Near-automated quantification of prenatal aortic intima-media thickness from ultrasound images

Aortic intima-media thickness (aIMT) is an early marker for atherosclerosis and cardiovascular diseases risk assessment in children and young adults. Recent studies have underlined the potential usefulness of its estimation at the fetal stage from ultrasound (US) images. However, this measurement currently relies on tedious and error-prone manual tracing. The aims of this study were to develop and test a near-automated technique for aIMT quantification from US images. The proposed technique is based on narrow-band level-set methods to identify blood-intima and media-adventitia interfaces, thus allowing aIMT estimation. The technique was tested on images acquired from 11 subjects at a mean gestational age of 29 weeks. Automatically estimated aIMT values were compared to reference ones manually extracted by an experienced interpreter. Quantitative comparisons were performed using Pearson's correlation coefficients, Bland-Altman and linear regression analyses. The results (R up to 0.92) indicate the high correlation between automatically and manually estimated values, suggesting that near-automated quantification of alMT from us images using level-set methods is feasible.