Experience overcomes difficulties of 3T MRI

April 14, 2006

Whole-body MR scanners that operate at high field strengths are becoming widely available, and new 3T models have been launched in recent years. This generation of scanners offers exciting possibilities for radiological diagnosis while also posing challenges.

Whole-body MR scanners that operate at high field strengths are becoming widely available, and new 3T models have been launched in recent years. This generation of scanners offers exciting possibilities for radiological diagnosis while also posing challenges.

The primary advantage of 3T scanning is increased signal-to-noise ratio. This can be used to cut MR scan times or to increase the resolution of images. MR spectroscopy performed at 3T should also benefit from a higher chemical shift separation, which manifests in spectral peaks that are farther apart. This advantage has to be balanced against an increase in chemical shift artifacts in imaging sequences.

Imaging at higher field strengths means that T1 relaxation times are longer. This results in better background suppression during MR angiography, and better vessel/tissue contrast. One drawback, however, is the decreased contrast between gray and white matter on T1-weighted images of the brain.

The T2* effect will also increase when moving from 1.5T to 3T. Both perfusion-weighted and functional MR studies will have higher sensitivities but also more susceptibility artifacts. Scanning at high field strengths can also cause problems with the specific absorption rate (SAR) limit, intended to avoid an excess of radiofrequency in the body, which can mean longer scan times.

An MR examination performed at 3T may generate a high level of acoustic noise, and patients should always wear earplugs and earmuffs. The higher field strength also increases the risk of incidents involving external objects.

Instruments and devices, including the respirator and monitoring equipment, must be 3T compatible. They must not be affected by the magnetic field. Active shielding is used to minimize the extension of the magnetic field. A consequence is that the force experienced 1.5 meters away from a 3T scanner is about the same as that generated by a 1.5T scanner. At close range, however, the force rapidly doubles, raising the risk of adverse incidents involving projectiles. Implants in the patient must also be 3T compatible.

Lund University Hospital installed its Intera 3T whole-body MR scanner (Philips Medical Systems) in September 2003. The scanner operates with a gradient strength of 30 mT/m, a slew rate of 150 mT/m/sec, and a field-of-view of 300 x 400 x 400 mm. Its short-bore magnet is considered more convenient and less claustrophobic for patients.

Because 3T scanners were still evolving when we first began use, we struggled with protocols and sequences. We optimized the flip angle to the TR to avoid the lower contrast between gray and white matter in T1-weighted images. But increased contrast means a weaker signal. A TR of 500 msec and standard flip of 90 degrees gave 100% SNR. Optimizing the flip to 76 degrees caused the SNR to drop to 84.6%.

Inversion recovery sequences give a better contrast between gray and white matter, but they can also miss contrast-enhanced tumors. We are assessing how a change to the inversion time from 400 to 800 msec will affect contrast. Preliminary results look promising.

Acquisition of T1-weighted sagittal lumbar spine images has caused many problems. We could not get rid of fuzzy, gray artifacts despite testing many parameters. We scanned with and without flow compensation, changed the phase encoding direction, and experimented with different TE and TR values. None of these strategies worked, and the artifacts remained.

Suggestions that the problem could be related to a dielectric effect led us to increase the flip angle and refocus the pulse for better penetration. This did not improve the image quality, however. We also had a few cases where hypointense bone marrow was misinterpreted as tumor infiltration. As a result, Lund University Hospital stopped performing spinal examinations on its 3T scanner in May 2005.

In summer 2005, a patch was introduced that makes it possible to scan with the phase direction feet to head. This reduces flow from cerebrospinal fluid, and the images are of good quality. We have since resumed spinal examinations.

Increasing the field strength at which MR examinations are performed means that artifacts will increase as well. Susceptibility artifacts, in which air meets brain tissue, become more severe in gradient-echo sequences. Metal from prostheses makes imaging almost impossible if the part in question is close to the area of interest. Breathing artifacts from abdominal scans show up as stripes all over the image. This makes faster scanning necessary and the use of rest slabs essential.

Motion artifacts can be limited if patients are advised to keep still and are placed in a position that is reasonably comfortable. Certain fixation devices may be helpful. Faster scanning also reduces the likelihood of motion artifacts. This is achievable at 3T with parallel imaging techniques, such as SENSE. The SENSE coil detects the signal from a fixed place in a reference scan and then reduces steps in k-space. The scanner subsequently uses information from the reference scan to avoid wraparound artifacts.

Metal artifacts will also increase as the magnetic field strength is raised. We inform patients with prostheses that they should tell us immediately if they feel any heat or pain (Figure 1).


Increasing the field strength can make a significant difference to MR angiography studies. Because the background is suppressed better at 3T compared with 1.5T, vessels can be seen more clearly. Also, when we performed interventional angiography on a patient with coils in an aneurysm, surprisingly few metal artifacts showed up on the T2-weighted turbo spin-echo sequence (Figure 2).

We run two fast phase-contrast sequences in two directions, with a velocity encoding of 20 cm/sec, when looking for thromboses. The merits of high-field scanning can be seen by comparing 1.5T and 3T imaging results from the same patient (Figure 3).

The higher signal generated by 3T systems can be used to yield higher resolution. When we want very high resolution pictures of the brain cortex, for instance, we scan first with the SENSE head coil, and then with the Flex Medium coil. The use of a surface coil (Flex M) located close to the area of interest makes it possible to observe individual layers on the cortex with 1.2-mm slice resolution. Signal will drop in the middle of the brain, however, owing to the large distance from the coil.

Diffusion-weighted imaging at 3T is rising to prominence. DWI relies on sensitivity to the movement of water molecules. Any event that restricts the molecules' motion, such as a stroke, produces a bright signal. Apparent diffusion coefficient maps, which illustrate the mean diffusion in three directions, can be produced from two or more different b-values. The b-value indicates the motion sensitivity.

Sensitivity of DWI to the movement of water molecules increases at 3T owing to the higher signal. In other words, higher b-values can be used. Use of SENSE to cut scan times improves DWI results, too. This is because there is less time for spins to "diphase" and produce distortions.

Diffusion tensor imaging involves many more diffusion directions. DTI will also benefit from the higher signal of the higher magnetic field strength (Figure 4).

We use an echo-planar, fast-field echo sequence for perfusion-weighted imaging. Injection of contrast lowers the MR signal. The main advantage of 3T scanning is the increased T2* effect. Data processing is aided by an algorithm derived from Leif Ostergaard's work and developed further at Lund.1-3 This enables us to produce maps of relative cerebral blood flow (Figure 5), relative cerebral blood volume, and relative mean transit time.

We have just started to perform MRS on our 3T whole-body scanner. MRS has an important clinical role to play in separating tumors from abscesses. We plan to start functional MR examinations on this system as well.

Cardiac imaging at 3T can be highly artifact-prone. We have experienced many such problems when scanning. When the images are good, however, they are extremely good (Figure 6).4

Future 3T scanners will have more coils, more channels, and SENSE capability. Such advanced technology is likely to carry a high price tag, however. Three-T scanners offer superior imaging capabilities, but achievement of the potential benefits is not automatic.

MS. HANSSON is an MR radiographer at Lund University Hospital in Sweden. She has experience working with 3T MR scanners manufactured by Siemens Medical Solutions, Philips Medical Systems, and GE Healthcare.


1. Ostergaard L, Weisskoff RM, Chesler DA, et al. High resolution measurement of cerebral blood flow using intravascular tracer bolus passages. Part I: Mathematical approach and statistical analysis. Magn Reson in Med 1996;36(5):715-725.

2. Thilmann O. LUPE: An extensible modular framework for evaluation of DSC-acquired perfusion images. Presented at annual meeting of the European Society for Magnetic Resonance in Medicine and Biology, Copenhagen; September 2004:537.

3. Andersson L, Wirestam R, Siemund R, et al. Colour-coding of MRI-based rCBV and rCBF maps by normalisation to cerebellum or whole brain. Presented at annual meeting of the European Society for Magnetic Resonance in Medicine and Biology, Paris; September 2000:125-126.

4. Markenroth K, Hjertberg-Kalman V, Carlsson M, et al. Quantitative flow measurements in the coronary sinus. Combining 3T with SENSE. Presented at the annual meeting of the European Society for Magnetic Resonance in Medicine and Biology, Copenhagen; September 2004:126.