Bruch'smembrane opening minimum rimwidth. Correlation and diagnostic accuracy in comparison to peripapillary retinal nerve fiber layer thickness

BackgroundGlaucoma is one of the main causes of blindness in the Western hemisphere. Because the disease often painlessly progresses it remains unnoticed until major optic nerve head damage occurs in many cases. That is why new, more sensitive diagnostic methods are needed. Bruch's membrane opening minimum rim width (BMO-MRW), measured with the new glaucoma module premium edition (GMPE) was recently introduced as amore accurate tool to detect glaucomatous changes. The purpose of this study was to assess the correlation of SPECTRALIS (R) spectral domain optical coherence tomography (SD-OCT) and the GMPE anatomic positioning module (APS module) for retinal nerve fiber layer thickness (RNFLT) measurements. The second aim was to assess the diagnostic accuracy of BMO-MRW.MethodsProspective study of 41eyes (41patients) with glaucoma and 26eyes from 26healthy controls. Scans were obtained using SPECTRALIS (R) SD-OCT and RNFLT was measured with both modules and compared using Spearman's rank test. The BMO-MRW was assessed by GMPE. Sensitivity, specificity and area under receiver operating characteristics curves (AUROC) of each sector of the optic nerve were calculated and compared using the method of Delong etal.ResultsWe found apositive correlation (0.694-0.955, p< 0.0001) between RNFLT measurements by standard SD-OCT and all diameters of RNFLT of APS module within all sectors. The AUROC of RNFLT in standard SD-OCT was 0.693 for the inferior nasal sector (NI) and BMO-MRW was 0.85 inNI. The difference in AUROC was statistically significant (p= 0.0049). No other sector showed statistically significant differences.ConclusionThe RNFLT measurements of both modules showed apositive correlation and appear to be comparable. The BMO-MRW in one sector (NI) showed a significantly higher accuracy of measurement than standard RNFLT. All other sectors showed a comparable accuracy of measurement.