Evaluating a new disc staging scale for glaucomatous damage: the ability to detect change over time

PURPOSE. To assess the ability of Disc Damage Likelihood Scale (DDLS) in detecting glaucomatous changes in the optic nerve on sequential optic disc stereo-photographs as compared with three other different cup/disc grading systems (C/D). METHODS. Eighty-three pairs of stereo-photographs with a minimum of 5-year follow-up were examined by one group of three observers masked to chronological order by using their clinical expertise. Their conclusions were considered gold standard. Another group of three glaucoma experts examined the same masked pairs to evaluate glaucomatous changes using four different optic nerve grading systems: DDLS, vertical, horizontal, and maximal cup/disc. Ratings of these four methods in the second group were then compared with the gold standard. RESULTS. Baseline and follow-up photographs of 83 eyes were included. Mean follow-up was 6 years. The sensitivities for detecting glaucomatous nerve changes were 71.4%, 63.9%, 66.7%, and 47.2% for DDLS, vertical, horizontal, and maximal cup/disc, respectively. The specificity for finding nerve change was 70.9%, 51.1%, 40.4%, and 48.9%, respectively. The proportion of photographs identified as showing glaucomatous changes differed significantly between the DDLS and vertical C/D (VCD) (chi(2)=7.900, p=0.005), between the DDLS and horizontal C/D (HCD) (chi(2)=13.89, p=0.000), and between DDLS and maximum C/D (MCD) (chi(2)=4.192, p=0.041). The largest area (74.4%) under the receiver operator characteristic curve was obtained by using the DDLS. CONCLUSIONS. DDLS had higher sensitivity and specificity for detecting glaucomatous changes versus the other three C/Ds when considering sequential stereo-photographs. (Eur J Ophthalmol 2009; 19: 404-10)