameters CAREA (Cluster area) and CVOL (Cluster volume). Sensitivity and specificity of cluster area ranged from 28,6% to 50% and 65,5% to 86,7% respectively. Sensitivity and specificity of cluster volume were 57,1% to 66,7% and 72,4% to 73.3% respectively.
Conclusions: :
The implemented HRT definition of progression (cluster of 20 superpixels) identified progression in one forth of normal eyes. The parameters CAREA and CVOL, compared to visual field, were able to identify progression with a specificity of 65.5% to 86.7%, and sensitivity of 28.6% to 66.7%.
Keywords: optic disc • imaging/image analysis: clinical