Intraobserver And Interobserver Variability In Choroidal Thickness Measurements By Two Different Techniques Using Heidelberg Spectralis Sd-oct In Diabetic And Non Diabetic Patients

Abstract

Purpose: : To assess intraobserver and interobserver variability in choroidal thickness measurements by two different techniques using Heidelberg Spectralis SD-OCT

Methods: : Prospective longitudinal study. Fifty-two eyes from 26 subjects (14 diabetics), underwent conventional raster scanning within a 5 X 30-degree area centered at the fovea, with 100 scans averaged using SD-OCT. Enhanced Depth Imaging (EDI) OCT inverted images technique were also obtained. The choroid was measured from the outer border of the retinal pigment epithelium to the inner sclera border at 500 µm intervals of a horizontal section from 2.5 mm nasal to the fovea to 2.5 mm temporal to the fovea. All measurements were performed three times by 2 independent observers. Intraclass correlation coefficient (ICC) was used to assess the level of intraobserver agreement. For assesing inter-observer variability the Limit of agreement (LoA95%) Bland and Altman, ICC and Pearson correlation coefficients were used

Results: : Choroidal thickness measurements had a mean ICC of 0.756 for observer 1 and 0.857 for observer 2. The mean 95 % Limit of Agreement (LoA) width was shorter for EDI (-69 to 77 µm) than for conventional technique (-120 to 94 µm). Choroidal thickness measurements had a stronger inter-observer correlation for EDI (r= 0.80) than for the conventional technique (r=0.35)

Conclusions: : The EDI technique demonstrated lower variability and a stronger inter-observer correlation in choroidal thickness measurement than conventional technique

Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • choroid • imaging/image analysis: non-clinical