Abstract
Purpose: :
To apply a novel event analysis–based progression technique to a longitudinal HRT data set. Agreement with an event analysis–based visual field (VF) technique is also assessed.
Methods: :
Interobserver–intervisit global rim area (RA) repeatability coefficients (RC) have previously been estimated (Strouthidis NG – BJO 2005). The RC values are stratified according to image quality (based on Mean Pixel Height Standard Deviation – MPHSD) – 0.08 mm2 for good quality images (MPHSD < 21), 0.13 mm2 for medium quality images (MPHSD 21 – 35) and 0.19 mm2 for poor quality images (MPHSD > 35). Progression is suggested if follow–up minus baseline global RA is > RC value for that image quality (taken as the mean MPHSD of the 2 images) and is confirmed if this is observed in a further 1 of 2 consecutive follow–up images. The technique was applied to data acquired from 198 ocular hypertensive (OHT) and 21 control subjects examined prospectively (1993–2001) with regular VF and HRT testing. Specificity was estimated as the proportion of progressing control subjects and as the proportion of significantly improving subjects. Within the OHT group, a comparison was made between HRT progression and VF progression based on a change of Advanced Glaucoma Intervention Study VF score from 0 to > 1 in 3 consecutive fields.
Results: :
Specificity was estimated at 95.2 – 98.6 % (3 improving subjects, 1 progressing control). Within the OHT group, 16 (8.1%) subjects were identified as progressing by both RA and VF; a further 12 (6.1%) progressed by RA alone and 27 (13.6%) by VF alone.
Conclusions: :
The novel event analysis performs similarly to other published HRT progression techniques, in both specificity and agreement with functional progression. It represents a simple technique that may prove useful where only minimal longitudinal data are available and minimizes false positive flagging of progression by taking image quality into account.
Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • optic disc • visual fields