00-f5 Title and Logo 00-f6

 guide Contents
 siteinfo Site Info

Chapter 17


Field Perturbations

Local Inhomogeneities
Radiofrequency and Gradient Artifacts

Slice Profile
Multiple Spin-Echo
Line Artifacts
Motion and Flow Artifacts

Respiratory Motion
Cardiac Motion
Flow Artifacts
Signal Processing and
Signal Mapping

Black Boundary
Quadrature Artifacts
k-Space Artifacts

Chapter Seventeen
Common Artifacts in MR Imaging

17-01 Introduction

hortly after the introduction of MR as a new imaging modality, it was
 hailed as a technique which did not suffer from the common beam hardening artifacts which can destroy images in x-ray computed tomography. However, it was soon realized that an unfortunate side effect of the complex nature of MR imaging was a whole new set of artifacts.

Artifacts in MR images can take the form of variations in signal intensities or mispositioning of signals (Figure 17-01).

Figure 17-01:
If the sagittal MR image of a head that you get on your screen reminds you of Donald Duck and his nephews, there is something wrong – either with your patient or with your MR equipment.
It might be Donald Duck artifacts caused by dentures.

Such artifacts can mimic pathology to such an extent that examinations have to be redone or other diagnostic modalities have to be used to exclude pathology. Figures 17-02 and 17-03 show two examples of such artifacts. Usually artifacts can be easily recognized when their causes are known. However, cases have been described where artifacts led to surgical intervention because pathology was falsely described [⇒ FDA 1997].

Figure 17-02 (left):
Sagittal midline images through a head. (a) Intermediately weighted image, and (b) T2-weighted image. On the left image, a low signal intensity area is seen in the pons, suggesting a lesion. This lesion is not visible on the T2-weighted image. It is caused by image distortion created by ferromagnetic implants.

Figure 17-03 (right):
(a) Transverse intermediately, and (b) T2-weighted images depict an ill-defined high signal intensity lesion in the right lung. Follow-up studies on another day and the use of CT did not show such a lesion. The cause of the artifact remained unclear.

Artifacts can be categorized into four main groups:

spaceholder 600 magnetic field perturbations;
spaceholder 600 RF artifacts and gradient-related artifacts;
spaceholder 600 motion and flow artifacts;
spaceholder 600 signal processing and mapping artifacts.

We will discuss each of these categories.

17-02 Field Perturbations

The main magnetic field (or more precisely, the lines of magnetic flux) can be distorted by a number of factors outside the MR imaging suite, for instance by large stationary or moving ferrometallic objects such as elevators or passing vehicles.The field has to be protected by shimming or shielding, which is generally performed properly by the manufacturer of the MR equipment during installation. Therefore, field inhomogeneities from the outside are at present seldom responsible for image artifacts.

17-02-01 Local Inhomogeneity Artifacts

Any local internal distortion of the magnetic field cannot be corrected by shimming. The most common causes of such local distortions are the presence of ferromagnetic foreign bodies and susceptibility effects. Ferromagnetic objects usually cause an area of total signal loss around the object and distort the signal intensity at the edge of this region (Figure 17-04 and Figure 13-16).

Figure 17-04:
Artifacts resulting from the presence of ferromagnetic material. Signal loss and signal distortion associated with a ferromagnetic prosthesis in the right leg.

For this reason, all external metallic objects such as jewelry and watches should be removed from the patient, and a change of clothing is advised to avoid problems with metallic zippers, etc.

Any implanted ferromagnetic material obviously has to be tolerated, but gradient-echo scans should be avoided in such cases since they are affected to a greater extent than spin echoes. Owing to their conductive properties, some nonferrous metal implants can disturb the magnetic field by low-level eddy currents.

Artifacts can also be caused by the ferromagnetic pigments used in eye and other makeup (e.g., mascara) and tattoos. This can result in a significantly reduced image quality, particularly in the case of studies of the orbit.

17-02-02 Susceptibility Artifacts

The susceptibility of a tissue tells us how easily it can be magnetized. The susceptibility values for most tissues fall within a fairly narrow range.

However, the presence of ferromagnetic material (e.g., localized concentrations of hemoglobin after a hemorrhage or high concentration of ferromagnetic contrast agents) or tissue-air interfaces lead to local variations in the susceptibility which result in a reduction of the quality of the local field (Figure 17-05).

Figure 17-05:
Susceptibility artifacts created by ingested ferromagnetic particles used as an oral contrast agent. Because the concentration of the particles is too high, the local magnetic field is disturbed and image artifacts are created.

The form of the susceptibility artifact depends on the local conditions, and both increases and decreases in signal intensity are possible. Tissue-air interfaces which give rise to such artifacts can be found in the lungs, around the sinuses and in the nasopharynx. The lack of signal from the lungs is caused by the air in the lungs and the susceptibility artifacts produced by the interfaces between air and lung tissue. The distortions increase with field strengths and can become a nuisance at high and ultra-high fields [⇒ Farahani].

The effect of obtrusive susceptibility artifacts is reduced by using spin-echo sequences rather than gradient-echo sequences. The presence of local field variations can be determined by using either a phase image or a special pulse sequence which exploits the interaction between two echoes to produce a local field map.

In contrast-enhanced MR angiography contrast agents can induce susceptibility artifacts. If these artifacts are severe, they can be reduced by acquiring the full k-space; partial (asymmetric) echo sampling (and/or halfscan) should be reduced or avoided.

spaceholder gray