Purchase this article with an account.
Kevin K Ma, C Gustavo De Moraes, Abinaya Thenappan, Daiyan Xin, Ravivarn Jarukasetphon, Dana Blumberg, Jeffrey M Liebmann, Robert Ritch, Donald Hood; A comparison of circumpapillary retinal nerve fiber and macular ganglion cell measures in detecting early glaucoma. Invest. Ophthalmol. Vis. Sci. 2017;58(8):700.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Circumpapillary retinal nerve fiber layer (cpRNFL) summary metrics using optical coherence tomography (OCT) can miss macula damage in early glaucoma. Here we compare macular ganglion cell to cpRNFL thickness measures in eyes with early glaucoma.
102 eyes of 102 patients, which were previously categorized as glaucomatous (57 eyes) and healthy (45 eyes), had spectral domain OCT macula and disc cube scans (3D-OCT 2000, Topcon, Inc). Using the commercial segmentation algorithm, average retinal ganglion cell plus inner plexiform layer (RGC+) thickness was calculated for the central ±8° (macula), as well as the more (MVM) and less (LVM) vulnerable macular regions (Fig. 1). In addition, the average thickness was calculated for global cpRNFL, each of the 4 quadrants (superior, temporal, inferior, nasal), as well as the regions susceptible to damage (Fig. 2; superior (SVZ), superonasal (SnVZ) and inferior (IVZ) vulnerability zones). Logistic regression models were used to evaluate the performance of multiple, combined measures. The area under the ROC curve (AUC) was compared using the method of DeLong, and statistical classifications were optimized for accuracy.
Combined cpRNFL quadrants using logistic regression had a better optimized accuracy, 91.2%, than did global cpRNFL thickness, 87.3% (p=0.041). Macular measures had a lower accuracy; 82.4% for overall RGC+ thickness and 86.3% for combined MVM and LVM RGC+ thickness. On the other hand, combining the 4 cpRNFL quadrants and the 2 macula regions led to a slightly improved accuracy of 92.2%, as compared to cpRNFL quadrants (p=0.157). However, combining the 3 vulnerable cpRNFL regions gave the best accuracy, 97.1%, and an AUC greater than overall cpRNFL (p=0.018), overall macula RGC+ (p=0.002), combined MVM and LVM RGC+ (p=0.007), and combined cpRNFL quadrants (p=0.098), although the latter only reached borderline significance.
Combining macula and cpRNFL metrics can improve the accuracy of glaucoma diagnosis. However, using the cpRNFL regions most susceptible to glaucomatous damage proved more effective than global cpRNFL thickness or combined macular and cpRNFL measures. 1. Wang et al. TVST, 2015, 4, 6; 2. Hood et al. TVST, 2016, 5,4; 3. Hood et al. PRER, 2013, 32, 1; 4. Hood et al. IOVS, 2013, 54, 7338; 5. DeLong et al. Biometrics, 1988, 44, 837.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.
This PDF is available to Subscribers Only