Perfusion quantification using Gaussian process deconvolution

• Published in: Magnetic resonance in medicine Vol. 48; no. 2; pp. 351 - 361
• Main Authors: Andersen, I.K., Szymkowiak, A., Rasmussen, C.E., Hanson, L.G., Marstrand, J.R., Larsson, H.B.W., Hansen, L.K.
• Format: Journal Article
• Language: English
• Published: New York Wiley Subscription Services, Inc., A Wiley Company 01.08.2002; Williams & Wilkins
• Subjects: Biological and medical sciences; Cerebrovascular Circulation; Contrast Media; dynamic susceptibility contrast; Gadolinium DTPA; Gaussian process deconvolution; Humans; Image Processing, Computer-Assisted; impulse response function; Investigative techniques, diagnostic techniques (general aspects); Magnetic Resonance Imaging - methods; Medical sciences; Miscellaneous. Technology; Nervous system; Normal Distribution; perfusion quantification; Radiodiagnosis. Nmr imagery. Nmr spectrometry; singular value decomposition; Human; Central nervous system; Nuclear magnetic resonance imaging; Image quality; Time resolution; Deconvolution; Gaussian process; Perfusion; Signal processing; Technique; Comparative study; Quantitative analysis; Singular value decomposition; Brain (vertebrata); Signal to noise ratio
• ISSN: 0740-3194; 1522-2594
• DOI: 10.1002/mrm.10213

The quantification of perfusion using dynamic susceptibility contrast MRI (DSC‐MRI) requires deconvolution to obtain the residual impulse response function (IRF). In this work, a method using the Gaussian process for deconvolution (GPD) is proposed. The fact that the IRF is smooth is incorporated as a constraint in the method. The GPD method, which automatically estimates the noise level in each voxel, has the advantage that model parameters are optimized automatically. The GPD is compared to singular value decomposition (SVD) using a common threshold for the singular values, and to SVD using a threshold optimized according to the noise level in each voxel. The comparison is carried out using artificial data as well as data from healthy volunteers. It is shown that GPD is comparable to SVD with a variable optimized threshold when determining the maximum of the IRF, which is directly related to the perfusion. GPD provides a better estimate of the entire IRF. As the signal‐to‐noise ratio (SNR) increases or the time resolution of the measurements increases, GPD is shown to be superior to SVD. This is also found for large distribution volumes. Magn Reson Med 48:351–361, 2002. © 2002 Wiley‐Liss, Inc.