Combining measurements from three anatomical areas for glaucoma diagnosis using Fourier-domain optical coherence tomography

Aims To improve the diagnostic power for glaucoma by combining measurements of peripapillary nerve fibre layer (NFL), macular ganglion cell complex (GCC) and disc variables obtained with Fourier-domain optical coherence tomography (FD-OCT) into the glaucoma structural diagnostic index (GSDI). Methods In this observational, cross-sectional study of subjects from the Advanced Imaging of Glaucoma Study, GCC and NFL of healthy and perimetrical glaucoma subjects from four major academic referral centres of the Advanced Imaging of Glaucoma Study were mapped with the RTVue FD-OCT. Global loss volume and focal loss volume parameters were defined using NFL and GCC normative reference maps. Optimal weights for NFL, GCC and disc variables were combined using multivariate logistic regression to build the GSDI. Glaucoma severity was classified using the Enhanced Glaucoma Staging System (GSS2). Diagnostic accuracy was assessed by sensitivity, specificity and the area under the receiver operator characteristic curve (AUC). Results We analysed 118 normal eyes of 60 subjects, 236 matched eyes of 166 subjects with perimetrical glaucoma, and 105 eyes from a healthy reference group of 61 subjects. The GSDI included composite overall thickness and focal loss volume with weighted NFL and GCC components, as well as the vertical cup-to-disc ratio. The AUC of 0.922 from leave-one-out cross validation was better than the best component variable alone ( p=0.047). The partial AUC in the high specificity region was also better ( p=0.01), with a sensitivity of 69% at 99% specificity, and a sensitivity of 80.3% at 95% specificity. For GSS2 stages 3-5 the sensitivity was 98% at 99% specificity, and 100% at 95% specificity. Conclusions Combining structural measurements of GCC, NFL and disc variables from FD-OCT created a GSDI that improved the accuracy for glaucoma diagnosis.