The first focused meeting on the use of 3T MR for cardiovascular applications convened under National Institutes of Health sponsorship last September. Luminaries in MR technology and applications met in Washington, DC, for two days under the leadership of National Institute of Biomedical Imaging and Bioengineering director Dr. Roderic Pettigrew.
The workshop was organized into seven sessions, each led by two key figures: a moderator and a rapporteur. The moderator introduced the speakers and summarized each talk at the end. The rapporteur provided a critical review at the end of each session as a means of preparing the audience for a brief discussion.
At the close of the conference, Paul Bottomley, Ph.D., and Dr. Robert Edelman provided scholarly overviews of what they considered the technical and clinical outcomes of the workshop, respectively. Several conclusions emerged:
- General agreement held that 3T is a worthwhile improvement over 1.5T for imaging the heart and blood vessels. Some technical challenges, however, need further optimization to realize the full potential of 3T.
- Three-T is not simply twice as good as 1.5T, however. The signal-to-noise ratio at 3T was estimated to be 1.4- to 1.6-fold better, leading to clinically relevant improvement in spatial resolution with resultant ability to depict the coronary arteries and other vessels with exquisite results.
- Improvement is also gained in contrast to noise. Delayed enhancement is seen with better sensitivity. Radiofrequency tag lines are better defined and persist substantially longer at 3T, which facilitates the quantification of diastolic cardiac motion.
- Ventricular function studies provide more information depicting myocardial stress and strain. Steady-state free precession (SSFP) imaging will take more work for optimum results. While more traditional imaging approaches (e.g., fast gradient-echo imaging) have improved, spiral imaging approaches have been shown to be very promising at 3T as well.
- An increase in chemical shift frequency dispersion was achieved for better detection of perturbations that change frequency, even allowing single cell evaluation.
- Spectroscopic studies using phosphorus-31 and other nuclei, as well as the blood oxygen level-dependent effect, improve with 3T.
- Morphology, perfusion, coronary, and other vascular imaging and spectroscopy were all thought to be improved at 3T. Enthusiasm was more muted for studies of ventricular function, since SSFP is more challenging at 3T than at 1.5T. Approaches to improve studies at 3T were suggested, but all agreed that using the fast gradient-echo approach would produce better images of function at 3T than at 1.5T.
- The improvements in signal to noise should lead to better molecular and nanotechnology-based imaging.
Accompanying these improvements with 3T were several problems:
- Comments frequently asserted that 3T was not as robust as 1.5T.
- Setup time was thought to be longer, although the suggestion was made that 3T would allow studies to become simpler using 4D acquisitions.
- Homogeneity in B0 and susceptibility were problematic. The "India ink artifact," a contiguous reduction in signal over the epicardium resembling the outlining of structures with India ink over a portion of a cardiac slice, has been more commonly observed at 3T. This was thought to represent a phase change between the water and fat interface that occurs at the epicardium.
- The substantial increase in specific absorption rate could exceed FDA limits, particularly using the SAR-intensive SSFP pulse sequence. This was a major concern. There are also potential problems using adiabatic pulses. Solutions to the SAR problem could include decreasing the flip angle, decreasing the repetition time (TI), and increasing the pulse lag. Alternatively, RF pulses could be redesigned to lead to a reduction in peak SAR (e.g., using the VERSE method). Another possibility is replacing SSFP with contrast-enhanced FLASH for coronary artery imaging. Parallel imaging was considered especially useful at 3T.
- Another concern involved use of 3T in patients with implanted devices. Problems of greater potential for displacement and/or heating of a device were noted. Burns are more likely due to increased possibility of RF heating. It is essential when working at a 3T field that is not as widely tested as 1.5T to use appropriate protocols.
- Nonuniformity in B1 was another problem. This could be overcome using multichannel parallel imaging or adiabatic pulses. But the latter would lead to an increase in SAR.
- Since SSFP is presently the standard for producing the highest quality images of function at 1.5T, it was suggested that strategies should be developed to take advantage of this ideal pulse sequence and the higher field. One approach would use phase cycling to allow a decrease in bandwidth to generate higher SNR. Other approaches could include reducing TR using asymmetric echoes, adjusting the center frequency, or simply using the older FLASH approach.
It was generally agreed that 3T imaging for cardiovascular studies leads to effective improvement in almost all aspects of cardiovascular imaging and would become the standard within the next few years. The few problems were thought to be surmountable. One is difficulty in using the gold standard of pulse sequences (SSFP) for imaging of ventricular function. A second is the increase in SAR at 3T, which could lead to deleterious effects. The problem of accurate SAR determination was something that industry could address with a little more investment.
Three-T could lead to the realization of the so-called one-stop shop, since all of the versatility of cardiovascular MR could be optimally expressed at 3T. The September workshop concluded with an almost universally high level of enthusiasm that 3T will become the new standard to replace 1.5T for cardiovascular imaging.
Dr. Pohost is a professor of cardiovascular medicine at the University of Southern California in Los Angeles and chief medical officer for Salick Cardiovascular Centers in Beverly Hills.