Assessment of Retinal Thickness with and without Manual Determination of the Foveal Center on SDOCT Volume Scans

Abstract

Purpose: : To compare retinal thickness measurements with and without manual determination of the foveal center point on spectral domain OCT (SDOCT)volume scans.

Methods: : SDOCT volume scans of 62 eyes with macular edema from retinal vascular disease (49 with diabetic retinopathy and 13 with retinal vein occlusion) and 64 eyes with neovascular AMD were retrospectively collected. Retinal thickness measurements were performed using computer-assisted manual grading software (3DOCTOR). Foveal center point (FCP) thickness as well as the mean foveal central subfield (FCS) thickness were calculated with and without manual determination of the foveal center and adjustment of the position of the ETDRS grid within the volume scan.

Results: : The mean absolute difference between FCP measurements with and without manual adjustment of the grid in eyes with retinal vascular disease - associated macular edema was 55 ± 75 µm (19 ± 35 %, max 251%, max 450 µm, median 29 µm). Only 10 out of 62 eyes (16%) demonstrated a difference of less than 10 µm. Eyes with neovascular AMD demonstrated a mean absolute difference of 39 ± 51 µm (21 ± 50 %, max 380%, max 276 µm, median 27 µm). Twenty out of 64 eyes (31%) demonstrated a discrepancy of < 10 µm. The difference between measurements with and without manual centration of the grid was statistically significant for all groups. The effect of manual centration was less pronounced for mean FCS thickness, as 28 out of 62 eyes (45%) with retinal vascular disease and 37 out of 64 eyes (58%) with neovascular AMD demonstrated discrepancies of ≤ 10 µm. The mean absolute difference for mean FCS thickness values was 34 ± 67 µm (8 ± 14 %) for retinal vascular disease and 13 ± 16 µm (5 ± 8%) for neovascular AMD.

Conclusions: : Our results indicate that decentration of SDOCT volume scans in eyes with retinal disease may result in erroneous central retinal thickness measurements. Manual grading of OCT scans may improve measurements in clinical trials not only by correcting segmentation errors but also by adjustment of the position of the foveal center point.

Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • retina