Background/aims Optical coherence tomography (OCT) imaging of the optic nerve head minimum rim width (MRW) has recently been shown to sometimes contain components besides extended retinal nerve fibre layer (RNFL). This study was conducted to determine whether excluding these components, termed protruded retinal layers (PRLs), from MRW increases diagnostic accuracy for detecting glaucoma.

Methods In this cross-sectional study, we included 123 patients with glaucoma and 123 normal age-similar controls with OCT imaging of the optic nerve head (24 radial scans) and RNFL (circle scan). When present, PRLs were manually segmented, and adjusted MRW measurements were computed. We compared diagnostic accuracy of adjusted versus unadjusted MRW measurement. We also determined whether adjusted MRW correlates better with RNFL thickness compared with unadjusted MRW.

Results The median (IQR) visual field mean deviation of patients and controls was −4.4 (−10.3 to −2.1) dB and 0.0 (−0.6 to 0.8) dB, respectively. In the 5904 individual B-scans, PRLs were identified less frequently in patients (448, 7.6%) compared with controls (728, 12.3%; p<0.01) and were present most frequently in the temporal sector of both groups. Areas under the receiver operating characteristic curves and sensitivity values at 95% specificity indicated that PRL adjustment did not improve diagnostic accuracy of MRW, globally or temporally. Furthermore, adjusting MRW for PRL did not improve its correlation with RNFL thickness in either group.

Conclusion While layers besides the RNFL are sometimes included in OCT measurements of MRW, subtracting these layers does not impact clinical utility.