Doing Without
By Beth W. Orenstein
Radiology Today
Vol. 20 No. 12 P. 18

New methods of noncontrast MRI are becoming key tools for diagnosing cardiovascular diseases.

Cardiovascular disease (CVD) is the leading cause of death worldwide, according to the World Health Organization (WHO). CVD accounts for nearly one-third of mortality, the WHO reports. Most CVDs involve dysfunction of the heart muscles and heart rhythm. Interest in assessing the heart noninvasively continues to grow, with 3D MRI taking a strong lead. MRI is now routinely used to diagnose CVDs such as cardiomyopathy, heart attack, and irregular heartbeat, among other heart diseases. However, patients are often given a dose of the contrast agent gadolinium to enhance and improve the quality of the images.

A few years ago, radiologists and patients began to question the safety of gadolinium, after a study was published in Radiology that showed the contrast agent is deposited and retained in the brain. In addition, a small number of patients came forward and claimed their health had been harmed by gadolinium. Gadolinium can also have an effect on kidney function and is likely contraindicated in patients with kidney disease. That’s why, according to Dawn Berkeley, manager of medical affairs, clinical development MR at Canon Medical Systems, “For the last three to four years, there has been a move to do more MR imaging with less contrast or no contrast when possible.” That effort includes MRI of the heart and blood vessels, Berkeley says.

Unique Technique
Researchers at the University of Warwick in Coventry, England, have developed hierarchical template matching (HTM), a new 3D MRI computing technique that can be used to diagnose heart disease without contrast. They report on it in a recent edition of Scientific Reports. The technique is used to track the cardiac muscles and measure strain, a mechanical property. In this context, strain represents shortening and lengthening of muscles. For example, in a healthy heart, wall muscles shorten and lengthen during a heartbeat, whereas in a diseased heart, such as those of heart attack patients, some of the muscles lose their ability to shorten and lengthen. Finding these muscles, clinically referred to as infarcted muscles, that have lost their ability is an important diagnostic criterion for physicians, explains Jayendra M. Bhalodiya, a doctoral candidate at the university who conducted the research.

“HTM tracks the muscles of the heart wall during a heartbeat and records the coordinates,” Bhalodiya explains. “These coordinates are used for estimating the displacement and strain values of the muscles. Based on how much the heart wall [myocardium] is moving, we identify muscles which are not moving enough, which are identified as infarcted muscles in heart attack patients.”

Bhalodiya notes that HTM is an image processing method that performs analysis on collected MRI. It is not, he says, an image acquisition method and, therefore, it can’t be directly compared with acquiring contrast images. For this reason, the researchers didn’t acquire scans with and without contrast for comparison. It can, however, be used as a comparison for specific clinical conditions such as myocardial infarction, he says.

Bhalodiya believes eliminating contrast is a significant benefit because it overcomes safety concerns and can be used in patients with renal insufficiency. Also, he says, use of a contrast agent increases treatment costs for patients.

Another advantage of the technique, Bhalodiya says, is that it reduces scanning time, something that can help to relieve stress on patients. In a worst-case scenario, a clinician may have to stop the scan. Often, when patients are highly uncomfortable and not able to follow all of the scanning instructions, it may take longer than usual to acquire quality images. In such cases, the administration of gadolinium may lead to complex situations.

“This technique doesn’t require a dosage of gadolinium, as it tracks the heart naturally,” Bhalodiya says.

The biggest challenge in developing this technique, he says, was learning multiple technical and clinical skills. Also, he says, it required “a lot of patience to understand how clinical organization works and what the important criteria are while diagnosing a patient.” The researchers overcame both challenges by having discussions with clinical staff and consulting experts, Bhalodiya says.

More Depth, Less Strain
Mark Williams, PhD, MSc, a professor and leader of the Centre for Imaging, Metrology and Additive Technology at WMG at the University of Warwick, believes that the HTM method offers physicians the opportunity to “see in more depth what is happening to the heart, more precisely to each heart muscle, and diagnose any issue, such as remodeling of heart, that causes heart failure.”

Currently, Bhalodiya says, HTM can locate larger segments of infarcted muscles in the left ventricle of the heart. “In future, we are planning to further improve HTM to determine exact size and transmurality of an infarct,” he says. A transmural myocardial infarction refers to a myocardial infarction that involves the full thickness of the myocardium; a nontransmural myocardial infarction does not involve the full thickness of the myocardium.

Where did the researchers at Warwick get the idea to develop this technique? Bhalodiya explains that HTM is an image processing method inspired, in part, by the way humans observe images. It has three steps: matching points, transforming points, and calculating strain.

“The idea of matching is inspired from human observation behavior: The way we first see a bigger picture and then narrow down our focus for smaller parts of the picture. Similarly, we match bigger parts of images first and then match smaller locations among pictures. HTM performs this procedure in a hierarchical manner and among the frames of a cardiac cycle. Ultimately, this matching procedure helps to calculate the transformation of myocardial points,” he says. “The transformation of points is performed using a local weighted mean [LWM] function. LWM is a mathematical function that we introduced in the myocardial tracking task due to its mathematical benefits. We adopted the strain calculation step from literature.”

How does HTM eliminate the need for contrast? “Infarcted muscles in heart attack patients do not shorten and lengthen like healthy muscles do,” Bhalodiya says. “We utilized this material property, which is quantified as strain, to identify infarcted muscles—quite reduced or closer to zero strain values refer to infarcted muscles—through our image computing method.” Hence, infarcted muscles could be located by estimating the material property and without using a contrast agent–based scan, he says.

Vascular Studies
Some facilities are now switching to noncontrast cardiac MR exams. “We have several facilities that try, whenever possible, to switch to noncontrast techniques,” Berkeley says.

Kanae Mukai, MD, medical director of noninvasive cardiovascular imaging at Salinas Valley Memorial Healthcare System in California, says that more than 90% of vascular MR angiograms at the Ryan Ranch Center for Advanced Diagnostic Imaging are now performed without contrast. She says that Canon has developed a broad range of sequences or protocols, which are proprietary and unique to the Canon MR system and can be used to obtain high-quality noncontrast images. These include Fresh Blood Imaging (FBI); Flow Spoiled (FS)–FBI; Time-Spatial Labeling Inversion Pulse (TIME-SLIP); and mASTAR, which helps clinicians visualize time-resolved hemodynamic flow.

Because Salinas Valley obtains nearly all of its vascular MR angiograms (MRAs) without contrast, it’s difficult to provide a one-to-one comparison of results for contrast vs noncontrast, Mukai says. However, she’s quite comfortable with the noncontrast applications.

“We have performed noncontrast MRA studies to evaluate lower-extremity vascular access strategies for structural intervention cases, such as transcatheter aortic valve replacement and endovascular stent procedures, as well as the thoracoabdominal aorta, neck, and intracranial vessels,” Mukai says. “Thus far, follow-up scans with fluoroscopy and/or CT have shown strong correlation, confirming the original diagnosis on the noncontrast MRA.” When artifacts are seen, she says, “they can be frequently addressed by optimizing the protocol or by using the appropriate combination of sequences to delineate the anatomy.”

One disadvantage of acquiring images without contrast is that the acquisition itself does take slightly longer than with contrast, Berkeley says. Rather than seconds, it may take one or two minutes to acquire the images. “So, if you have a patient who is extremely uncooperative, it’s not going to be possible,” she says. “Patients have to be still, as they would for any other MR exam.”

Although performing noncontrast MRA does pose some unique challenges, especially for technologists who may be unfamiliar with the tools that are available on the equipment, Mukai says Salinas Valley has been able to overcome these challenges with training support from Canon.

Berkeley says Canon’s automation of the pulse sequences for FS-FBI has benefited technologists. “Now, technologists don’t have to calculate the values based on the patient’s heart rate. The system automatically calculates it for them,” she says. When technologists had to calculate the delay times for an acquisition, they would steer away from these exams; they didn’t like having to calculate the timing of when to image. “But once you say the word ‘automated’ to them, they would say, ‘Oh, I can do this,’” she adds. Automation, using delay tracker, was introduced in 2011 to the Canon MR systems, which fully automated this part of the exam, she notes.

Reliable Method
A multicenter international study, Renal Artery Contrast Free Trial (REACT), using the Time-SLIP sequence demonstrated a high correlation between noncontrast MRA images and contrast-enhanced CT angiography (CTA) images. The trial showed no statistical differences between noncontrast MRA and CTA in terms of their ability to visualize renal artery stenosis. The results from the trial were published in the January 2015 issue of the American Journal of Roentgenology.

“We are looking into getting involved in additional research initiatives with noncontrast vascular imaging and noncontrast cardiac imaging,” Mukai says.

Time-SLIP is an arterial spin labeling (ASL) variant that is typically used to depict blood vessels within a targeted region, Berkeley notes. Time-SLIP has been evaluated for use in the renal arteries, carotid arteries, pulmonary system, and portal-venous system. It produces angiograms using blood as an endogenous tracer, instead of a contrast agent, she explains. Because Time-SLIP does not require contrast, it can be used on patients who have advanced kidney disease.

Berkeley says further advances in software and hardware allow users to “acquire several variations of ASL or Time-SLIP techniques and apply multiple timings for tag pulses during one sequence so you can target a vessel and watch as blood fills or moves through that vessel. That can indicate whether or not there are blockages or malformations.”

Mukai believes that noncontrast MRA has a definite place in the future “as we continue to strive toward less invasive, high-quality vascular imaging.” Peripheral vascular disease is commonly found in patients with end-stage renal disease or dialysis patients in whom the use of a contrast is contraindicated, she says. “While gadolinium can be a valuable agent for visualizing the vascular anatomy, it is important for referring physicians and imagers to weigh the risks and benefits,” Mukai says. “Without an effective means of capturing the same information without contrast, this can be a difficult dilemma for the referring physician or imager. Having a reliable method for noncontrast cardiac MRA allows us to safely avoid having to unnecessarily expose patients to this risk.”

— Beth W. Orenstein of Northampton, Pennsylvania, is a freelance medical writer and regular contributor to Radiology Today.