HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Golish, S. R. Articles by Gambhir, S. S. Search for Related Content PubMed PubMed Citation Articles by Golish, S. R. Articles by Gambhir, S. S.

A Fast Nonlinear Method for Parametric Imaging of Myocardial Perfusion by Dynamic ¹³N-Ammonia PET

Crump Institute for Molecular Imaging and Departments of Computer Science, Molecular and Medical Pharmacology, and Biomathematics, University of California, Los Angeles, Los Angeles, California; and Rigshospitalet, Copenhagen, Denmark

View larger version (15K): [in a new window]   FIGURE 1. Example of 1 simulated pair of blood and tissue time–activity curves. Curve that spikes is blood time–activity curve. Smooth line is continuous activity; = integrated activity; = noisy activity (final time–activity curve). Curve that does not spike is tissue time–activity curve. Smooth line is continuous activity; = integrated activity; x = noisy activity (final time–activity curve). For simulation: perfusion = 1.0; spillover = 0.0; blood time–activity curve noise level = 2%; tissue time–activity curve noise level = 20%; all other parameters are drawn at random from Table 1. Note high noise level in tissue time–activity curve and sampling error at 5 s for blood time–activity curve = .

View larger version (16K): [in a new window]   FIGURE 2. Graphical representation of sigmoidal networks. Here, r is dimensionality of input, s is number of basis functions, and each node represents dot product of input and local weights, followed by a sigmoidal transfer function. Output node has linear transfer function (linear regression on basis functions). Sigmoidal network for perfusion imaging has 20 inputs (10 samples each of blood and tissue time–activity curve), 52 basis functions, and 1 output (perfusion). Use of single output is explained in Discussion.

View larger version (44K): [in a new window]   FIGURE 3. Parametric images of myocardial perfusion by NRE (A), WNLR (B), and Patlak analysis (C). Pseudocolored scale bar shows units of mL/min/g. Ischemic region at upper right is caused by vessel ligation.

View larger version (18K): [in a new window]   FIGURE 4. Scatter plot of NRE versus WNLR () and Patlak analysis versus WNLR (x) for 32 ROIs taken from 4 canine subjects. Least squares line fits are also shown.