Purpose:
To determine the repeatability and reproducibility of retinal thickness measurements with a custom-built OCT Retinal Image Analysis software (OCTRIMA).
Methods:
Ten eyes (5 healthy subjects) underwent 3 scanning sessions during the same visit by 2 experienced examiners. Healthy subjects also received a fourth scanning session during a second visit. Repeatability and reproducibility per scan and for each of 9 ETDRS like-regions were calculated by their repeatability and reproducibility coefficients and intraclass correlation coefficients (ICCs).
Results:
Repeatability and reproducibility coefficients obtained for each layer per scan were:The highest ICCs were found in the pericentral ring for GCL+IPL, INL and ONL (intervisit variability test). The lowest ICC values (0.85% to 0.92%) in the macular region were obtained for the interobserver variability test. Coefficients of repeatability per layer for the data segmented twice by one of the examiners were less than 2.03% (0.30% to 2.03%) and 0.17% for the macular region. In the analysis for each ETDRS region, coefficients of repeatability and reproducibility were less than 14% (ranged from 4% to 14%, intraobserver), 17% (6% to 17%, interobserver), and 13% (7% to 13%, intervisit) in the pericentral region; and at worst it was less than 28% (8% to 28% , intraobserver), 32% (7% to 32%, interobserver), and 29% (5% to 29%, intervisit) in the peripheral regions for all the layers except the RPE.
Conclusions:
Our results indicate that RT measurements per scan using OCTRIMA are repeatable and reproducible. In addition, these measurements for each of the 9 ETDRS like-regions are more repeatable and reproducible in the pericentral than in the peripheral ring. Change of examiner or segmentation operator did not significantly affect the reproducibility of the measurements in healthy eyes.
Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • image processing • retina