Assessment of cardiovascular risk markers from ultrasound images: System reproducibility

Abstract

In this work the reproducibility of a system for the automatic assessment of carotid intima-media thickness and stiffness from ultrasound image sequences is evaluated. Sequences of the right/left common carotid arteries of 10 healthy volunteers were analyzed repeatedly in order to evaluate the reproducibility between observers and sessions; data were presented as the coefficients of variation. Results regarding the inter-observer and intra-observer intra-session variabilities were comparable. Intra-observer inter session variability was similar to those previously mentioned for the parameters which were derived directly from the B-mode image analysis, but larger for the others. The variability of the data obtained with the system is comparable to that reported in literature for other techniques used for the evaluation of the mechanical properties of large arteries. Thus, the reproducibility of our device, its low-cost and flexibility make it suitable for large population studies.