Clinical Positron Emission Tomography/ Magnetic Resonance Imaging Applications

Gustav K. von Schulthess, MD, PhD, MD hon., Felix Pierre Kuhn, MD, MS, Philipp Kaufmann, MD, and Patrick Veit-Haibach, MD
Although clinical positron emission tomography (PET)/computed tomography (CT) applications were obvious and have completely replaced PET in oncology, clinical applications of PET/ magnetic resonance (MR) are currently not clearly dened. This is due to the lack of clinical data, which is mainly because PET/MR technology is not clinically mature at this point. Open issues are technical and concern ease of obtaining PET attenuation correction maps, dealing with, for example, MR surface coil metal in the PET eld-of-view and appropriate workows leading to a cost-effective examination. All issues can be circumvented by using a shuttleconnected PET/CT-MR system, but the penalty is that simultaneous PET and MR imaging are not possible and potential motion between examinations may occur. Clinically, some systems installed worldwide start to have a reasonable bulk of clinical data. Preliminary results suggest that in oncology, PET/MR may have advantages over PET/CT in head and neck imaging. In liver imaging, more PET-positive lesions are seen on MR than on CT, but that does not mean that PET/MR is superior to PET/CT. Possibly in some settings where a contrast-enhanced PET/CT is needed to be diagnostic, PET/MR can be done without contrast media. Although PET/CT has virtually no role in brain imaging, this may be an important domain for PET/MR, particularly in dementia imaging. The role of PET/MR in the heart is as yet undened, and much research will have to be done to elucidate this role. At this point, it is also not clear where the simultaneity afforded by a fully integrated PET/MR is really needed. Sequential data acquisition even on separate systems and consecutive software image fusion may well be appropriate. With the increasing installed base of systems, clinical data will be forthcoming and dene more clearly where there is clinical value in PET/MR at an affordable price. Semin Nucl Med 43:3-10 2013 Elsevier Inc. All rights reserved.

ositron emission tomography (PET)/magnetic resonance (MR) has been used in experimental setups for almost 20 years.1 The long time it has taken until this imaging modality has started to enter the clinical arena has to do with technical and practical hurdles that the clinical applications of PET/MR are facing. Technically, implementation of PET/MR faces 3 major problems.

Third, MR data, unlike those acquired using computed tomography (CT), are not readily usable for attenuation correction (AC).2-4 This limits quantication of PET data, which is particularly problematic when using PET in therapy response monitoring.

First, the photomultiplier-based PET scanners currently used do not work within or near the magnetic environment of an MR scanner, Second, metallic objects such as surface coils used to get best MR image quality interfere with the gamma rays from PET, causing unwanted attenuation, and

Department of Medical Radiology, University Hospital, CH-8091 Zurich, Switzerland. Address reprint requests to Gustav K. von Schulthess, Department of Medical Radiology, University Hospital, CH-8091, Zurich, Switzerland. E-mail: gustav.vonschulthess@usz.ch

Clinically, PET/MR does not arrive into a blue ocean, but into an environment where PET/CT has proven its clinical capabilities in thousands of publications and millions of clinical examinations worldwide.5-8 Because of the relative technical ease of producing clinically effective PET/CT scanners, the focus shifted away from PET/MR, and in 2001, effective PET/CT scanners were introduced for clinical use. The integration of PET and CT has resulted in much data that give us insights into the need for clinical integrated imaging. The data have also helped identify the strengths and weaknesses of PET/CT, the latter being hoped to be partly overcome by PET/MR. In contrast, the technical challenges in implementing PET/MR clearly demonstrate its major shortcomings. Integrated imaging with PET/CT is highly synergistic: the sum 3

0001-2998/13/$-see front matter 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1053/j.semnuclmed.2012.08.005

4 of PET and CT is more than its parts. PET/CT provides clinically critical anatomic correlation, resulting in increased sensitivity and specicity, AC, and an increase in productivity at once. Whether such synergy can be attained with MR remains to be demonstrated.

G.K. von Schulthess et al

them qualitative images only. Somehow, MR data have to be acquired such that they can be readily transformed into quasi CT data and from there into attenuation maps. The major challenge in using MR data for AC of PET data is related to the inability of conventional MR pulse sequences to unequivocally show bone. Various studies have demonstrated that omission of bone and its replacement by soft tissues in attenuation maps lead to erroneous standard uptake values (SUVs) in the brain and in the bones and their vicinity, which differ by as much as 25% from the correct values. This is the current approach taken obtaining body data in integrated systems without CT.2,3,10 Various approaches to overcome this problem are currently under investigation and are subject of another contribution in this issue. Although for simultaneous and 1-room sequential PET/MR systems, the inability to unequivocally visualize bone poses a serious problem regarding clinical utility, the PET/CT-MR sequential system avoids this problem, as in such systems, CT data are available from the PET/CT examination.

Problem 1: Photomultipliers Unt for the Magnetic Environment in MR

Current photomultiplier (PMT) detectors in PET scanners do not function within or near MR scanners. Hence, new detectors have to be designed, such as avalanche photodiode or silicon PMT based PET detector systems. This is discussed elsewhere in this issue. With such detector systems, full integration and simultaneous image acquisition of PET and MR data are possible; such systems are called simultaneous systems.9 Currently, only Siemens provides such systems. Alternatively, PET and MR scanners are placed far enough apart that the magnetic eld does not interfere with the PMTs. In such systems, a shuttle and a transfer board is required to transfer patient from PET(/CT) to MR scanner or vice versa when hardware coregistration of images is to be achieved. Such systems are called sequential systems.9 All PET/CT (and single-photon emission computed tomography [SPECT]/CT) systems are sequential integrated systems, in which hardware image registration is achieved by the patient system table serving as shuttle transfer system. Sequential PET/MR systems have been implemented by GE Healthcare and Philips either as 2-room PET/CT-MR (GE Healthcare) or as 1-room PET-MR (Philips Medical Systems). For completeness sake, it has to be mentioned that in many situations, software fusion of images achieves excellent results, particularly in the brain and heart. Systems not connected by hardware are called separate systems.9

Using a PET/CT-MR to Avoid Technical Issues and Focus on Clinical Data Acquisition
The PET/CT-MR conguration thus has major advantages at this point to address the key issue of clinical PET/MR, namely, to demonstrate the clinical value of this technology relative to PET/CT. As in this setup, PET and MR scanners do not interfere, the best state-of-the-art PMT-based PET and 3-T MR scanners can be connected by a shuttle transfer system and used to acquire clinical data. AC can be done with the CT of the PET/CT scanner, avoiding the use of the still awed ways of obtaining AC maps from MR data for correcting PET. This is a unique feature of PET/CT-MR systems. Artifacts introduced by MR coils can also be avoided. If the surface coils are arranged in such a way around the patient that he/she is inserted into and removed from them when he/she is moved into and out of the MR scanner, no coils are on the patient within the PET(/CT) scanner. This is similar to moving a hand into and out of a glove, and we have termed this the glove approach (Fig. 1). Consequently, the PET/CT-MR 2-room system approach has the unique property to provide hardware-coregistered data sets of PET, CT, and MR, thereby permitting to perform clinical studies, in which PET/CT vs PET-MR data can be subject to blind comparative evaluation. PET-MR congurationswhether fully integrated or shuttle connectedprovide only PET and MR data sets; thus, for comparison, they have to undergo a PET/CT rst and then a PET/MR or vice versa. This results in a PET(1)/CT data set from one system and a PET(2)/MR data set from a second system. As PET(1) and PET(2) scanners differ from each other in make and are acquired at least 30 minutes apart, resulting in some changes in, for example, uorodeoxyglucose (FDG) tissue distribution and related alterations in SUV, a clean comparison of PET/CT and PET/MR data is difcult, as differences observed

Problem 2: Interference of Metallic Structures Used in MR with the Gamma Rays of PET
For best MR image quality, surface coil arrangements are needed. In extended body imaging, this is achieved with array coils. Coils contain various metal parts such as the metallic conductors and the amplier electronics. These cause major artifacts in the PET images and have to be corrected for. As will be seen, this problem can be circumvented in sequential PET/MR systems. Both development of new detectors and modication of surface coils to reduce their metal content are major tasks, which currently are being addressed by the system manufacturers and are tasks that few academic institutions can address successfully.

Problem 3: AC of PET Data by MR Data

As MR does not provide an attenuation map for ionizing radiation, PET data cannot be subject to AC, which makes

Clinical integrated PET/MR applications

5 thumb is that data acquisition on an integrated system using only 1 of the 2 systems should not exceed the sum of the patient upload and download times onto the imaging table, that is, 10-15 minutes.9 The development of much of a workow for integrated PET/MR can already be done using shuttle-connected systems by choosing an MR data acquisition protocol adhering to a prescription in which the table is translated every 2-4 minutes by 15-25 cm.

Figure 1 Patient lying on table with built-in posterior coil is slipped into head coil, and anterior torso coil is placed on patient. All these operations do not interfere with the patients position, thereby maintaining coordinate registration between positron emission tomography (PET), computed tomography (CT), and magnetic resonance (MR) data. (Color version of gure is available online.)

OncologyPartial-/Whole-Body Workow
In PET/CT, usually a low-dose CT without contrast media (CM) is acquired, followed by the PET scan andif required by a dedicated full-organ or region-focused CMenhanced CT. It is unlikely that an oncological PET/MR workow will look much different. Hence, the goal in a simultaneous PET/MR must be to acquire axial MR data matching or surpassing in quality the low-dose CT data during the time the PET scanner acquires PET data in a given bed position. In other words, one needs to develop a low-dose MR protocol. The axial PET scanner FOVs in the existing or evolving fully integrated PET/MR scanners of the major vendors will be 25 cm. Thus, MR data acquisition will need to acquire an adequate axial slice number of approximately 50 slices in 2-4 minutes. Currently, MR can accomplish this task in 1 minute using Dixon-type in- and out-of-phase images (Fig. 2). Compared with low-dose CT images, these images provide a superior soft-tissue contrast, and therefore are suitable for imaging the body from head to pelvic oor, with substantial quality limitations in the lung. For lung imaging, current pulse sequences provide clearly poorer images than CT. Much work is currently being done on ultrashort and zero-echo-time MR imaging.11,12 Such MR pulse sequences with their low sensitivity to tissue susceptibility changesso relevant in the lungmight prove better than those currently available. As such sequences should also be able to image bone with its very short T2 relaxation times, they may also provide MR data, which will be useful to segment bone for AC maps. In addition and as stated previously in the text, a clinical PET/MR workow may contain some additional organ-focused MR data acquisition at the end of the examination, not lasting more than 15 minutes.

in PET/CT vs PET/MR data may not only be due to the 2 technologies used, but could also be due to the different make of the PET scanners or a difference in FDG distribution at the 2 differing imaging times.

Development of a Clinical Workow for PET/MR

Although the PET workow for various clinical imaging studies is usually straightforward, in MR, anything goes, that is, the variety of pulse sequences to be used on a patient is virtually unlimited. In clinical PET/CT, the dominant workow is that of oncologic staging in which 6-10 typically 15-cm axial FOVs are acquired in 2- to 4-minute steps, leading to a partial- or full-body examination. In the brain and the heart, single eld-of-view scans are coupled depending on needwith a dynamic data acquisition, which takes 10-60 minutes. In a fully integrated PET/MR scanner, this PET data acquisition strategy has to be matched: the agile horse of MR has to be in tune with the slowly trotting PET. Hence, PET calls the shot, and the technology is appropriately called PET/MR. If MR is used for imaging sequences much beyond the time it takes to acquire the PET data, a PET/MR scanner ends up being mostly an MR scanner, and imaging cost becomes prohibitive on an integrated system. A rule of

Figure 2 Sixty-ve year old male status post rectal cancer and new liver metastases. (A) Contrast-enhanced CT. (B) Contrast-enhanced PET/CT. (C) Non contrast-enhanced T1 fat-saturation (fat-sat) MR. (D) PET/MR with non contrast-enhanced T1 fat-sat MR. One of the uorodeoxyglucose (FDG)-positive liver metastasis is almost invisible on the contrast-enhanced CT (A, white arrow). On MR and PET/MR (C and D), the FDG-positive lesion can be detected on the non contrast-enhanced T1-weighted image as well as in PET/MR.

G.K. von Schulthess et al

Figure 3 Fourty-six year old male with (A) Contrast-enhanced CT. (B) Contrast-enhanced PET/CT. (C) Contrastenhanced T1 fat-sat MR. (D) PET/MR with contrast-enhanced T1 fat-sat MR. A large right-sided tongue cancer as well as an FDG-positive Rouvire lymph node (white arrow) is detectable. However, the lesion conspicuity (especially the lymph node) in (A) and (B) are inferior compared with contrast-enhanced MRI and PET/MRI.

Brain and Heart Imaging

For static brain and heart imaging, the same considerations probably apply, with the MR data acquisition not exceeding that of the PET data acquisition by more than 15 minutes. In contrast, dynamic/static PET organ data acquisition can take 30-50 minutes and will allow for much more MR data capture. The length of dynamic examinations will prove costly owing to the high cost of a fully integrated PET/MR scanner.

Clinical Data and Potential Uses of PET/MR in Oncology

Various oncological indications and concepts are being discussed in the literature and also evaluated in our department. Firstly, one has to differentiate between whole-body imaging (head to upper thighs or to feet) or partial-body/organ imaging, where PET/magnetic resonance imaging (MRI) will be only done in a specic anatomical area. We will discuss some major examples, with no entitlement to completeness. One indication always mentioned as natural for PET/MRI is the visualization of liver lesions (primary and metastatic). Based on the relatively high liver 18F-FDG-activity in PET/ CT, the detection of focal liver lesions with only slightly elevated FDG activity can be impaired, and the intrinsically lower soft-tissue contrast of CT compared with MRI in the liver is well known. Contrast media in CT can increase detection rates for liver lesions, but generally MRI is still considered superior. However, there are only few studies evaluating contrast-enhanced (CE) PET/CT compared with MRI, and a large meta-analysis did not nd statistically signicant differences in lesion detection.13 Initial comparison of data acquired in our PET/CT-MRI system shows no difference in lesion detection when comparing CE PET/CT and non-CE MRI. Lesion conspicuity was signicantly better on the MRI component when compared with the CE CT component14 (Fig. 2). Furthermore, comparisons of standard low-dose PET/CT and non-CE PET/MRI showed signicantly better results concerning lesion conspicuity in the liver.15 It is obvious that primary liver malignancies will also prot from PET/MRI imaging. For example, hepatocellular carcino-

mas are always evaluated by CE MRI. In contrast, large trials have indicated that detection and characterization can be signicantly improved when imaged with 18F-choline PET/ CT.16 Thus, the combination of other tracers than 18F-FDG with MRI imaging may improve imaging diagnostics in liver lesions beyond current stand-alone PET/CT and MRI, and overall PET/MRI even partially without CMmight be more accurate in the diagnosis and characterization of liver lesions. Another indication requiring only regional extent imaging and expected advantages of PET/MRI over PET/CT is head and neck cancer. In these tumors, surrounding soft-tissue inltration is important for local staging as well as surgical and radiotherapy planning (Fig. 3). Because of its higher soft-tissue contrast, MRI has become the most signicant morphologic modality for imaging head and neck cancer. However, PET/CT has proven to be superior for the detection of lymph node metastases in head and neck cancer.17 Because MRI and PET/CT are currently used for primary staging and restaging in many of these patients, a combined PET/MRI here is likely resulting in a one-stop shop. PET/MRI will have an additive value in patients with metallic dental implants, as existing and new MRI sequences can minimize metal dental artifacts, which can impair the diagnosis on CT signicantly. Examples and potential indications for whole-body PET/ MRI are manifold. Several publications already compared, for example, PET/CT vs diffusion weighed imaging (DWI) in lymphoma patients.18 Most of the studies currently available generally found satisfying agreement when evaluating the diagnostic accuracy of those 2 methods. MRI publications also indicated that the DWI-MRI might be sufcient for follow-up on this patient population, especially when evaluating the apparent diffusion coefcient (ADC) values. However, at this point, there is no established routine clinical staging and therapy follow-up protocol using DWI-MRI in this setting. Thus, in PET/MRI, FDG uptake data (SUV) and tumor lesion glycolysis as well as the diffusion data and ADC values can be combined and correlated. A conceivable future approach for primary staging and for follow-up in this patient group would be to have PET/MRI for primary staging and

Clinical integrated PET/MR applications

only DWI-MRI imaging for follow-up. This could substantially reduce radiation dose from imaging, but rst, it has to be demonstrated that the high sensitivity of DWI coupled with relatively low specicity will result in an accurate enough examination. Staging of bronchial carcinoma might also become an important indication for PET/MRI. Currently, CT provides unparalleled image quality and tissue contrast in the lung and is worldwide the method of choice for lung nodule detection and evaluation of parenchymal texture. MR is not even close in matching this ability of CT in lung imaging, even with breath-hold gradient recalled echo sequences or gating. However, patients with the suspicion of lung cancer referred to PET/CT or PET/MRI have already undergone almost always a chest CT. Thus, with a chest CT available, it might be less important that MR within the context of a PET/MR examination provides accuracy comparable with CT if the rest of the information gained by PET/MR is superior to that of PET/CT. There are initial promising results for PET/MRI at least for bronchial carcinoma detection,19 but CT in PET/CT can certainly be no match to brain imaging afforded by PET/ MR, and MR is recommended for staging in all patients with stage II tumors.20 In fact, in this setting, it probably makes no difference if a PET/CT and an MR or a CT and a PET/MR are acquired, as all modalities are needed for proper staging. It is noteworthy that our current PET/CT-MRI setting is able to provide all those staging steps within one imaging procedure. Here, the CT from PET/CT can be used for state-of-the-art lung imaging, the PET/CT provides the standard wholebody staging, and MRI can be performed in the brain and/or any other area of interest (eg, liver). Overall, we do see several advantages in PET/MRI imaging in oncologic indications in imaging of body regions/organs as well as in partial-/whole-body imaging. Currently, the PET/CT-MRI approach offers the highest exibility to provide complete staging concepts. To dene where PET/MRI can be used to better advantage compared with PET/CT, PET/MRI will have to be dened in sizable patient series for various disease entities.

7 PET/MR in clinical routine does not preclude the usefulness of an integrated scanner. To read PET and MR images and integrate the complementary information of the 2 modalities, both scans should take place at roughly the same time, at least on the same day. This is often impossible because of having 2 separate scanners in separate imaging divisions (neuroradiology/ nuclear medicine) leads to scheduling conicts. It is disagreeable for the patient to have 2 separate appointments and laborious for the referring physician to coordinate 2 examinations. Furthermore, to promote combined PET/MR neuroimaging, the multiplatform image acquisitions have to be perceived as 1 examination. In clinical routine, it is desirable to provide only 1 medical report for the referring physician at the end. Hence, integrated PET/MRI has a good potential to become a powerful tool for quantitative imaging and objective decision making as a one-stop-shop examination of the brain. On the technical side, the integration of PET and MR, rather than PET and CT, allows for a dose reduction to the patient by omitting the CT scan. In PET imaging, the CT is in most cases only used for AC, which in the future probably can also be accomplished by MR. For implementing automated image analysis tools to quantify PET data, it is helpful to have MR images always from the same MR device, as this increases the robustness of the processing algorithm. Nowadays, much of quantitative image analysis is executed by transforming the MR images into stereotactic space using the Montreal Neurological Institute/International Consortium of Brain Mapping standard brain template. In a second step, this calculated transformation matrix is then applied to the PET images. Using atlases or other predened regions of interest allows reproducible and user-independent quantitative data analysis. Automated quantitative analysis of brain pathologies is of major importance for the standardization of treatment and therapy assessment. Furthermore, partial volume correction of the PET data based on MR images might increase the exactness of radiotracer quantication. Hence, although often there is no absolute need for an integrated PET/MR scanner, there are important arguments why such an integrated system may improve data quality in brain imaging. Multimodal PET/MR neuroimaging has various interesting applications, that is, in neuro-oncology, neurodegeneration, or epilepsy.

Potential Applications in Neurology

In neuroimaging, morphologic images are often a prerequisite for processing and analysis of functional data. Furthermore, data from PET and MR are often complementary, and as a consequence, the integration of the information from both modalities yields the most accurate diagnosis. This can and has been achieved by software image fusion. Hence, an integrated PET/MR is often not mandatory in clinical multiparametric neuroimaging. Strictly speaking, simultaneous PET and MR imaging is only necessary if the visualized functional processes are time dependent in PET and MR. Examples are functional neuropsychiatric examinations measuring task-induced changes of blood ow and oxygen consumption rate or the measurement of the arterial input function by MRI for its use in compartment modeling in quantitative PET studies. The rare physiological necessity of simultaneous

Neuro-Oncology
In vivo visualization of molecular targets inuences clinical decision making and outcome. Brain tumors are often heterogeneous, and stereotactic biopsies of the highest tumor grade is crucial for further treatment planning.21 It is known that hypoxic areas in tumor tissue are more resistant to radiotherapy, and that the local microenvironment strongly inuences the effectiveness of chemotherapy.22 Therefore, to determine the most effective therapy, it is important to take relevant imaging biomarkers of PET and MR into consideration, that is, cell proliferation, metabolic activity, cellular density, cell death, growth factor expression, vascular structure, neoangiogenesis, oxygenation status, pH, and drug delivery.

G.K. von Schulthess et al

Figure 4 (A) Pittsburgh compound B is a specic radiotracer for -amyloid plaques. (B) FDG visualizes hypometabolism and atrophy. In Alzheimers disease, FDG is typically reduced in the posterior cingulate gyrus, as well as in the parietal and temporal cortex. (C) Vascular dementia. The diagnosis is based on clinical and radiologic ndings summarized in the National Institute of Neurological Disorders and Stroke/Association Internationale pour la Recherche et lEnseignement en Neurosciences criteria.

Neurodegenerative Disease
In Alzheimers disease, an early and accurate detection of amyloid plaques by PET is of major importance, as there is evidence that these deposits provoke the regional neurodegeneration. For the evaluation of disease progression and therapy response, integrating morphologic and functional information from PET and MR is mandatory. Interesting biomarkers are, for instance, the metabolic activity measured in PET23 and the functional connectivity as measured by resting-state functional MRI.24 Morphologic images are essential to exclude other causes of dementia like multi-infarct dementia or normal-pressure hydrocephalus (Fig. 4). In movement disorders, morphologic imaging does not allow for early differentiation of Parkinson disease from atypical parkinsonian syndromes. However, the addition of functional studies permits to recognize degeneration patterns specic to Parkinson disease, multiple system atrophy, progressive supranuclear palsy, and corticobasal degeneration.25 Once again, to evaluate the effectiveness of treatment, quantitative imaging is crucial.

Epilepsy
Patients with refractory epilepsy despite medical treatment may undergo surgery to resect the epileptogenic focus. Before surgery, it has to be veried that the suspected focus indeed is causing the epilepsy. Hence, for optimal therapy outcome, it is required to combine the available localization techniques, that is, morphologic MRI, functional MRI, PET, and EEG.26

Clinical Data and Potential Uses of PET/MR in Cardiology

Nuclear myocardial perfusion imaging (MPI) offers well-established tools for the assessment of coronary artery disease, and a large body of literature underlines the strong prognos-

tic value of nuclear MPI. Although superiority of PET over SPECT MPI has been suggested for accuracy and can be expected regarding study duration, staff radiation exposure, reduced patient dose, improved patient comfort, and potentially economic benet, in reality, PET still plays a limited role in daily clinical routine for MPI compared with SPECT. This may be due on the one hand to the lack of solid data to support a superior accuracy of PET over SPECT and on the other hand to the limited availability of PET for cardiac studies. Although with the widespread use of oncology PET scanning, the availability of PET scanners has dramatically increased, this was not yet paralleled by a substantial rise in cardiac PET applications. The main drawback has been the lack of a ow tracer for PET MPI, with a half-life long enough to allow for shipment to satellite PET centers without cyclotron. This may change with the advent of new 18F-tracers such as 18F-urpiridaz,27 and clinical phase 3 trials are under way, but results are not yet available. Although PET is the most reliable noninvasive tool for identication of myocardial viability in ischemic dysfunctional left ventricles and the only method validated in a prognostic interventional outcome study meeting all criteria of solid comparative effectiveness research,28 its clinical role has remained modest for several reasons. One reason is that there are competing radiation-free methods in the hands of cardiologists, such as echocardiography and MRI, which allow identication of viability. Another reasonpotentially with even more impactis recent controversial results of the Surgical Treatment for Ischemic Heart Failure trial,29 which have questioned the clinical benets of viability assessment on patient outcome in ischemic left ventricular dysfunction.30 It is in this context that the potential importance of a sequential cardiac PET and MR scanning or a combined PET/MR imaging tool for cardiac application has to be discussed. Several years ago, the introduction of hybrid imaging systems with multislice CT and PET devices, which is now

Clinical integrated PET/MR applications

considered the imaging modality of choice for a broad clinical spectrum, was primarily driven by oncological indications. The interest in cardiac hybrid imaging is increasing, as this allows rapid and comprehensive evaluation of anatomic lesion severity (CT coronary angiography), with its functional relevance with regard to perfusion in the subtended myocardial territory. Also, in cardiac imaging, the perfect match of functional and anatomic information is probably less crucial than in other body areas; however, in several clinical situations this is important, such as in patients with multivessel disease or after bypass surgery, and it may confer an added diagnostic and prognostic value and have an impact on patient management.31,32 In less advanced coronary artery disease stages, however, no ow-limiting lesions are present due to eccentric vessel remodeling and, potentially, rupturing of a vulnerable plaque. More than 60% of cases with sudden cardiac death originate from rupture of such vulnerable plaque, characterized by several features, including a thin brous cap, large lipid core, and the presence of active inammation. Although CT angiography may assess the overall vascular risk burden, identication of vulnerable plaques is currently impossible with this technique. MRI is a versatile tool for investigating many aspects of cardiovascular disease, offering evaluation of cardiac structural disease, myocardial function, thrombus formation after infarction, and detection of myocardial brosis. A modern application of MR is the evaluation of the vascular wall and the identication of plaques prone to rupture. One of the strengths of MR is its high anatomic resolution, which, in combination with the high sensitivity of PET for the detection of molecular targets, may help to explore evolving and investigational elds of cardiovascular imaging, such as vulnerable plaque characterization, in vivo stem cell tracking, and assessment of angiogenesis (vasa vasorum of vulnerable plaques). Such applications are highly investigational so far, and a direct implementation in the daily clinical routine is not around the corner. However, the importance of these potential applications lies in the fact that perfect alignment and therefore a hybrid device seems mandatory. A more clinical application is the complementary use of PET with FDG and MR with gadolinium late enhancement, including the high-resolution assessment of regional contractility from MR, although the superiority of a hybrid approach over a combination of data sets obtained from stand-alone devices may be limited. Whether the currently less well accepted clinical importance of viability detection will justify such technological and nancial efforts remains to be evaluated. The combined availability of PET and MR may also help to evaluate in a proper study the relative merits and the comparability of myocardial perfusion with both techniques, and the temporal simultaneity afforded by a combined device would allow coinjection of both PET tracer and MR contrast medium for a solid head-to-head comparison. Thus, although in oncology imaging, a combined PET/MR may have a great potential, a clinical utility in cardiac imaging seems currently less obvious.

Conclusions
Clinical PET/MR will have to establish itself in a world where PET/CT reigns. Even assuming that all current technical issues in PET/MR will be overcome, the huge clinical task of dening in which clinical applications PET/MR may be superior to PET/CT has hardly begun. Any statements on the clinical utility of PET/MR are thus speculative. Nevertheless, our group believes that PET/MR used under strict workow protocol may eventually become an excellent and even costeffective imaging tool in many indications. As a simultaneous implementation, it combines the value of a one-stop shop with simultaneous data acquisition, which roughly cuts imaging time in half, and requires a single patient appointment. This is indeed in itself an excellent start bonus for the patient.

Acknowledgments
This work was supported in part by grants from the canton of Zurich, several Swiss foundations and a research grant from GE Healthcare.

References
1. Shao Y, Cherry SR, Farahani K, et al: Simultaneous PET and MR imaging. Phys Med Biol 1997;42:1965-1970 2. Hofmann M, Bezrukov I, Mantlik F, et al: MRI-based attenuation correction for whole-body PET/MRI: Quantitative evaluation of segmentation- and atlas-based methods. J Nucl Med 2011;52:1392-1399 3. Martinez-Mller A, Souvatzoglou M, Delso G, et al: Tissue classication as a potential approach for attenuation correction in whole-body PET/ MRI: Evaluation with PET/CT data. J Nucl Med 2009;50:520-526 4. Hofmann M, Steinke F, Scheel V, et al: MRI-based attenuation correction for PET/MRI: A novel approach combining pattern recognition and atlas registration. J Nucl Med 2008;49:1875-1883 5. Hany TF, Steinert HC, Goerres GW, et al: PET diagnostic accuracy: Improvement with in-line PET-CT system: Initial results. Radiology 2002;225:575-581 6. Von Schulthess GK, Hany TF: Imaging and PET-PET/CT imaging. J Radiol 2008;89:438-447; quiz 48 7. Maziak DE, Darling GE, Inculet RI, et al: Positron emission tomography in staging early lung cancer: A randomized trial. Ann Intern Med 2009; 151:221-228; W-48 8. Ben-Haim S, Ell P: 18F-FDG PET and PET/CT in the evaluation of cancer treatment response. J Nucl Med 2009;50:88-99 9. von Schulthess GK, Burger C: Integrating imaging modalities: What makes sense from a workow perspective? Eur J Nucl Med Mol Imaging 2010;37:980-990 10. Samarin A, Burger C, Wollenweber SD, et al: PET/MR imaging of bone lesionsImplications for PET quantication from imperfect attenuation correction. Eur J Nucl Med Mol Imaging 2012;39:1154-1160 11. Keereman V, Fierens Y, Broux T, et al: MRI-based attenuation correction for PET/MRI using ultrashort echo time sequences. J Nucl Med 2010;51:812-818 12. Catana C, van der Kouwe A, Benner T, et al: Toward implementing an MRI-based PET attenuation-correction method for neurologic studies on the MR-PET brain prototype. J Nucl Med 2010;51:1431-1438 13. Niekel MC, Bipat S, Stoker J: Diagnostic imaging of colorectal liver metastases with CT, MR imaging, FDG PET, and/or FDG PET/CT: A meta-analysis of prospective studies including patients who have not previously undergone treatment. Radiology 2010;257:674-684 14. Assiri Y, Schmid D, Pietsch C, et al: Contrast-enhanced PET/CT vs. non-contrast enhanced PET/MRI in abdominal oncological lesions Initial results. Annual Congress of the European Association of Nuclear Medicine, Milan, Italy, 2012

10
15. Crook DW, Mader C, Kuhn FP, et al: Diagnostic performance of PET/MR versus PET/CT in the abdomen. Presented at: European Congress of Radiology, Vienna, 2012 16. Talbot JN, Fartoux L, Balogova S, et al: Detection of hepatocellular carcinoma with PET/CT: A prospective comparison of 18F-uorocholine and 18F-FDG in patients with cirrhosis or chronic liver disease. J Nucl Med 2010;51:1699-1706 17. Hustinx R, Lucignani G: PET/CT in head and neck cancer: An update. Eur J Nucl Med Mol Imaging 2010;37:645-651 18. van Ufford HM, Kwee TC, Beek FJ, et al: Newly diagnosed lymphoma: Initial results with whole-body T1-weighted, STIR, and diffusionweighted MRI compared with 18F-FDG PET/CT. Am J Roentgenol 2011;196:662-669 19. Schwenzer NF, Schraml C, Mller M, et al: Pulmonary lesion assessment: Comparison of whole-body hybrid MR/PET and PET/CT imagingPilot study. Radiology 2012;264:551-558 20. NCCN. Clinical Practice Guidelines in Oncology. NCCN. Available at: http://www.nccn.org/professionals/physician_gls/f_guidelines.asp. Accessed October 12, 2012 21. Asselin MC, OConnor JP, Boellaard R, et al: Quantifying heterogeneity in human tumours using MRI and PET. Eur J Cancer 2012;48:447-455 22. OConnor JP, Jackson A, Asselin MC, et al: Quantitative imaging biomarkers in the clinical development of targeted therapeutics: Current and future perspectives. Lancet Oncol 2008;9:766-776 23. Ossenkoppele R, Tolboom N, Foster-Dingley JC, et al: Longitudinal imaging of Alzheimer pathology using [11C]PIB, [18F]FDDNP and [18F]FDG PET. Eur J Nucl Med Mol Imaging 2012;39:990-1000 24. Kenny ER, Blamire AM, Firbank MJ, et al: Functional connectivity in

G.K. von Schulthess et al

cortical regions in dementia with Lewy bodies and Alzheimers disease. Brain 2012;135:569-581 Teune LK, Bartels AL, de Jong BM, et al: Typical cerebral metabolic patterns in neurodegenerative brain diseases. Mov Disord 2010;25: 2395-2404 Duncan J: The current status of neuroimaging for epilepsy. Curr Opin Neurol 2009;22:179-184 Yu M, Nekolla SG, Schwaiger M, et al: The next generation of cardiac positron emission tomography imaging agents: Discovery of urpiridaz F-18 for detection of coronary disease. Semin Nucl Med 2011;41:305313 Beanlands RS, Nichol G, Huszti E, et al: F-18-uorodeoxyglucose positron emission tomography imaging-assisted management of patients with severe left ventricular dysfunction and suspected coronary disease: A randomized, controlled trial (PARR-2). J Am Coll Cardiol 2007; 50:2002-2012 Bonow RO, Maurer G, Lee KL, et al: Myocardial viability and survival in ischemic left ventricular dysfunction. N Engl J Med 2011;364:16171625 Cortigiani L, Bigi R, Sicari R: Is viability still viable after the STICH trial? Eur Heart J Cardiovasc Imaging 2012;13:219-226 Pazhenkottil AP, Nkoulou RN, Ghadri JR, et al: Prognostic value of cardiac hybrid imaging integrating single-photon emission computed tomography with coronary computed tomography angiography. Eur Heart J 2011;32:1465-1471 Pazhenkottil AP, Nkoulou RN, Ghadri JR, et al: Impact of cardiac hybrid single-photon emission computed tomography/computed tomography imaging on choice of treatment strategy in coronary artery disease. Eur Heart J 2011;32:2824-2829

25.

26. 27.

28.

29.

30. 31.

32.