A comparison of macular thickness measurements across three different spectral-domain optical coherence tomography (SD-OCT) machines

Shaun Ewe

Meeting:  2015 RANZCO


Date:      -

Session Title: Poster

Session Time:      -

Purpose: To evaluate the comparability and reliability of macular scans performed on three different SD-OCT machines.

Method: This prospective study recruited 17 eyes (11 patients) with diabetic macular oedema (DME); 26 eyes (17 patients) with neovascular age-related macular degeneration (nAMD), and 16 healthy eyes (8 patients) as controls; all with visual acuity _6/30. Two scan-types were performed twice for each SD-OCT machine: Dense Scan and 7-line raster (Spectralis, Heidelberg); Macular Cube 512×128 and 5-line raster (Cirrus, Carl-Zeiss); and 3D-Macular Scan and 5-line raster (Dri-OCT Triton, Topcon). Central macular thickness (CMT) measurements were obtained and compared between machines. Repeated scans were assessed for reliability and presence of artefacts (segmentation; fixation).† Statistical analyses performed include non-parametric and paired t-test.

Results: There were significant differences across average CMT measurements performed on all SD-OCT machines (p<0.0001) or when compared between groups (p<0.05); with largest mean difference seen between Spectralis and Triton scans (55.15±33.7µm). Repeated CMT measurements performed across all SD-OCT modalities in all groups were comparable with the exception of Spectralis SD-OCT in the nAMD group which was significantly different on repeated scans (p=0.038); Higher incidences of artefacts occurred in nAMD (84.6%; compared to DME (41.2%; p>0.05) and controls (0%; p=0.0005) with Spectralis SD-OCT.

Conclusion: Macular thickness measurements varied across the three SD-OCT machines. Therefore, care needs to be taken when comparing measurements from different machines. Reliability of macular scans performed across all three SD-OCT modalities were similar, but may be subject to artefacts in the presence of significant retinal pathology.