Abstract
Abstract: :
Purpose: For topographic optic nerve head (ONH) analysis with a scanning laser ophthalmoscope (SLO), it is necessary to define the ONH with a contour line (CL). We investigated the reliability of drawing CLs freehand, with the aid of stereoscopic disk photographs (SDP) and with monoscopic disk photographs (MDP). Methods: We recruited 45 patients from a glaucoma clinic. Each had SDPs taken and their ONH imaged using the Heidelberg Retinal Tomograph (HRT I). Two experienced observers drew ONH CLs. CLs were drawn freehand, with a SDP, and with one image of the SDP covered to produce a MDP, in random order on 6 separate days. A subgroup of 23 patients had CLs redrawn using MDPs. Six ONH parameters generated by the SLO were selected for analysis (disk area, cup disk ratio, cup volume, rim volume, cup shape measure and height variation in contour). We analysed three comparisons for the 6 ONH parameters after CL drawing: (1) intraobserver reliability for CLs drawn freehand versus with SDPs and with MDPs versus with SDPs , (2) interobserver reliability for CLs drawn freehand, with SDPs and with MDPs, and (3) intraobserver reliability for CLs drawn on 2 occasions with MDPs. Reliability was assessed using intraclass coefficients (ICC) and method error variation coefficients (MEVC). An ICC of >0.81 was regarded as ‘almost perfect agreement’. Results: ICCs for comparison (1) ranged from 0.94 to 0.99 for observer A and from 0.93 to 0.99 for observer B. MEVCs ranged from 3.18% to 9.00% for observer A and from 3.77% to 6.01% for observer B. For comparison (2), ICCs ranged from 0.89 to 0.99 for CLs drawn freehand, 0.86 to 0.99 for SDP and 0.81 to 0.99 for MDP. MEVCs ranged from 6.82% to 13.5% for freehand, 7.55% to 12.45% for SDP and 6.84% to 15.23% for MDP. When CLs were drawn freehand and with MDPs, mean disk area for observer A was significantly smaller than that for observer B (Bonferroni corrected t-test). This result was heavily biased by disagreements for ONHs which were tilted and/or had peripapillary atrophy (PPA). For comparison (3), ICC ranged from 0.93 to 0.99 for observer A and 0.97 to 0.99 for observer B. MEVC ranged from 3.80% to 11.85% for observer A and from 2.44% to 8.82% for observer B. Rim volume was the most variable parameter. Conclusions: A MDP is an acceptable alternative to a SDP as an aid to drawing CLs around the ONH for SLO analysis. Observers with impaired stereopsis may be able to draw reliable CLs. Care should be exercised when the ONH is tilted or there is PPA.
Keywords: optic disc • imaging/image analysis: clinical • imaging methods (CT, FA, ICG, MRI, OCT, RTA, S