Background Assessment of the perifoveal capillary network (PCN) might indicate macular function and could reflect the systemic microcirculation. The quantification and reliability of this measurement is currently unknown. The aim of this study was to validate quantification of the PCN by a non-invasive technique from high-resolution retinal images.

Methods Ten healthy volunteers were included in this validation study. At least 320 high-resolution retinal images were used for assessment of inter- and intra-observer reliability. Non-invasive capillary perfusion mapping was performed using a retinal function imager. After the images were enhanced and segmented, the reproducibility was verified by comparing the values of two independent examiners and of a single examiner at two different time points.

Results The inter-observer concordance coefficients were highly significant for PCN (intraclass correlation coefficient (ICC)=0.901, 95% CI 0.655 to 0.975, p<0.001) and normalised PCN (ICC=0.727, 95% CI 0.262 to 0.923, p=0.004). The intra-observer measurements at two different time points were also highly concordant for PCN (ICC=0.879, 95% CI 0.598 to 0.968, p<0.001) and for normalised PCN (ICC=0.960, 95% CI 0.851 to 0.990, p<0.001).

Conclusions The reliability of PCN measurement is reproducible and could be used as a new tool to quantify the capillary perfusion network of the macular area.