Because of the modality’s excellent soft-tissue contrast, multiplanar capability, and good spatial resolution, the heart was an early target of clinical MRI. Most of the early work focused on structural abnormalities, such as those seen in congenital or valvular diseases and some cardiomyopathies, but myocardial perfusion and functional studies were soon introduced. The tremendous burden of atherosclerotic cardiovascular disease in developed countries drove attempts to perform coronary angiography using MRI, but the technical problems of obtaining satisfactory images proved complex. Not only does the motion of the heart make it difficult to obtain images without blurring, but many of the vessels are small. Now, however, it is possible to study the morphology, function, perfusion, and blood flow of the heart,1 although the smallest blood vessels remain out of reach, and routine imaging of atherosclerotic plaques is not yet possible.

Wen et al2 of the US National Heart, Lung, and Blood Institute demonstrated in 1997 that higher magnetic fields (of 3T and 4T) provided better signal-to-noise ratios (SNRs) and, thus, even better cardiac images than are possible at 1.5T. Higher field strengths can, however, be associated with artifacts caused by the characteristics of the interface between the heart and the air-containing lung.3

Figure 1. Short axis cine images of the anterior myocardium (slice planes are located within 1 cm of each other) obtained on same patient at 1.5T (a) and 3T (b) using single loop receiver coils of comparable dimensions. In order to preserve the temporal resolution and obtain the same number of images over the cardiac cycle, the timing parameters were identical at both field strengths. Contrast to noise (CNR) measures between the anterior myocardium and left ventricular blood pool measured in these end systole images are 12 at 3T and 8 at 1.5T. Courtesy of Denise P. Hinton, PhD, Massachusetts General Hospital, Boston, Mass.

Denise P. Hinton, PhD, of Massachusetts General Hospital, Boston, and the MGH/MIT/HMS Athinoula A. Martinos Center for Functional and Structural Biomedical Imaging, Boston, is an instructor in radiology at Harvard Medical School. She is working on ultrahigh-field cardiac MRI. She says, “In order to take full advantage of the SNR gain provided by high magnetic fields, both hardware and software optimization are critical. The higher field tends to be more sensitive, and we have to invest more time at the outset of a 3T body examination to remove field inhomogeneities. We do not have a full range of body radiofrequency (RF) coils as we do for lower-field scanners; RF coils that are optimal for the heart are still being developed.”

Comparative Images

The challenge of optimizing 3T cardiac MRI may be great, but the payoff will be enormous. In May 2002, Hinton and her colleagues will present some of the early findings from a direct comparison of 1.5T and 3T imaging using commercially available scanners equipped with the same gradient coil and geometrically identical built-in body coils.4 They imaged three healthy men who were 43 to 55 years of age, using cine steady-state, free-precession imaging TrueFISP with balanced gradients to obtain images with fat suppression in the short axis of the heart. Such images are used clinically to evaluate heart morphology. The investigators also used a three-dimensional, segmented, fast, low-angle shot sequence FLASH with a 14-second breath hold for coronary angiography. With both scanners, electrocardiographic gating was used. Analysis of the same regions of interest in the anterior myocardium showed a 40% improvement in SNR at 3T using comparable surface-coil receivers. The contrast-to-noise ratio of the blood and tissue was also higher at 3T. Significantly, satisfactory non-contrast-enhanced angiograms could be obtained at 3T using only the built-in body coil, whereas a more sensitive phase-array coil and contrast medium typically are needed for satisfactory coronary angiography at 1.5T. The research team expects phased-array coils for the 3T scanner to make coronary angiography clinically feasible and preferable at 3T rather than 1.5T. The greater inherent contrast between tissue and blood (about 30% better at 3T) is expected to improve measurements of vessel diameter and, hence, appraisal of stenoses.

Yuri Wedmid, PhD, manager of ultrahigh-field MRI for Siemens Medical Solutions USA, Iselin, NJ, explains that developments in 3T cardiac MRI are very promising. He says, “At 1.5T, we can see the major trunks of the coronary arteries (what is usually described as the top half of the arterial tree). With 3T scanners, we can see all of the major parts of the coronary tree.” He adds, “We have to optimize our protocols for various types of cardiac pathology. We also need to develop MRI sequences for children, but there do not appear to be any roadblocks that would prevent such optimization. We do not need new inventions to achieve this goal; we simply need to polish the current technology.”

The rewards will be substantial. “Historically, MRI has found the coronary arteries very challenging,” Hinton says. “We know, however (from cerebrovascular angiography at 3T) that the overall quality is better than at 1.5T, and the conspicuousness of smaller vessels is significantly enhanced. If we extrapolate this result, 3T MRI will improve the imaging of the main coronary vessels, as well as the branch structures.”

Gerhard Laub, PhD, Cardiovascular MR Program Manager at Siemens, reports that research is continuing on the imaging of plaque. “It is becoming clear that lumen imaging simply is not enough,” he says. “We know that there can be rupture of a plaque and vessel blockage that could not have been predicted on the basis of examination of the lumen. A couple of years of intensive research will be needed to bring plaque imaging to the clinic because we are looking at such fine detail, but this could change the types of studies that will be done in clinical practice.”

The Martinos Center at Massachusetts General Hospital has grants from the US National Institutes of Health and the American Heart Association to develop plaque imaging at 3T. Early trials have focused on the carotid arteries, where 3T scanning can acquire plaque images in half the time required at 1.5T. Development of plaque imaging requires very thin slices of less than 2 mm and high in-plane resolution of less than 0.5 mm, but the determination of plaque structure and composition may permit the identification of those lesions that are at risk of rupture.

Laub also sees greater use of MR spectroscopy. At present, with voxel sizes of 1 cm or larger, MR spectroscopy is relatively insensitive for measuring myocardial viability. That situation could well change, however, with the greater SNR available using 3T scanners. As 3T scanners become more widely available, many other applications for them are sure to appear.

Judith Gunn Bronson, MS, is a contributing writer for Decisions in Axis Imaging News.


  1. Sakuma H, Takeda H, Higgins CB. Fast magnetic resonance imaging of the heart. Eur J Radiol. 1999;29:101-113.
  2. Wen H, Denison TJ, Singerman RW, Balaban RS. The intrinsic signal-to-noise ratio in human cardiac imaging at 1.5, 3, and 4 T. J Magn Reson. 1997;125:65-71.
  3. Atalay MK, Poncelet BP, Kantor HL, Brady TJ, Weisskopf RM. Cardiac susceptibility artifacts arising from the heart?lung interface. Magn Reson Med. 2001;45:341-345.
  4. Hinton DP, Wald LL, Holmvang G, Chan R, Kirsch J, Schmitt F. Comparison of 1.5 T and 3 T cardiovascular MRI: preliminary results. Proceedings of the Society of Magnetic Resonance in Medicine. In press.