Blind Multichannel Deconvolution for Estimation of a Parametric AIF in DCE-MRI of Mice

0434531 - ÚPT 2015 US eng A - Abstract
Jiřík, Radovan - Souček, Karel - Dražanová, Eva - Krátká, Lucie - Standara, M. - Kratochvíla, J. - Macíček, O. - Malá, Aneta - Taxt, T. - Starčuk jr., Zenon
Blind Multichannel Deconvolution for Estimation of a Parametric AIF in DCE-MRI of Mice.
Proceedings of the International Society for Magnetic Resonance in Medicine (ISMRM). Vol. 22. Berkeley: Society of Magnetic Resonance, 2014. s. 2754. ISSN 1545-4428.
[Joint Annual Meeting ISMRM-ESMRMB 2014 /23./. 10.05.2014-16.05.2014, Milan]
R&D Projects: GA ČR GAP102/12/2380; GA MŠMT ED0017/01/01
Institutional support: RVO:68081731
Keywords : Blind Multichannel Deconvolution * DCE-MRI
Subject RIV: FS - Medical Facilities ; Equipment

The main challenge in dynamic contrast-enhanced (DCE) MRI, especially when applied to mice and rats, is to use a correct arterial input function (AIF). Measuring the AIF from an artery within the slice of interest leads to partial-volume, flow and dispersion artifacts. Measuring the AIF in a slice intersecting heart which is acquired in addition to the slice of interest minimizes the partial volume effect but at the cost of the time resolution in the acquired image sequence. An alternative avoiding these problems is to use blind multichannel deconvolution where the AIF is estimated directly from the measured contrast-agent concentration curves in several tissues. Compared to clinical DCE-MRI 1, application of this approach to small animal imaging is limited only to a nonparametric (model-free) AIF 2. To improve the noise-robustness, a model-based AIF approach is introduced to multichannel blind deconvolution in small-animal DCE-MRI.
Permanent Link: http://hdl.handle.net/11104/0238563