Diagnostic Performance of a Novel Three-Dimensional Neuroretinal Rim Parameter for Glaucoma Using High-Density Volume Scans

• Published in: American journal of ophthalmology Vol. 169; pp. 168 - 178
• Main Authors: Shieh, Eric, Lee, Ramon, Que, Christian, Srinivasan, Vivek, Guo, Rong, DeLuna, Regina, Pandit, Sumir, Simavli, Huseyin, Seevaratnam, Rajini, Tsikata, Edem, de Boer, Johannes, Chen, Teresa C.
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01-09-2016; Elsevier Limited
• Subjects: Aged; Area Under Curve; Confidence intervals; Diabetic retinopathy; False Positive Reactions; Female; Glaucoma; Glaucoma, Open-Angle - diagnosis; Healthy Volunteers; Humans; Imaging, Three-Dimensional; Intraocular Pressure; Male; Medical diagnosis; Medical imaging; Middle Aged; Nerve Fibers - pathology; Ophthalmology; Optic Disk - pathology; Optic nerve; Optic Nerve Diseases - diagnosis; Optics; Predictive Value of Tests; Retina; Retinal Ganglion Cells - pathology; ROC Curve; Sensitivity and Specificity; Tomography; Tomography, Optical Coherence - methods; Tonometry, Ocular; United States > US; Massachusetts
• ISSN: 0002-9394; 1879-1891
• DOI: 10.1016/j.ajo.2016.06.028

To evaluate the diagnostic performance of a 3-dimensional (3D) neuroretinal rim parameter, the minimum distance band (MDB), using optical coherence tomography (OCT) high-density volume scans for open-angle glaucoma.
Reliability analysis.
setting: Institutional. study population: Total of 163 patients (105 glaucoma and 58 healthy subjects). observation procedures: One eye of each patient was included. MDB and retinal nerve fiber layer (RNFL) thickness values were determined for 4 quadrants and 4 sectors using a spectral-domain OCT device. main outcome measures: Area under the receiver operating characteristic curve (AUROC) values, sensitivities, specificities, and positive and negative predictive values.
The best AUROC values of 3D MDB thickness for glaucoma and early glaucoma were for the overall globe (0.969, 0.952), followed by the inferior quadrant (0.966, 0.949) and inferior-temporal sector (0.966, 0.944), and then followed by the superior-temporal sector (0.964, 0.932) and superior quadrant (0.962, 0.924). All 3D MDB thickness AUROC values were higher than those of 2D RNFL thickness. Pairwise comparisons showed that the diagnostic performance of the 3D MDB parameter was significantly better than 2D RNFL thickness only for the nasal quadrant and inferior-nasal and superior-nasal sectors (P = .023–.049). Combining 3D MDB with 2D RNFL parameters provided significantly better diagnostic performance (AUROC 0.984) than most single MDB parameters and all single RNFL parameters.
Compared with the 2D RNFL thickness parameter, the 3D MDB neuroretinal rim thickness parameter had uniformly equal or better diagnostic performance for glaucoma in all regions and was significantly better in the nasal region.