PET/MR: Finding its Niche
The final diagnostic form the evolving hybrid PET/MR tree will take is uncertain, but it has already begun to blossom. In fact, low-hanging clinical fruit has lured a number of major medical centers to early adoption.

The attraction is combined PET and MRI images, with the potential of opening new diagnostic opportunities. Fused PET and MRI data, for instance, allow temporal registration and visualization of diseased tissue as it proliferates over time at the molecular level. This is expected to help identify various disease states sooner.      

Despite the potential clinical roles, the high cost of PET/MR units, as well as the increased size of the machine, could inhibit many hospitals from embracing the technology as soon as they may wish.  

Frost & Sullivan estimates the cost of a PET/MR hybrid to be between $2.5 and $3 million, slightly higher than the $1.9 to $2.4 million price tag for a PET/CT scanner.

Best of both worlds

PET/MR appeared on the diagnostic scene nearly a decade after PET/CT. Design changes in PET and MRI scanners were necessary to overcome a number of obstacles; for instance, MRI magnets distorted images from nearby PET scanners, and there wasn't space enough inside an MRI system to house a PET scanner.

Among the design changes, LCD workstations prevent image distortion and higher-performing gradients free up enough space to accommodate a PET inside the MR gantry, while at the same time widening the bore. FDA approval in June paved the way for clinical use in the U.S.

The fusion of PET and MRI images and data will capitalize on the strengths of each, providing a hybrid technology that is greatly superior to the sum of its parts, says Habib Zaidi, PhD, head of the PET instrumentation and neuroimaging laboratory at Geneva University Hospital.

With PET/MR, MRI images provide anatomy and tissue characterization combined with metabolic imaging obtained from PET, making it possible to see tissue function and metabolism.

"More importantly, using simultaneous rather than sequential scanning will enable the resolution of many of the impediments to precise co-registration of anatomo-molecular information," Zaidi says.

The immediate impact of PET/MR will likely be most significant in oncology, where the combination of PET and MR in one imaging device offers advantages over conventional imaging modalities. Oncologic application of PET/MR allows imaging of the four main processes in cancer formation: apoptosis resistance, angiogenesis, proliferation and metastasis.

"PET/MR will take functional characterization of tissue to another level, and allow one to marry the beautiful soft-tissue contrast of MRI with the tissue characterization of PET," says Vikas Gulani, MD, PhD, director of MRI at Case Western Reserve University in Cleveland.

Moreover, Gulani says, study contrasts will not necessarily be limited to customary FDG. The goal is to produce contrast agents that will light up in particular disease or physiological states.

"While molecular imaging with MR is of major interest worldwide, particularly because of the high tissue contrast of MR, it is made difficult by the inherently low sensitivity of MR," Gulani says. It is anticipated that PET/MR will allow radiologists to combine the best of both worlds in molecular imaging, beyond FDG.

In the long run, various molecular imaging possibilities could open up, Gulani says.

"The limited role of PET/CT in certain clinical indications, such as central nervous system disorders, orthopedic infections, inflammatory disorders and in the evaluation and follow-up of metastatic disease, is well-established," Zaidi says. "The better soft-tissue contrast observed on MRI is also well-established and emphasizes the ineffectiveness of PET/CT for this indication, as well as the potential role of PET/MR."

Constellation of configurations

Breast cancer. Source: Geneva Hospital/Philips Healthcare
PET/MR has been rolled out in three different design configurations.

In one design, the PET and MR systems are operated separately, but are positioned in the same room (Ingenuity TF PET/MR, Philips Healthcare), about three meters apart at opposite ends of a patient turntable. The patient, fitted with MR coils, is first scanned by the 3T system, then withdrawn, rotated 180° and slid into the PET bore. This allows for sequential acquisition and later fusion of data.

"The advantage is the patient will be in the same reference frame when imaged in PET and in MRI, so you don't need to conduct image registration per se," says Osama Mawlawi, PhD, of the department of imaging physics at MD Anderson Cancer Center in Houston. Combining PET and MRI in one single, superimposed image is an improvement over current methods of acquiring images at different times and then attempting to fuse them later, he adds.

Mawlawi says the disadvantage with this configuration is both instruments are occupied to scan one patient sequentially, potentially inhibiting throughput. When the patient is being imaged by one machine, the other machine is standing idle.

The second design is a tri-modality configuration (Discovery PET/CT+MR, GE Healthcare), where the PET/CT 690 machine is in one room and the MR750 in an adjoining room. The transport table is key here also, as the patient is positioned only once between exams. As with the Philips model, software then registers and fuses the datasets from the two exams.

While configurations one and two do not offer true simultaneous PET/MR acquisition, they do provide researchers and clinicians reasonable PET/MR viability at a somewhat reduced cost.  

The third design is simultaneous PET/MR (Biograph mMR, Siemens Healthcare), in which a PET ring detector fitted inside a 3T magnet combine to acquire MR anatomical data and PET functional data simultaneously.

"Simultaneous imaging might open a new area of research whereby a radiologist can view function and structure at the same time or validate function using two different modalities," Mawlawi says.

Still, he notes the promise of PET/MR has yet to be realized, that the question everyone is trying to answer is what unmet clinical need PET/MR addresses. "Vendors, as well as most clinical researchers, are trying to assess what PET/MR can provide that justifies its development," Mawlawi says. "From a clinical perspective, we do not yet have a killer application for PET/MR."

One possible PET/MR application would be looking at prostate cancer because of the difficulty in pinpointing malignancies in this area. When conducting PET imaging using standard FDG, Mawlawi says it is difficult to see uptake in the prostate region because most of the FDG compound ends up in the bladder. "You need to be able to differentiate the uptake in the bladder and the uptake in the prostate and that's not simple to do," he says.

Potential applications

While PET/MR technology may still be in an embryonic stage, advances are imminent on multiple fronts, specifically cardiology and oncology.

As far as cardiology applications, PET is the current gold standard for non-invasive assessment of myocardial viability and allows accurate detection of coronary artery disease by assessment of myocardial perfusion. MRI provides high-resolution anatomical images that allow evaluation of ventricular structure and function together with detection of myocardial infarction. Hybrid PET/MR imaging may facilitate the characterization of atherosclerotic plaques and aid in the evaluation of angiogenetic or stem cell therapies.

Heinz-Peter Schlemmer, MD, of the German Cancer Research Center and chair of radiology, University of Heidelberg, says the bottleneck of diagnostic progress will likely be the development of new radiotracers. "PET alone does not produce images, it only displays what has been injected."

PET/MR could deliver an advantage with respect to radiation exposure. "While PET/CT is faster and cheaper, MR provides better soft-tissue contrast, plus functional information and what is particularly important, works without radiation exposure to the patient," Schlemmer says. PET/MR produces up to 70 percent less ionizing radiation than PET/CT.

PET/MR is thus indicated where there is concern about radiation risk, such as pediatrics, where studies can be obtained without the higher radiation risk associated with CT. Another use for PET/MR would be areas of the body—pelvis, head and neck—where CT provides suboptimal tissue contrast.

"The capability to use CT as an anatomical reference is compromised by its relatively poor soft-tissue contrast," says Bruce Rosen, MD, director of the Martinos Center for Biomedical Imaging at Massachusetts General Hospital in Boston. "MR-based soft tissue contrast is superior to CT, so it's actually a better anatomical reference."

Another area ripe for PET/MR is lesion characterization where the contrast properties of MR are superior to that of CT, including the breast and liver, Rosen says.

"Often, patients imaged with PET/CT also are being sent for an MRI because of MR's ability to provide additional tissue characterization," Rosen says. This duplication could change with PET/MR, saving time, money and patient inconvenience.

For example, in patients with rectal carcinomas, Rosen says it may be possible to complete staging in a single PET/MR exam. "We'll be able to look for nodal involvement and see sufficient detail of local tissue spread in a single exam that combines PET and MR in a way that we really can't assess with a single PET/CT exam," Rosen says. "We often end up having to refer those patients for additional studies with MRI."

Rosen anticipates more advanced physiological applications for PET/MR, which would evaluate antiangiogenic effects, where the capability to perform metabolic imaging with PET and physiological imaging with MR are naturally complementary to understanding the potential long-term treatment response of novel therapeutic regimes.

"Oncologic applications are good areas for that," Rosen says. Other applications where metabolic information from PET combined with the physiologic information of MR will be complementary are areas of neuroplasticity and stroke recovery, as well as cases of epilepsy and traumatic brain injury, circumstances where advanced physiological and biophysical properties of MR imaging, when combined with direct quantitative metabolic imaging of PET, could be extremely helpful.

"These aren't studies that are routinely being conducted today with PET, but the combination of the two technologies are extremely synergistic and could be quite valuable," Rosen says.

Zen radiology?

When the disease is ready, the modality appears. The biggest advances for PET/MR may come in neuroscience, neurobiology and neurological disease, specifically for Alzheimer's disease. PET/MR arrived just as a National Institutes of Health study group changed the definition of Alzheimer's from a clinical syndrome to a biological phenomenon. Rosen predicts significant implications for PET/MR.

The new modality could combine amyloid imaging agents with quantitative brain morphometry, and perhaps, functional resting-state studies, as a way to provide a single exam in patients at the earliest stages of Alzheimer's.

"Here, there's an opportunity for radiologists to be able to define for clinicians the biochemical disease prior to the loss of working brain," Rosen says. "PET/MR is ideally situated to make a contribution."

That leads to several other potential applications in the neurologic and psychiatric disease domains.

Brain abnormalities are understood more often as abnormalities of brain networks. PET/MR may provide the capability to pinpoint abnormalities in these networks, then provide a map to target a pharmacological treatment or to intervene with direct deep brain electrical stimulation of certain key nodes.

"The opportunity to conduct receptor-based, targeted imaging with PET, combined with functional connectivity mapping with either resting-state MR or diffusion tractography or some combination really becomes potentially a paradigm changer in terms of the way we assess, diagnose and ultimately treat mental illnesses," Rosen explains.

He says the required fundamental capabilities are present in PET/MR technology.

"Emerging neuroscience suggests that that's going to be an important future direction [for] this technology," Rosen says.

As researchers and early adopters strive to define and refine applications, it seems clear that PET/MR could make a mark in multiple clinical areas and offer clinical and efficiency improvements for targeted applications.