Blood vessel structural morphology derived from 3D dual-frequency ultrasound images

The method of dual-frequency contrast ultrasound imaging can produce high resolution, high contrast 3D maps of the microvasculature. In this manuscript, we demonstrate an extension of the technique, showing the potential to segment individual blood vessels from 3D datasets in order to quantify their structural morphologies. There is a known relationship between abnormal blood vessel structure and the presence of cancer, with diseased vasculature generally showing a high degree of tortuosity and irregularity. These irregular vessel paths have been shown to normalize with successful anti-cancer treatment, effects which generally precede gross tumor volume changes. Thus, we hypothesize that our ultra-broadband imaging technique could be used in conjunction with vessel segmentation and morphological analyses to assess in-vivo therapeutic response with greater sensitivity than previously possible. In this paper, we demonstrate the consistency of our segmentation algorithm, and estimate the limitations of detectable tortuosity levels.