Abstract
Purpose: :
To compare the difference with or without reading center manual correction in automated optical coherence tomography (Stratus OCT) retinal center point thickness (CPT) measurements in patients with neovascular age-related macular degeneration (AMD).
Methods: :
1454 OCT scans from 6 visits of 440 subjects from a study of AMD were analyzed. Manual measurements were performed using digital calipers when the automated center point measurements were considered erroneous or if the CPT standard deviation (SD) was > 10% of the thickness value. An automated measurement error was defined as an automated measurement that differed from a manual measurement by more than 12%. Change in CPT from baseline to Month 12 was compared with or without manual correction. The association of center point standard deviation (SD) and center point thickness (CPT) with the automated measurement error was analyzed.
Results: :
Manual measurement was performed in 893 scans (61%). The mean difference between manually measured CPT and automated CPT was 41.3 microns ([32.2, 50.4], p<0.01). The automated CPT measurement was greater than the manual measurement when the retina was thinner. 713 of 893 scans (80%) had automated measurement error according to the definition above. The mean (SEM) reduction in CPT from baseline to Month 12 was 26.2 microns (13.6) and 26.0 microns (14.7) using automated CPT and manually corrected CPT, respectively. SD was only a modest predictor of an automated measurement error (OR=1.11 [1.09, 1.12]).
Conclusions: :
In age-related macular degeneration studies, the error rate of the automated CPT measurement with Stratus OCT is very high, consistent with past reports. Quality assessment and manual re-measurement of retinal thickness is important when change in retinal thickness is an outcome of interest.
Keywords: imaging/image analysis: clinical • clinical research methodology • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)