Advances in Human PET-MRI

Thursday 15 May 2014

Chang Kim¹, William T Peterson¹, Tesfaye Kidane¹, Sri Harsha Maramraju¹, and Craig S Levin^2
1GE Healthcare, Waukesha, WI, United States, ²Department of Radiology, Stanford University School of Medicine, Stanford, CA, United States

We evaluated the impact of gradient-induced thermal loads on the PET detector ring in a hybrid time-of-flight PET/MR system developed for simultaneous whole body PET/MR imaging. Varying temperature profiles generated due to the high intensity gradient fields compromise the performance of PET detectors that are in close proximity of the RF shield. Gradient echo EPI sequences were used to study the drifts in the 511 keV energy peak of the PET detectors. A thermal regulation algorithm was developed and applied to correct the energy peak drift. Results show that the method developed compensates well for MR induced thermal drift.

Daniel H Paulus¹, Daniela Thorwarth², and Harald H Quick^1
1Institute of Medical Physics, University of Erlangen-Nuremberg, Erlangen, Germany, ²Section of Biomedical Physics, University Hospital for Radiation Oncology, Eberhard-Karls-Universität Tübingen, Tübingen, Germany

MR and PET imaging have both become an important part in radiation therapy (RT) treatment planning to improve the accuracy of target volume delineation. A prototype RT table and two RF coil holders that each fix one flexible body matrix RF coil for head imaging are introduced for hybrid PET/MR imaging and tested towards MR and PET compatibility. MR and PET image quality has been validated with phantom scans and attenuation correction methods are presented. Furthermore, an in vivo study on two patients has been included to illustrate the clinical usage of these RT devices in PET/MR hybrid imaging.

Christin Y. Sander^1,2, Ciprian Catana¹, Aijun Zhu¹, Cathy Poulsen³, Bruce R. Rosen^1,4, and Giorgio Bonmassar^1
1A. A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Harvard Medical School, Charlestown, MA, United States, ²Electrical Engineering, Massachusetts Institute of Technology, Cambridge, MA, United States, ³Electrical Geodesics, Inc., Eugene, OR, United States, ⁴Health Sciences and Technology, Harvard-MIT, Cambridge, MA, United States

The simultaneous acquisition of non-invasive human brain imaging techniques, including EEG, PET and fMRI can give insights into the connections between neuronal activity, receptor occupancy and function. However, standard dense-array EEG caps can cause artifacts in both MR and PET images, making simultaneous imaging a challenge. We thus designed a 256-electrode EEG cap based on conductive polymers, and demonstrate its superior performance in RF invisibility and reduced photon attenuation for the use of simultaneous EEG-PET/MR imaging.

Lingzhi Hu¹, Christian Stehning², Holger Eggers², Zhiqiang Hu¹, and Lingxiong Shao^3
1Philips Healthcare, Cleveland, OH, United States, ²Philips Research, Hamburg, Germany, ³Philips Healthcare, San Jose, United States

MR-based Attenuation Correction (MRAC) is crucial for PET quantitation and image quality in a hybrid PET/MR system. An ideal MR sequence for attenuation correction should at least satisfy three criteria: (1) whole-body scan capabilities; (2) good anatomical localization; (3) robust tissue classification. UTE-mDixon (if it is still meant “generally” here, it should be Dixon, not mDixon, and you should add some references) sequences have demonstrated great potential for enhancing contrast of cortical bone and for water/fat separation within a single acquisition. Herein we report our initial experience of utilizing a UTE-mDixon sequence for whole-body PET/MR attenuation correction

Mootaz Eldib¹, Jason Bini^1,2, Philip M Robson¹, David Faul³, and Zahi A Fayad^1
1Translational and Molecular Imaging Institute, Ichan School of Medicine at Mount Sinai, New York, NY, United States, ²Biomedical Engineering, The City College of New York, NY, United States, ³Siemens Healthcare, New York, New York, United States

In this abstract we investigated a novel method for attenuation correction for flexible surface coils in MR/PET imaging. We utilized the UTE sequence to localize a carotid coil which allowed for accurate registration of a coil attenuation map. Phantom and clinical data are presented.

Meher Juttukonda¹, Yasheng Chen², Yi Su³, Tammie Benzinger³, Brian Rubin³, David Lalush¹, and Hongyu An^2
1Biomedical Engineering, University of North Carolina, Chapel Hill, North Carolina, United States, ²Radiology, University of North Carolina, Chapel Hill, North Carolina, United States, ³Radiology, Washington University, St. Louis, Missouri, United States

In this study, we have developed and evaluated a UTE- and R2*-based segmentation method for performing attenuation correction and compared it to a published method. Our results show that the proposed method outperforms the existing method in the accuracy of bone/air segmentation as well as in the accuracy of simulated PET images reconstructed using the attenuation maps derived from the segmentation methods.

Andrew Peter Aitken¹, Charalampous Tsoumpas¹, Daniel Giese¹, Sebastian Kozerke^1,2, Claudia Prieto¹, and Tobias Schaeffter^1
1Division of Imaging Sciences and Biomedical Engineering, King's College London, London, London, United Kingdom, ²Institute for Biomedical Engineering, University and ETH Zurich, Zurich, Switzerland

UTE sequences have been used to provide segmented attenuation correction (AC) maps for PET. Such sequences are prone to eddy current artifacts, which lead to deviations from nominal k-space trajectories and to misclassifications in the AC maps. These effects can be corrected for by measuring the true k-space trajectories using a magnetic field camera. In this study the effect of misclassifications on PET reconstruction are investigated using a PET simulation. Uptake in the brain was overestimated by 12.2% when AC-maps derived using nominal k-space trajectories are used, which reduces to 0.34% when AC maps derived from measured trajectories are used.

Sven Prevrhal¹, Susanne Heinzer², Bénédicte Delattre^2,3, Steffen Renisch⁴, Christian Wülker⁵, Osman Ratib³, and Peter Börnert^1
1Philips Research, Hamburg, Germany, ²Philips AG Healthcare, Zurich, Switzerland, ³University Hospital of Geneva, Geneva, Switzerland, ⁴Philips Research, Hamburg, Deutschland, Germany, ⁵University of Heidelberg, Germany

The fusion of the information from PET and MRI can increase the diagnostic value of both modalities. Compared to MRI, PET has limited spatial resolution, which could potentially be improved by using the information MRI can deliver. In this work the hypothesis was employed that fatty tissue is metabolically not very active in terms of glucose consumption. Thus, a PET tracer like 18F-deoxyglucose (FDG) has limited accumulation in fatty tissue and should have no significant PET signal in this compartment. Thus, the fat distribution measured in water/fat-resolved mDixon MRI can be used as a prior in an appropriately modified PET image reconstruction. Phantom and in vivo results show the basic feasibility of this approach.

Christoph Kolbitsch¹, Claudia Prieto¹, Charalampos Tsoumpas¹, and Tobias Scheaffter^1
1Division of Imaging Sciences and Biomedical Engineering, King's College London, London, London, United Kingdom

Positron emission tomography (PET) is commonly used to detect tumours and assess the progress of cancer treatment by measuring standardized uptake values (SUV). Physiological motion such as bulk-motion shifts during data acquisition can negatively impair obtained SUV. Here we evaluate a simultaneous PET-MR acquisition to detect and correct for any non-rigid bulk motion shifts directly from acquired gradient or spin echo MR data without the need of additional motion information. Volunteer studies show an accuracy of 1.71 ± 0.29mm for the obtained motion information reducing SUV errors in motion corrected PET simulations from 65% to below 10%.

Thomas Koesters^1,2, Li Feng¹, Kai Tobias Block¹, Michael Fieseler³, Klaus P Schäfers³, David Faul⁴, Daniel K Sodickson¹, and Fernando Emilio Boada^1,2
1Center for Biomedical Imaging, Department of Radiology, NYU Langone Medical Center, New York, New York, United States, ²CAI2R, Center for Advanced Imaging Innovation and Research, NYU Langone Medical Center, New York, New York, United States, ³European Institute for Molecular Imaging, Westfälische Wilhelms-Universität Münster, Münster, Germany, ⁴Siemens Medical Solutions, New York, New York, United States

This work presents a new technique for motion compensation of simultaneous PET/MR exams. Motion information is obtained from the MR data by extracting a respiration signal, binning the MR data into respiratory states, and generating a motion operator through cross-registration of images reconstructed for the different respiration states. The motion operator is then incorporated into a modified expectation-maximization algorithm to directly reconstruct motion-free PET images without calculation of intermediate gated PET images. Results from first in-vivo reconstructions are presented and demonstrate a clear reduction of image blurring in PET exams acquired during normal breathing.