Integrated PET/MRI (positron emission tomography / magnetic resonance imaging) systems have only recently been introduced into routine clinical practice. The vast majority of PET/MRI installations have been performed within the last five years, mostly in tertiary care centers / university hospitals. Ever since the conception of PET/MRI as an imaging technique, the search for one or more “killer applications” – i.e., applications where PET/MRI would yield a significant clinical benefit, in comparison with the most widely established hybrid imaging technique, PET/CT (PET/computed tomography), or in comparison with MRI – has been a major topic of discussion in the radiology and nuclear medicine communities alike. This is not only due to the considerable costs associated with the installation and maintenance of a PET/MRI system, which clearly exceed those of PET/MRI; notably, also the fact that patient through-put is generally lower on PET/MRI systems than on the less cost-intensive PET/CT systems, requires a solid justification. The latter is particularly true for institutions in countries where financial reimbursement is a critical issue. Therefore, optimization of scanning protocols is a key issue, with the goal of minimizing total examination as well as in-room time, while providing/preserving all clinically relevant information.

In this presentation, we compare the work-flows of two clinical research institutions that are equipped with the same PET/MRI scanner model (a fully-integrated, simultaneous device): the Technical University of Munich (scanner installation: November 2010), and the Medical University of Vienna (scanner installation: December 2013). Oncologic imaging, neuroimaging and cardiac imaging are performed at both sites. With regard to oncologic imaging, prostate cancer represents a main research focus at both sites, in accordance with the international trend. In Vienna, lymphoma is another top indication, followed by melanoma, lung cancer, rectal cancer, pancreatic cancer, and gynecological tumors.

In order to optimize work-flow and patient management, as well as patient through-put, different approaches are presently used by different centers. One such approach is to perform only a limited number of simultaneous whole-body or single-region PET/MRI scans per day, and use the remaining scanner time for (mostly single-region) MR-only examinations. Another strategy is to perform only whole-body PET/MRI scans in all patients, but with a limited number of MR sequences, since the MR component of PET/MRI, rather than the PET component, is the time-limiting factor. The last approach is to focus on single- or (at maximum) dual-region PET/MRI, making use of the full range of MR pulse sequences, including functional techniques such as perfusion-weighted imaging and MR spectroscopy. Needless to say, several centers use a mix of the aforementioned PET/MRI strategies. Below, the work-flows at the Technical University of Munich and the Medical University of Vienna are described in more detail.

At both Munich and Vienna, PET/MRI for neurological/neurooncological applications as well as cardiac applications is performed with just one bed position; here the total examination time depends on the number of MR sequences, but is usually less than/equal to one hour. For cardiac applications, the examination time increases to at least 80 min if a dual-tracer PET scan (i.e., with consecutive injections of [18F]-FDG and [13N]-NH3) is performed.

For whole-body oncologic PET/MRI without a focus on a special site (i.e., for systemic diseases such as lymphoma or multiple myeloma), both Munich and Vienna perform the same MR sequences for each bed position, from head to pelvis, with simultaneous PET acquisition. For instance, in Vienna, the MR protocol for lymphoma comprises an axial T1 VIBE Dixon, a free-breathing EPI SPAIR DWI sequence, and a coronal T2 HASTE sequence; with this strategy, a whole body examination consisting of 4-5 bed positions requires a total examination time of about 40 min.

For whole-body oncologic PET/MRI with a focus on one anatomic region – the region of the primary tumor – an extensive MR protocol, involving contrast-enhanced sequences, is used by Munich and Vienna for this particular region, with an additional, minimalistic protocol for all 4-5 bed positions. The main difference between Munich and Vienna is that Munich performs (1) the minimalistic whole-body MR protocol (axial T2 HASTE fs and coronal T1 TSE) with simultaneous PET acquisition first, followed by (2) the extensive, 15-20 min, single-region MR protocol, with simultaneous PET, followed by (3) additional axial T1 VIBE sequences, from the lungs to the pelvis, without simultaneous PET. By contrast, Vienna performs (1) the extensive, 15-20 min, single-region MR protocol, with simultaneous early time point PET (e.g., 30 min after [18F]-FDG injection) first, followed by (2) the minimalistic whole-body MR protocol (axial T1 VIBE Dixon, coronal T2 HASTE) with simultaneous PET acquisition. For prostate cancer patients, and cancers that are known to affect the liver (e.g. pancreatic cancer), both centers will consider the use of dynamic contrast-enhanced sequences; for the liver, a liver-specific agent such as gadoxetate may be used.

The question of whether or not PET/MRI may also be used as just a single-region examination, but with an extensive, multi-parametric MR protocol, is still subject to debate.