Towards Standardization of Quantitative Retinal Vascular Parameters: Comparison of SIVA and VAMPIRE Measurements in the Lothian Birth Cohort 1936

Purpose: Semiautomated software applications derive quantitative retinal vascular parameters from fundus camera images. However, the extent of agreement between measurements from different applications is unclear. We evaluate the agreement between retinal measures from two software applications, the Singapore I Vessel Assessment (SIVA) and the Vessel Assessment and Measurement Platform for Images of the Retina (VAMPIRE), and examine respective associations between retinal and systemic outcomes. Method: Fundus camera images from 665 Lothian Birth Cohort 1936 participants were analyzed with SIVA and VAMPIRE. Intraclass correlation coefficients (ICC) and Bland-Altman plots assessed agreement between retinal parameters: measurements of vessel width, fractal dimension, and tortuosity. Retinal-systemic variable associations were assessed with Pearson's correlation, and intersoftware correlation magnitude differences were examined with Williams's test. Results: ICC values indicated poor to limited agreement for all retinal parameters (0.159-0.410). Bland-Altman plots revealed proportional bias in the majority, and systematic bias in all measurements. SIVA and VAMPIRE measurements were associated most consistently with systemic variables relating to blood pressure (SIVA r's from -0.122 to -0.183; VAMPIRE r's from -0.078 to -0.177). Williams's tests indicated significant differences in the magnitude of association between retinal and systemic variables for 7 of 77 comparisons (P < 0.05). Conclusions: Agreement between two common software applications was poor. Further studies are required to determine whether associations with systemic variables are software-dependent. Translational Relevance: Standardization of the measurement of retinal vascular parameters is warranted to ensure that they are reliable and application-independent. This would be an important step towards realizing the potential of the retina as a source of imaging-derived biomarkers that are clinically useful.