advertisement
Editors Selection IGR 11-1
Disc Evaluation: Topographic change analysis
Comment by Claude Burgoyne on:
22538 Performance of confocal scanning laser tomograph Topographic Change Analysis (TCA) for assessing glaucomatous progression, Bowd C; Balasubramanian M; Weinreb RN et al., Investigative Ophthalmology and Visual Science, 2009; 50: 691-701
In 2000, Chauhan and colleagues introduced a super-pixel (4 x 4 = 16 pixels) based strategy for Topographic Change Analysis (TCA) within longitudinal Confocal Scanning Laser Tomographic (CSLT) images of the optic nerve head and demonstrated encouraging sensitivity and specificity by requiring that a group of at least 20 contiguous pixels demonstrate change from baseline on at least three occasions. More recently, in order to better separate 'clinically important' from 'statistically significant' change, Artes and Chauhan (Artes PH, et al. IOVS 2006; 47: ARVO E-Abstract 4349) investigated TCA change in glaucomatous (n = 172) and healthy eyes (n = 60) followed for at least 8 years and explored new TCA change criteria that required variably sized cluster and depth requirements for the superpixels demonstrating change. Their results suggested that the ideal parameters for separating these groups were small clusters of contiguous pixels (between 1 and 2% of disc area) demonstrating shallow (20 to 50 micron) depth changes. In the present article, Bowd et al. (171), first established 0.90, 0.95 and 0.99 specificity cutoffs for n = 70 TCA parameters that were based on Artes work, in a new series of healthy patients from the Moorfields Eye Hospital that had been imaged on at least 4 occasions (n = 18 eyes cleverly permuted into 1000 'healthy' topographic series). They then applied these cutoffs to topographic series from the following groups of UCSD patients each of whom had been imaged at least 4 times over 4 years: n = 36 known glaucoma progression eyes (by stereo disc photo and/or visual field (VF) criteria); a subset of the 36 progressors (n = 20) that progressed by disc only; n = 21 healthy volunteers; and n = 210 glaucoma suspect patients who had demonstrated no evidence of progression by disc photos or VF testing. In general they found that sensitivity was greatest when the depth of defect was not considered and specificity was greatest when a depth defect greater than 50 micron was required. The parameter with the marginal best diagnostic accuracy overall (sensitivity .778/specificity .809) was cluster area within the optic disc with all significantly change pixels accepted (without a depth requirement) using the 0.90 specificity cutoff of 0.036 mm2. However, the authors noted that some of parameters that included depth, performed almost as well. Not unexpectedly, a substantial subset of the 210 glaucoma suspect patients who had not demonstrated change by stereo disc photo or VF criteria, achieve change by one or more criteria. The authors legitimately suggest that at least a portion of these patients are likely to represent true disease progression. This study is important for the clarity and rigor of its approach. Its limitations, as acknowledged by the authors, include the relatively small number of healthy and progressing eyes and its inability to address the question of age-related changes in aging healthy eyes. Additional, high-quality, multi-centered assessments of these and other TCA- (and non-TCA-) based progression strategies that potentially include rate of change criteria are urgently needed.
Comments
The comment section on the IGR website is restricted to WGA#One members only. Please log-in through your WGA#One account to continue.Log-in through WGA#One
