Angiogenesis imaging by spatiotemporal analysis of ultrasound-contrast-agent dispersion kinetics

M. Mischi, M.P.J. Kuenen, H. Wijkstra

Research output: Contribution to journalArticleAcademicpeer-review

51 Citations (Scopus)

Abstract

The key role of angiogenesis in cancer growth has motivated extensive research with the goal of noninvasive cancer detection by blood perfusion imaging. However, the results are still limited and the diagnosis of major forms of cancer, such as prostate cancer, are currently based on systematic biopsies. The difficulty in the detection of angiogenesis partly resides in a complex relationship between angiogenesis and perfusion. This may be overcome by analysis of the dispersion kinetics of ultrasound contrast agents. Determined by multipath trajectories through the microvasculature, dispersion permits a better characterization of the microvascular architecture and, therefore, more accurate detection of angiogenesis. In this paper, a novel dispersion analysis method is proposed for prostate cancer localization. An ultrasound contrast agent bolus is injected intravenously. Spatiotemporal analysis of the concentration evolution measured at different pixels in the prostate is used to assess the local dispersion kinetics of the injected agent. In particular, based on simulations of the convective diffusion equation, the similarity between the concentration evolutions at neighbor pixels is the adopted dispersion measure. Six measurements in patients, compared with the histology, provided a receiver operating characteristic curve integral equal to 0.87. This result was superior to that obtained by the previous approaches reported in the literature.
Original languageEnglish
Pages (from-to)621-629
JournalIEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control
Volume59
Issue number4
DOIs
Publication statusPublished - 2012

Fingerprint Dive into the research topics of 'Angiogenesis imaging by spatiotemporal analysis of ultrasound-contrast-agent dispersion kinetics'. Together they form a unique fingerprint.

Cite this