Abstract
PURPOSE: To assess the reliability of manual choroidal thickness measurements by comparing different posterior boundary definitions of the choroidal-scleral junction on enhanced depth imaging optical coherence tomography (EDI-OCT). DESIGN: Reliability analysis. METHODS: Two graders marked the choroidal-scleral junction with segmentation software using different posterior boundaries: (1) the outer border of the choroidal vessel lumen, (2) the outer border of the choroid stroma, and (3) the inner border of the sclera, to measure the vascular choroidal thickness (VCT), stromal choroidal thickness (SCT), and total choroidal thickness (TCT), respectively. Measurements were taken at 0.5-mm intervals from 1.5 mm nasal to 1.5 mm temporal to the fovea, and averaged continuously across the central 3 mm of the macula. Intraclass correlation coefficient (ICC) and coefficient of reliability (CR) were compared to assess intergrader and intragrader reliability. RESULTS: Choroidal thickness measurements varied significantly with different posterior boundaries (P < .001 for all). Intergrader ICCs were greater for SCT (0.959-0.980) than for TCT (0.928-0.963) and VCT (0.750-0.869), even in eyes where choroidal-scleral junction visibility was <75%. Intergrader CRs were lower for SCT (41.40-62.31) than for TCT (61.13-74.24) or VCT (72.44-115.11). ICCs and CRs showed greater reliability for averaged VCT, SCT, or TCT measurements than at individual locations. Intragrader ICCs and CRs were comparable to intergrader values. CONCLUSIONS: Choroidal thickness measurements are more reproducible when measured to the border of the choroid stroma (SCT) than the vascular lumen (VCT) or sclera (TCT).