Grader learning effect and reproducibility of Doppler Spectral-Domain Optical Coherence Tomography derived retinal blood flow measurements

PURPOSE: To investigate grader learning effect and to quantify intergrader reproducibility of Doppler Spectral-Domain Optical Coherence Tomography (SD-OCT) derived retinal blood flow measurements.

METHODS: Fifteen healthy young subjects (mean age 28.44; SD 3 years) underwent Doppler SD-OCT scans of one eye using the circumpapillary double circular scan protocol of the Optovue RTVue by one of two experienced operators. One trained (i.e. having undergone certification) and one novice (i.e. preliminary training comprising five standard practice data sets) individual then graded a standardized set of scans, consisting of 15 data sets (session 1) using custom Doppler Optical Coherence Tomography of Retinal Circulation (DOCTORC) software. One week later (session 2), the novice grader underwent further training by grading an additional 15 practice data sets and then both graders subsequently regraded the original 15 data sets.

RESULTS: Measurements achieved by a novice grader during session 1 showed a trend to be higher in terms of total retinal venous blood flow (TRBF) and also to be significantly (p = 0.03) higher for venous area, compared with a trained grader. Session 2 results were not significantly different for either grader. The mean TRBF for session 2 for the trained and novice grader was 45.29 ± 9.28 μl/min and 44.39 ± 7.36 μl/min, respectively. The coefficient of repeatability (COR) of session 2 TRBF values between the trained and novice grader was 8.09 μl/min.

CONCLUSIONS: There is a grader learning effect which impacts the venous area measurements. Reproducible and repeatable retinal blood flow measurements were achieved among trained graders using DOCTORC software.