By Topic

Fully automated partial volume correction in PET based on a wavelet approach without the use of anatomical information

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

4 Author(s)

A well-known consequence of the limited spatial resolution in positron emission tomography is partial volume effects (PVE). These latter hamper accurate quantification by attenuating intensity in areas of size similar to the point spread function, and by inducing cross-contamination in adjacent tissues. Most existing PVE correction methods necessitate the segmentation of tissues either from MRI or CT images. Moreover, adequate tissue correlation between PET and MRI or CT is mandatory and difficult to achieve especially because of physiological motion of organs, patient uncontrollable movement in a more general sense, and possible difference in tissue homogeneity. In a previous work we proposed a method aiming at producing PVE corrected images without drawing regions of interest. On the other hand, this methodology requires an anatomical image and consequently, the correction suffers from the limitations related to the tissue spatial correlation described above. In the present article, we describe a modification of this algorithm eliminating the need for the use of anatomical images but keeping all of the advantages of the previously described approach. The basic modification is based on the use of iterative deconvolution of the PET image for the creation of the high resolution information instead of the anatomical image. In addition, noise propagation typical of deconvolution is suppressed by using a specific thresholding in the wavelet domain. An initial evaluation of this new methodology was performed on a series of 39 PET images (24 from patients and 15 simulated) in both oncology and brain domains. Results demonstrate large improvements in quantitative accuracy along with significant image enhancement. The approach lead to images with better delineated tissues and organs, without losing the functional information of emission tomography.

Published in:

Nuclear Science Symposium Conference Record, 2007. NSS '07. IEEE  (Volume:4 )

Date of Conference:

Oct. 26 2007-Nov. 3 2007