Purpose:
To evaluate the reproducibility and inter-grader agreement of total retinal blood flow (TRBF) measurements obtained by semi-automated grading of Doppler Spectral Domain OCT (SD-OCT) scans.
Methods:
Doppler OCT scans from 20 eyes of 20 subjects were obtained using a circumpapillary dual circular scan protocol on the Optovue RTVue SD-OCT. These scans were analyzed using the Doppler OCT of Retinal Circulation (DOCTORC) software (version2) which automatically identifies candidate vessel locations. Human graders can then manually refine and correct the scans by adding or deleting vessels, changing vessel boundaries and identifying the vessels as veins or arteries. After this manual step, the software then calculates the venous blood flow (and hence TRBF) automatically using the Doppler signal and the vessel area. In this study, two masked graders independently refined the retinal vessel detections, and the agreement in TRBF values generated by the two graders was analyzed using SPSS software.
Results:
Mean (± SD, Max) TRBF of the subjects in this series was 38.49 µl/min (± 10.84, 57.14) for grader 1 and 39.04 µl/min (± 10.65, 57.14) for grader 2, with an intra-class correlation coefficient of 0.933 and a coefficient of variability of 0.96. Bland-Altman Plot (Figure 1) illustrates the agreement between graders. Limits of agreement were -11.1 and +10.0. Venous area measurements showed similar levels of agreement with a mean (± SD, Max) area of 47.27 mm2 (± 8.35, 61.01) for grader 1 and 50.18 mm2 (± 8.74, 67.84) for grader 2, with an intra-class correlation coefficient of 0.933.
Conclusions:
Using semi-automated software with manual refinement, reproducible measurements of TRBF can be obtained from Doppler OCT data. These findings should be of value in future studies evaluating retinal blood flow in various diseases.
Keywords: imaging/image analysis: non-clinical • image processing