ABSTRACT
Max Planck Institute for Biological Cybernetics, Tu¨bingen, Germany Laboratory for Preclinical Imaging and Imaging Technology, University of Tu¨bingen, Tu¨bingen, Germany Department of Engineering Science, University of Oxford, Oxford, United Kingdom
Bernd Pichler
Laboratory for Preclinical Imaging and Imaging Technology, University of Tu¨bingen, Tu¨bingen, Germany
Thomas Beyer
cmi-experts GmbH, Zu¨rich, Switzerland
11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 11.2 MR-AC for brain applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
11.2.1 Segmentation approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 11.2.2 Atlas approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
11.3 Methods for torso imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224 11.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
11.4.1 The presence of bone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230 11.4.2 MR imaging with ultrashort echo time (UTE) . . . . . . . . . . 231 11.4.3 Required PET accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232 11.4.4 Validation of MR-AC methods . . . . . . . . . . . . . . . . . . . . . . . . . . 232 11.4.5 Truncated field-of-view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232 11.4.6 MR coils and positioning aids . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 11.4.7 User intervention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 11.4.8 Potential benefits of MR-AC . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234 11.4.9 Additional potential benefits of simultaneous PET/MR
acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234 11.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
Current concepts of combined PET/MR tomographs do not allow for separate CT-like transmission sources. Therefore, corresponding PET attenuation coefficients must be calculated from the available MR images. MR-based attenuation correction (MR-AC) is far more challenging than the well-established algorithms for CT-based attenuation correction (CT-AC) since MR image voxel values correlate with the hydrogen nuclei density in tissues and tissue relaxation properties rather than with electron density-related mass attenuation coefficients on CT (Figure 11.1). Therefore, a direct transformation from available MR images to CT-like attenuation values is challenging [3].
