As MR and CT jockey for position in the race to dominate noninvasive angiography, few are ready to predict which imaging method will make it to the winner’s circle. There is one safe bet, however: Invasive angiography, while still in the race, is fading fast.

Noninvasive angiographic methods are proving to be both swift and powerful in imaging arteries from the cranium to the foot, and may one day make invasive angiography a distant memory, except in special situations.

“We’re going to be looking back at these days and thinking, ‘Can you believe that we used to diagnose disease in that crude way? Can you believe we jabbed people in the groin to make diagnoses?’” said Dr. Martin Prince, chief of MRI at New York Hospital.

CT angiography and MR angiography can be sized up on the basis of their technological characteristics, their innate strengths in particular clinical applications, and such practical issues as access, ease of use, and cost. No matter how those calculations add up, however, institutional expertise must be thrown into the mix. Even the most prestigious academic centers typically build a reputation for excellence in CTA or in MRA, but not both. It’s a bias that not only influences which exam will be done at a given medical center, but also skews interpretation of the scientific literature.

“I can look at an article and, just by seeing the name of the first author, usually tell what the results will be,” said Dr. Elliot Fishman, a professor of radiology and oncology at Johns Hopkins Hospital in Baltimore. “No one who’s doing MR research ever publishes an article that says CT is better, and no one doing CT research ever publishes an article that says MR is better.”

In noninvasive angiography, spatial resolution is critical for discriminating smaller blood vessels, as well as for appreciating variations in luminal contour that indicate occlusive disease. When a large area must be imaged, the radiologist must make a trade-off between length of coverage and spatial resolution. That’s where CTA’s speed offers a strong advantage, said Dr. Geoffrey Rubin, an associate professor of radiology at Stanford University.

“CT is capable of acquiring many more voxels per second than is MR, and the range in speed increase can be anywhere from three to 20 times, depending upon exactly how one is acquiring the data and what type of coverage is critical,” Rubin said. “If it’s important to cover a large distance, this is when CT takes full advantage of its speed.”

Examples of studies that call for rapid imaging of large distances include the entire aorta (which traverses the thoracic and abdominal cavities), the iliac arteries (chest, abdomen, and pelvis), and lower extremity inflow and runoff (from the diaphragm to toes). The introduction of multidetector technology, which enables acquisition of four slices of data in the time it used to take to acquire one, has been the key to CT’s prowess in whole body imaging, advocates say.

“With the multidetector technology, we can do CT angiograms, which run from the celiac artery all the way down to the ankles. With neuro CTA, we can go from the aortic arch all the way up into the intracranial circulation. That’s different from the past when tube cooling and scanner hardware limited our coverage,” said Dr. Lawrence Tanenbaum, chief of MRI, CT, and neuroradiology at the New Jersey Neuroscience Institute in Edison.

The increase in CTA’s speed—whether done by multidetector helical scanner or electron-beam scanner—has increased throughput, so patients don’t have to hold their breath as long. And it has also delivered the option to improve spatial resolution by acquiring data in thinner slices.

“Take something as simple as a renal artery and a transplant donor. We were looking at the renal arteries with 3-mm-thick sections. Now we look at them with 1.25-mm-thick sections in a single breath-hold,” Fishman said.

Even MRA experts agree that CTA has come a long way as a result of multidetector technology and improved tube design. But don’t ask them to concede its superiority over MRA.

“CT’s coverage now begins to approach what MR has always had. Both technologies can now do total body angiography,” said Dr. Thomas Grist, an associate professor of medical physics at the University of Wisconsin in Madison. “One of the advantages of gadolinium-enhanced MR is the fact that coverage is so good, and MR scanners are definitely becoming faster. It’s possible to cover a much larger region with MR, or in the same time obtain higher spatial resolution.”

Both sides agree that MR holds the lead in imaging small body parts, such as the hand or foot, where speed is not an issue. The use of special coils designed for specific applications enables MR to acquire very small voxels—something CT can’t do—without sacrificing an optimal contrast-to-noise ratio.

“With CT, the voxel size is limited by physical constraints of the CT hardware, whereas with MR, once you get past the need to have very powerful gradient amplifiers, you’re just trading off between a voxel size that optimizes resolution versus having too much noise in the data,” Rubin said.

Safety Issue

“If you’re the patient, you often prioritize things differently than the doctor or healthcare industry,” Prince said. “When I’m thinking about what I’m going to do for my daughter or my mother, I’m thinking, ‘Before we take all this risk, why don’t we just do a quick MRA?’”

For Stanford’s Rubin, the risk posed by exposure to ionizing radiation tips the balance in favor of MRA as well, but only in certain cases. An example would be in imaging the aorta. Ordinarily, Rubin would prefer CT for aortic angiography, but not in a younger woman, whose breast tissue would be exposed to ionizing radiation.

“If I had a 25-year-old woman coming in with a suspected coarctation of her aorta, I would elect to perform an MR angiogram over a CT angiogram,” Rubin said. “Alternatively, if I have a 70-year-old woman who comes in with chest pain and an expanding mass on her chest x-ray that’s believed to be an expanding aneurysm, and she would be considered a candidate for stent-graft repair, then I would unquestionably get a CT scan.”

Concerns about nephrotoxicity also shape the choice of angiographic method, but decision-making appears to be swayed by technological allegiance more than clear-cut guidelines. Because of CT’s superior spatial resolution, Fishman and Rubin prefer to use CTA for renal studies unless the patient has renal dysfunction or a specific risk for nephrotoxicity.

“When you’re looking at small vessels, say, in mesenteric ischemia, or small branch vessels of the kidney, resolution is what it comes down to,” Fishman said.

Grist, however, believes MRA is the best choice in most renal cases.

“We prefer to evaluate our renal artery stenosis patients with MR because a lot of those patients have renal insufficiency,” Grist said. “We still do most of our peripheral lower extremity angiography studies with MR as well, because many of those patients are diabetic, and 40% or 50% of them will have renal insufficiency associated with peripheral vascular disease.”

Despite its overall lower risk, MRA does raise concerns about safety as well as comfort in some instances. Many patients with vascular disease have coexisting heart disease, with implantation of a pacemaker or cardioverter-defibrillator. These metal devices eliminate MR as an imaging option. Cochlear implants are also contraindicated.

In addition, because MRA scans tend to take longer—an hour or more if the exam involves multiple pulse sequences—patients may experience back pain or claustrophobia.

Fundamental Differences

“The problem in comparing MRA to CTA is that CTA really only does one thing. It gives you a kind of anatomic map, whereas MRA is a kind of broad encompassing term that includes additional sequences that look for information about flow and physiology,” Prince said. “Just the fact that MR is safer than CT makes it possible to do all kinds of additional sequences without having to be concerned that you’re inflicting injury on the patient or causing unnecessary exposure to ionizing radiation.”

For example, Prince said, a typical MRA of the carotid arteries might include a combination of several sequences that have different levels of resolution. If one carotid artery is occluded, a phase contrast sequence might be added to the lineup to evaluate the direction of blood flow and determine whether the anterior and posterior communicating arteries are permitting collateral blood flow from one side of the brain to the other.

Perfusion and diffusion imaging of the brain may also be a useful addition to a carotid exam in the case of suspected stroke, Grist said. He cited imaging of the kidneys as another example where MR offers key diagnostic information beyond that available from an arterial map.

“There are certain techniques with MR that measure renal function and the response of the kidneys to various pharmacologic interventions, like angiotensin-converting-enzyme inhibitors, which give some information about the physiologic significance of a stenosis,” Grist said.

For its part, CT offers supplemental information as well, though of a different nature than that available from MR. Rather than revealing physiologic and metabolic activity, CT can suggest alternate diagnoses, as a result of its greater field-of-view and spatial resolution.

Suspected pulmonary embolism offers a good example. Not only does CT have a natural advantage in the chest, where air causes little attenuation of the x-ray beam, but in addition, its high spatial resolution enables it to detect tiny pulmonary emboli more easily. Even more important is its utility in offering alternate diagnoses when an embolus is not found.

“Only about one-third of patients suspected of having a PE actually have one. CT is the most likely imaging test to offer you the alternative diagnosis, whether it’s pneumonia, pneumothorax, or lung cancer,” Rubin said.

Evaluation of patients for liver transplants also takes advantage of CT’s large field-of-view, Fishman said.

“When we do the liver for transplant, we not only look at vascular maps, but we look at the liver for tumor, we look outside the liver, we look at the nodes, we get the whole abdomen,” he said. “With CT, in a sense, the chance of serendipitous findings increases. It’s a big advantage for CT.”

One challenge that both techniques would like to master is evaluation of the arterial wall itself, not just the interior contours that are visible on conventional angiography. Because CT provides information over a broad range of x-ray attenuation values, it is able to characterize structures other than the lumen of the blood vessel. Imaging the outer wall can determine the extent of aneurysm, for example. Also, looking at the wall will show the amount of plaque burden and the shape of plaque, which might indicate the likelihood that the plaque will embolize.

CT’s sensitivity for identifying calcification in the arterial wall is also useful, particularly for surgical planning in patients who need surgical graft implantation, as calcification makes it more difficult to achieve durable anastomoses with graft material.

Perhaps the most pointed question about the arterial wall is which plaque—by virtue of being inflamed, filled with a lipid core, and capped by only a thin fibrous layer—is most vulnerable to rupture. It is an issue that is most pressing in the coronary arteries, where plaque rupture can cause a heart attack even in patients with only mild or moderate coronary stenoses.

Neither CT nor MR can today identify vulnerable plaque, but most bets are on MR.

“Looking for the stenosis may not be as important as looking for the vulnerable plaque,” Prince said. “MRA is good for looking at lipid content, and inflammation, and thickening.”

Need For Speed

With MR, a streamlined exam limited to luminal imaging takes about 20 minutes. A comprehensive exam may involve an additional five to 10 pulse sequences and stretch the total exam time to more than an hour. It is also a more complex exam to do.

“That is a general weakness. MR has tremendous potential, but it comes at a price, and that price is that it really requires somebody to understand the technology in detail,” Grist said.

MR makes up for lost time when it comes to doing 3-D reconstructions. For both techniques, 3-D reconstructions have been critically important not only in enabling diagnosis, but perhaps more importantly, in winning support for noninvasive angiography among referring physicians. That’s because the 3-D reconstructions enable a nonradiologist to instantly recognize pathology.

“If you present most clinicians with a stack of transverse CT sections or coronal MR sections, they don’t know what to do with them,” Rubin said.

With today’s technology, however, it takes far less time to do 3-D postprocessing with MR, because the brightest structures on the scan are blood vessels, particularly the lumen. With CT, the bones are brighter than the arteries. and must be removed in postprocessing. Arterial calcification can cause similar problems.

“In a flat-out race for processing speed, MR would unquestionably win,” Rubin said.

How long it takes to do CT postprocessing depends on the type of exam. In the abdomen and pelvis, it might take two or three minutes to remove bones from the image. The skull, or the distal legs and ankles are far more time-consuming, and might require an hour’s worth of work.

Automation may make CT postprocessing much easier, however. Tanenbaum has been working with GE Medical Systems on software called Direct 3D that automatically creates volume renderings as the image slices are acquired (see sidebar). Work on automated multiplanar reformations is also under way. Together, such automation should take virtually all of the post-processing out of the hands of the technologist and the physician.

“Then CT postprocessing will be almost as simple and hands-off as MR. It should significantly increase utilization,” Tanenbaum said.

Other Considerations

Ultimately, the choice between CTA and MRA may come down to the type of equipment each medical center has and the expertise and experience of its radiologists.

“Even if CT is shown to be vastly superior to MR for certain applications, if your CT scanner is of 1985 vintage and your MR scanner is a 1999 model, chances are you’re going to do better with the MR—and vice versa,” Rubin said.

“You can try to measure where things are at a specific point in time, but the speed of innovation is so rapid it’s amazing,” said Dr. Elliot Fishman of Johns Hopkins Hospital. “It used to take years to make a significant change. Now it may take months.”

That dilemma aside, interviews with leading researchers and industry insiders have produced the following sample of key technological developments, either new to the MRA and CTA markets or likely to be introduced soon.

• Eight-channel detectors. CTA insiders predict that the introduction of multidetector technology with eight channels is right around the corner. This leap forward will double the advances seen with today’s four-channel technology, enabling imaging at twice the speed or at higher resolution, by acquiring thinner slices in the same time.

• Volume rendering. Automated volume rendering technology will end the need to send CT image data over a network to a workstation, then manually load the images and select a volume rendering protocol. Instead, images will reconstruct right on the scanner as the data are acquired, building a volume-rendered display, slice by slice. Automatic multiplanar reformations, which are in a less mature stage of development, could take nearly all of the postprocessing tasks out of the hands of the technologist and physician.

• Contrast media. One of the most exciting areas of MRA research involves contrast agents. Those with higher relaxivity enable shortening of the T1 relaxation time and will generate a better signal-to-noise ratio (SNR). Blood pool agents, which are made of larger molecules than conventional gadolinium-based contrast agents and therefore resist leaking from the blood pool, will enable higher-resolution scanning.

  Contrast-enhanced peripheral angiography is also generating a buzz in the MR community. It involves injection of contrast, either slowly over about a minute or in three rapid injections, and then moving the patient through the gantry during imaging. This technique enables studies to be done in a few minutes, rather than 45 minutes to an hour.

• Pulse sequences. Improved pulse sequences could increase the sampling efficiency of MR scanners from the 25% to 30% range to 75% to 80%. This would mean the MR scanner could spend more time looking for signal, possibly shortening scan time to seconds.

• Software. Navigator and other motion compensation techniques could make it possible to obtain motion-free MR images over several minutes and build up the SNR through signal averaging. Alternatively, combining fast imaging with navigator techniques in some manner may further improve either the temporal or spatial resolution, particularly in the coronary arteries.

• Software advances may improve the postprocessing of images acquired using blood pool agents, enabling, for example, selective removal of the venous structures for an improved view of the arteries, or vice versa.

• Miscellaneous. Other areas of development in MRA promise higher field strength magnets (3 to 8 tesla), faster gradients, coils designed for imaging specific body parts, and the ability to do MRA in an open magnet.

“This is a fast-changing field, and especially in MR, there is so much happening that what you see today is going to be totally different very soon,” said Dr. Martin Prince, chief of MRI at New York Hospital.