Yannick Kuhl¹, Stephan Naunheim¹, Bjoern Weissler^1,2, Vanessa Nadig¹, David Schug^1,2, Florian Mueller¹, Harald Radermacher¹, Pierre Gebhardt¹, Nicolas Gross-Weege¹, Teresa Nolte¹, Martino Borgo³, Marcel de Koning³, Menno Mathlener³, Jeroen Koeleman³, Frank van Duin³, Arold Dekker³, Wout Schut³, Jos van den Berghe³, Daniel Gareis⁴, Turgay Celik⁴, André Salomon⁵, Dennis Schaart⁶, Dimitri Kuznetsov⁶, René Bakker⁶, Karl-Josef Langen⁷, Christiane Kuhl⁸, and Volkmar Schulz^1,2,9,10
1Physics of Molecular Imaging Systems (PMI), Experimental Molecular Imaging (ExMI), Aachen, Germany, ²Hyperion Hybrid Imaging Systems GmbH, Aachen, Germany, ³Futura Composites B.V., Heerhugowaard, Netherlands, ⁴NORAS MRI Products GmbH, Hoechberg, Germany, ⁵Philips Research, Eindhoven, Netherlands, ⁶Delft University of Technology, Delft, Netherlands, ⁷Forschungszentrum Juelich, Juelich, Germany, ⁸Department of Diagnostic and Interventional Radiology, University Hospital Aachen, Aachen, Germany, ⁹Physics Institute III B, RWTH Aachen University, Aachen, Germany, ¹⁰Fraunhofer Institute for Digital Medicine MEVIS, Aachen, Germany

Within the H2020 EU project HYPMED, we have developed a PET insert that can be moved into a 1.5 T MRI Philips Ingenia turning this into a simultaneous MRI-PET device for breast cancer applications. The PET insert consists of two independent PET rings with a field of view of 28$$$\times$$$10 cm² and two dual-channel local receive coils. Each ring consists of 14$$$\times$$$2 detector blocks of which each features a three-layer LYSO crystal array with 1.3 mm pitch. We will present the first PET and MRI results of our system.

Ria Forner¹, Woutjan Branderhorst¹, Debra Rivera², Alexander Raaijmakers^1,2, Dimitri Welting¹, and Dennis Klomp^1
1UMC Utrecht, Utrecht, Netherlands, ²Eindhoven University of Technology, Eindhoven, Netherlands

The feasibility of simultaneous 1H/31P PET/MR imaging added-on to an existing 7T MRI system was assessed. Space constraints preclude using whole-body birdcages. Instead, an 8-channel 1H dipole array stacked with additional 8-31P channels is set up. By positioning PET detectors between the RF-shield and gradient coil, further shielding of the PET-crystals and electronics can be omitted. 8-channel 31P scans had a higher SNR compared to a birdcage setup.  The 1H/31P setup gives ~5% degradation of PET sensitivity. Similarly, the impact of the new setup (proximity of the RF shield) was insignificant for MRI performance at both frequencies.

Yun-Jeong Stickle¹, Mark Giancola¹, Clyve Konrad Follante¹, Thomas Stickle¹, Fraser Robb¹, Holly Blahnik², and Tae-Young Yang^1
1GE Healthcare, Aurora, OH, United States, ²GE Healthcare, Waukesha, WI, United States

A PET optimized 21- Element Head Neck coil is described for acquiring high-sensitivity MR (Magnetic Resonance) images and good quality PET images of head/neck, carotids, and cervical spine at 3 Tesla simultaneous MR/PET system. This Head Neck prototype2 coil has been optimized to minimize the interference with -ray detection of the PET without sacrificing MR image quality.  In this study, we developed a foam technology to replace the rigid plastic coil formers, optimized arrangement of components (internal cables, internal cable baluns, feed boards and decoupling boards) and introduced new mechanical fastener and snap features.

Laiyin Yin¹, Nicolas Gross-Weege¹, Teresa Nolte¹, Bjoern Weissler^1,2, David Schug^1,2, and Volkmar Schulz^1,2,3,4
1Department of Physics of Molecular Imaging Systems, Institute of Experimental Molecular Imaging, RWTH Aachen University, Aachen, Germany, ²Hyperion Hybrid Imaging Systems GmbH, Aachen, Germany, ³Physics Institute III B, RWTH Aachen University, Aachen, Germany, ⁴Fraunhofer Institute for Digital Medicine MEVIS, Aachen, Germany

We present a preclinical PET/MRI system with highly-integrated PET modules in which the RF shielding is implemented on scintillator level for a 7 T Bruker BioSpec 70/20. The PET system's RF housing size is minimized by excluding parts of the scintillator while maintaining the RF shielding properties through the scintillator-integrated shielding. The modules are mounted inside the MRI coil, between resonator and RF screen. MRI compatibility is evaluated in a mouse phantom by means of spurious noise scans and SNR evaluations of standard morphologic sequences, while the PET system is off, idle and measuring. Additionally, postmortem mouse images are presented.

Zhuopin Sun¹, Georgios Angelis^1,2, Steve Meikle^3,4, and Fernando Calamante^1,4,5
1School of Biomedical Engineering, The University of Sydney, Sydney, Australia, ²Australian National Imaging Facility, The University of Sydney, Sydney, Australia, ³Faculty of Medicine and Health, The University of Sydney, Sydney, Australia, ⁴Brain and Mind Centre, The University of Sydney, Sydney, Australia, ⁵Sydney Imaging, The University of Sydney, Sydney, Australia

Diffusion MRI (dMRI) can provide a wealth of information about brain microstructure and connectivity, but its use in PET-MRI image reconstruction has not been exploited. We incorporated dMRI-derived prior information as a regulariser into the iterative PET One-Step Late Maximum A Posteriori (OSLMAP) reconstruction algorithm. The proposed regularisation method has unique advantages of providing more informative and targeted denoising and regularisation based on the complementary structural connectivity information from dMRI.

Confidence Raymond^1,2, Jurkiewicz Michael ^1,2, Akin Orunmuyi³, Dada Oluwaseun Michael ⁴, Claes Nøhr Ladefoged⁵, Jarmo Teuho⁶, and Udunna Anazodo^1,2
1Medical Biophysics, Western University Ontario, London, ON, Canada, ²Lawson Health Research Institute, LONDON, ON, Canada, ³Anaesthesia, College of Medicine, Ibadan, Nigeria, ⁴Physics, Federal University of Technology, Minna, Nigeria, ⁵Clinical Physiology, Nuclear Medicine and PET, Rigshospitalet, Denmark, ⁶Turku PET Centre, Turku University, Turku, Finland

The last decade has seen an increase in the application of machine learning (ML) methods to PET/MRI attenuation correction (AC). This systematic review provides a head-to-head comparison between state-of-the-art ML methods and clinical standards for AC to determine the clinical feasibility of ML approaches PET AC. We extracted numerical values for image quality, tissue classification, regional and global diagnostic performance. The pooled mean relative error for global performance was 0.87 ± 1.3%, the quality of evidence for all outcomes ranged from moderate to very low. Our findings suggest that ML-AC performance is within acceptable limits for clinical PET/MR neuroimaging. 

Alina Schneider¹, Camila Munoz¹, Alina Hua¹, Sam Ellis¹, Sami Jeljeli², Karl P Kunze^1,3, Radhouene Neji^1,3, Eliana Reyes¹, Tevfik F Ismail¹, René M Botnar¹, and Claudia Prieto^1
1School of Biomedical Engineering and Imaging Sciences, King's College London, London, United Kingdom, ²PET Centre, St Thomas’ Hospital, King’s College London & Guys and St Thomas’ NHS Foundation Trust, London, United Kingdom, ³MR Research Collaborations, Siemens Healthcare Limited, Frimley, United Kingdom

Simultaneously acquired ¹⁸F-FDG PET-MR imaging and quantitative 2D T₂-mapping have been suggested for improved diagnostic accuracy of cardiac sarcoidosis, however misregistration between imaging modalities makes clinical interpretation challenging. Here we used a recently proposed free-breathing motion-corrected 3D whole-heart T₂-mapping sequence acquired simultaneously with ¹⁸F-FDG PET at a 3T PET-MR system. This approach enables the non-rigid motion-correction for both the 3D T2-mapping and the PET data to the same respiratory position, resulting in aligned volumes for improved clinical interpretation. In this study, we tested this approach in a patient with suspected cardiac sarcoidosis.  

Josephine Tan¹, Alireza Akbari¹, Matthew Fox², Mena Gaed³, Madeleine Moussa⁴, Jose A Gomez⁴, Zahra Kassam⁵, William Pavlosky⁵, Joseph L Chin⁶, Stephen Pautler⁶, Nicholas Power⁶, Aaron D Ward^1,2, Glenn S Bauman^1,2, Jonathan D Thiessen^1,2,5, and Timothy J Scholl^1,3,7
1Medical Biophysics, University of Western Ontario, London, ON, Canada, ²Lawson Health Research Institute, London, ON, Canada, ³Robarts Research Institute, University of Western Ontario, London, ON, Canada, ⁴Pathology and Laboratory Medicine, University of Western Ontario, London, ON, Canada, ⁵Medical Imaging, University of Western Ontario, London, ON, Canada, ⁶Surgery, Division of Urology, University of Western Ontario, London, ON, Canada, ⁷Ontario Institute for Cancer Research, Toronto, ON, Canada

This abstract focuses on the development of an imaging assay based on multiparametric MRI and sodium (²³Na) MRI to discriminate between low- and high-grade prostate cancer (PCa) lesions, using Gleason grade defined by whole-mount histopathology as the gold standard and PET targeting the prostate-specific membrane antigen (PSMA) as a validation. Data from the first patient (Gleason score 7) demonstrates that PSMA-PET may be superior in detecting PCa lesions in the transition zone and distant metastases while the sensitivity of the current ²³Na radiofrequency system in this study is limited to lesions in the peripheral zone of the prostate.