MRI’s Expanding Frontiers: Scanning More, Scanning Faster

Numerous recent advancements in MR technology and technique are offering promising prospects for research and clinical use. The growth of PET/MR, parallel imaging and whole-body MRI means the modality is more versatile and efficient, but what does that mean for patient care?

Reaching the crest of the MRI wave

Hybridity is a beautiful thing, as it creates products that offer the best of both worlds. PET/MR scanners blend two imaging technologies into one, illustrating the many benefits that come from melding medical modalities.

Researchers at Washington University in St. Louis installed one of the first PET/MR scanners in the U.S. at an affiliating hospital in 2012. Unlike some other PET/MR scanners, this one is unique in that it simultaneously performs the two types of scans. This benefits both patients and referring physicians, says Bob McKinstry, MD, PhD, director of the Center for Clinical Imaging Research and Washington University School of Medicine radiology and pediatrics professor.

“PET/MR is a win for the patient and a win for the referring physician,” says McKinstry. Simultaneous PET/MR provides a 40 percent time reduction over separately acquired scans. Referring physicians often have the difficult task of scheduling multiple scans, particularly for pediatric patients. Children generally need sedation, and available anesthesiologists can be hard to find. PET/MR scanners help alleviate this and other pediatric patient issues by reducing radiation exposure, simplifying scheduling, and curtailing serial sedation.

Not only do PET/MR scanners profit patients and physicians, but they improve research methods as well. The hybrid modality allows for more creative and complex clinical research, according to McKinstry.

“This marriage goes beyond convenience and efficiency,” he explains. “MR provides superior soft tissue contrast. Simultaneous acquisition improves anatomical registration, which ultimately provides a better look at the cancer.”

Most experience thus far with PET/MR at the Washington University School of Medicine has been in screening the pelvis for cervical cancer. In the future, McKinstry hopes the combination will look at other malignancies. The duo could sub-characterize parts of tumors, providing a deeper level of characterization in areas like the brain, breast, and pelvis.

PET/MR hybrids also could eventually become the standard imaging technique for head and neck cancers. McKinstry points to brain cancer and cardiac imaging as two key directions in which the technology is headed next.

The number of scanners currently installed in the U.S. is relatively modest, with about 10 to 12 available nationwide, says McKinstry. While they’ve been acquired largely for research in the U.S., their adoption has been much more rapid in Europe where they are used for clinical purposes on a daily basis.

In the U.S., FDA approval for PET/MR scanners in 2011 came quickly after their advent. Traditional research is still needed to convince physicians of its clinical merits compared with other modalities.

“We are just at the start of seeing what PET/MRI scanners can do. The wave is just beginning to crest,” says McKinstry

Picking up the pace

While hybrid imaging looks to improve physiological insights, parallel imaging (PI) aims to make MRI scanners more efficient. PI makes quicker image acquisition possible, offering numerous new MR applications for cardiothoracic, abdominal, cardiovascular, and renal imaging. The speed boost is the result of collecting a reduced amount of data in k-space (the MR data space) and using an array of receiver coils to fill in missing data. This enables shortened breath-hold times for patients, and thus fewer motion-corrupted exams.

“Parallel imaging makes possible a huge jump in clinical practice, where imaging can be extended to body parts that could not be imaged without significant reductions in imaging times,” says Vikas Gulani, MD, PhD, of the Case Western Reserve University in Cleveland.

Gulani, and colleagues published a thorough review of PI in the July 2012 issue of the Journal of Magnetic Resonance Imaging. The amount of data collected in k-space is reduced, the authors explain, and thus the “undersampled” data are gathered more quickly. This undersampling, however, leads to aliasing, or repeated representations of the image.

This is where parallel imaging algorithms come into play. They reconstruct artifact-free images either directly in the image space by employing the sensitivity encoding (SENSE) method, or directly from the undersampled data in k-space, which is most commonly performed by generalized autocalibrating partially parallel acquisitions (GRAPPA).

SENSE and its variations are used to accelerate MRI scans, but coil sensitivity maps are required. In contrast, GRAPPA doesn’t require a coil sensitivity map, so it also can be used in situations where it’s hard to obtain a coil sensitivity map. This often occurs in images with areas of low signal and in regions where patient motion is common. Because GRAPPA is such a robust method, it is the more frequently used of the two algorithms.

SENSE and GRAPPA can be applied along with almost any sequence, reducing overall scan time or improving image quality.

Parallel imaging does not come without its challenges, though. The techniques need to be understood and applied in clinical practice, which may be no simple task.

“Parallel imaging techniques are complicated reconstruction methods that make use of complex MR technology and signal processing to improve imaging,” says Gulani. “Using PI in a creative and clinically efficacious manner requires knowledge and skill at operating the magnet. MR can be rather operator dependent because there are so many parameters the user can control. PI adds yet another layer of complexity.”

Other negatives to be aware of included degraded signal-to-noise ratio and residual aliasing artifacts. The need for speed must be balanced against these challenges, calling for constant consideration of the tradeoffs when determining the parameters in a PI accelerated image.

There is a growing movement toward using PI for non-Cartesian, or more complex, data collection strategies. Non-Cartesian methods are not commonly used in the clinic, but there is an increasing emphasis on the clinical utility of these methods.

Non-Cartesian PI will be especially advantageous in settings where exceptionally high speed imaging is a critical requirement—such as cardiac, abdominal and vascular imaging.

“Applying non-Cartesian parallel imaging to open clinical problems such as real-time cardiac imaging, or free-breathing abdominal imaging is an extremely attractive area of very active research,” Gulani says.

PI’s future also holds the potential to combine the technique’s benefits with those of compressed sensing, leading to very fast imaging without artifact setbacks from both.

Sweetening the future for diabetic patients

While PI provides a faster snapshot of specific sites, whole-body MRI takes a broader view, and has recently shown its ability to assess risk of heart attack and stroke in diabetic patients. Fabian Bamberg, MD, MPH, of Ludwig Maximilian University of Munich, and colleagues designed a study aimed to establish the predictive value of whole-body MRI for the occurrence of cardiovascular and cerebrovascular events in a cohort of patients with diabetes mellitus (DM). The study was published online Sept. 10, 2013, in Radiology.

Diabetic patients are known to develop atherosclerosis, which is the thickening of the arterial walls at an accelerated rate. This can result in a higher rate of major adverse cardiac and cerebrovascular events (MACCE).

The authors followed up with phone interviews for 61 patients with DM types 1 and 2 who underwent comprehensive, contrast-enhanced whole-body MR imaging protocol, including brain, cardiac, and vascular sequences at baseline. The primary endpoint was a MACCE, such as composite cardiac-cerebrovascular death, myocardial infarction, cerebrovascular event, or revascularization. Fourteen of the 61 patients suffered a MACCE during the follow-up period.

MR images were assessed for the presence of systemic atherosclerotic vessel changes, white matter lesions, and myocardial changes. The researchers found that while subjects without any pathologic findings on whole-body MRI did not experience an event over the follow-up period, the risk for MACCE was substantially higher among diabetic patients with any finding on whole-body MRI. Among those, a cumulative event rate of 20 percent at three years and 35 percent at six years was demonstrated.

“This discovery shows that a comprehensive, whole-body MR acquisition provides accurate assessment of the underlying disease burden in patients with diabetes mellitus that has very relevant prognostic information,” says  Bamberg. “Also, the prognostic value may be higher than only limited protocols, such as those tailored to the heart.” 

Special expertise in MR imaging is required for whole-body MR protocols. The cardiac sequences are generally the most challenging acquisitions, but can be easily performed by well-trained technologists.

“In my institution, these acquisitions are well-liked by most of the technologists because they cover more than one organ system, are interesting to perform, and take approximately an hour,” Bamberg explains.

Whole-body MRI is not an exclusive, non-invasive risk assessment tool only for those with diabetes. Rather, it could be incredibly beneficial for patients with other systemic diseases too.

“Diabetes is a systemic disease that affects all organ systems, so a whole-body imaging approach with a special emphasis on cardiovascular complications is just more appropriate,” says Bamberg. “There are a number of other systemic diseases that may benefit from a comprehensive assessment, including inflammatory disease states.”

Bamberg and others are currently executing whole-body MR in large population-based studies that include diabetic subjects in order to confirm their findings and illuminate differences in the predictive value of the imaging technique for type 1 and type 2 DM.

With continued research into PET/MR, PI and whole-body imaging MRI, the view with MRI looks to continue to improve in the years to come.