RESUMO
INTRODUCTION: Direct quantification of atherosclerotic plaque volume by three-dimensional vascular ultrasound (3DVUS) is more reproducible than 2DUS-based three-dimensional (2D/3D) techniques that generate pseudo-3D volumes from summed 2D plaque areas; however, its accuracy has not been reported. We aimed to determine 3DVUS accuracy for plaque volume measurement with special emphasis on small plaques (a hallmark of early atherosclerosis). METHODS: The in vitro study consisted of nine phantoms of different volumes (small and medium-large) embedded at variable distances from the surface (superficial vs. >5 cm-depth) and comparison of 3DVUS data generated using a novel volumetric-linear array method with the real phantom volumes. The in vivo study was undertaken in a rabbit model of atherosclerosis in which 3DVUS and 2D/3D volume measurements were correlated against gold-standard histological measurements. RESULTS: In the in vitro setting, there was a strong correlation between 3DVUS measures and real phantom volume both for small (3.0-64.5 mm(3) size) and medium-large (91.1-965.5 mm(3) size) phantoms embedded superficially, with intraclass correlation coefficients (ICC) of 0.99 and 0.98, respectively; conversely, when phantoms were placed at >5 cm, the correlation was only moderate (ICC = 0.67). In the in vivo setting there was strong correlation between 3DVUS-measured plaque volumes and the histological gold-standard (ICC = 0.99 [4.02-92.5 mm(3) size]). Conversely, the correlation between 2D/3D values and the histological gold standard (sum of plaque areas) was weaker (ICC = 0.87 [49-520 mm(2) size]), with large dispersion of the differences between measurements in Bland-Altman plots (mean error, 79.2 mm(2)). CONCLUSIONS: 3DVUS using the volumetric-linear array method accurately measures plaque volumes, including those of small plaques. Measurements are more accurate for superficial arterial territories than for deep territories.
