Reproducibility of retinal and optic nerve head perfusion measurements using scanning laser Doppler flowmetry

To evaluate the reproducibility of full-field perfusion analysis using scanning laser Doppler flowmetry (SLDF) for perfusion measurements of the neuroretinal rim of the optic nerve head and the peripapillary retina in patients with open-angle glaucoma or ocular hypertension, and in normal subjects.

SLDF perfusion measurements of the neuroretinal rim and the peripapillary retina were performed on 20 patients with open-angle glaucoma or ocular hypertension (group G) and 20 normal volunteers (group N). Each subject underwent two independent sessions, 30 minutes apart, each involving 5 high quality images. Intrasession and intersession reproducibility coefficients for flow, volume, and velocity were calculated for a single image and for means of 3 and 5 images using analysis of variance models. flow, the intrasession coefficient of reliability was 0.99 each for the rim, nasal retina, and temporal retina in group G and 0.93, 0.93, and 0.95, respectively, in group N. The intersession coefficient of reliability for flow was 0.99 for the rim, 0.95 for the nasal retina, and 0.87 for the temporal retina in group G and 0.87, 0.82, and 0.80, respectively, in group N. Compared with single image analysis, intrasession and intersession reproducibility were generally better when a mean of 3 images and substantially better when a mean of 5 images was used.

SLDF full-field perfusion analysis is markedly more reproducible than the original software using 10 x 10 pixel windows. Obtaining mean values for at least 3 images improves the intrasession and intersession reproducibility of this technique.