RadioGraphics
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

DOI: 10.1148/rg.273065087
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lee, M.-J.
Right arrow Articles by Suh, J.-S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Lee, M.-J.
Right arrow Articles by Suh, J.-S.
Related Collections
Right arrow Computed Tomography
Right arrow Magnetic Resonance Imaging
Right arrow Musculoskeletal Radiology
RadioGraphics 2007;27:791-803
© RSNA, 2007

EDUCATION EXHIBIT

Overcoming Artifacts from Metallic Orthopedic Implants at High-Field-Strength MR Imaging and Multi-detector CT1

Mi-Jung Lee, MD, Sungjun Kim, MD, Sung-Ah Lee, MD, Ho-Taek Song, MD, Yong-Min Huh, MD, Dae-Hong Kim, PhD, Seung Hwan Han, MD, and Jin-Suck Suh, MD

1 From the Department of Diagnostic Radiology and Research Institute of Radiological Science (M.J.L., S.K., H.T.S., Y.M.H., J.S.S.) and Brain Korea 21 Project for Medical Science (J.S.S.), Yonsei University College of Medicine, 134 Sincheon-Dong Seodaemun-Gu, Seoul 120-752, Republic of Korea; Department of Diagnostic Radiology, Hanyang University College of Medicine, Kuri City, Kyunggi-do, Republic of Korea (S.K.); Department of Diagnostic Radiology, Seoul Medical Center, Seoul, Republic of Korea (S.A.L.); Korea Basic Science Institute, Daejeon, Republic of Korea (D.H.K.); and Department of Orthopedic Surgery, Ajou University College of Medicine, Suwon, Republic of Korea (S.H.H.). Presented as an education exhibit at the 2005 RSNA Annual Meeting. Received May 3, 2006; revision requested August 22 and received October 5; accepted November 13. All authors have no financial relationships to disclose. Address correspondence to J.S.S. (e-mail: jss{at}yumc.yonsei.ac.kr).


   Abstract
Top
 Abstract
Introduction
 High-Field-Strength MR Imaging
 Multidetector CT
 Summary
 References
 
At magnetic resonance (MR) imaging and multidetector computed tomography (CT), artifacts arising from metallic orthopedic hardware are an obstacle to obtaining optimal images. Although various techniques for reducing such artifacts have been developed and corroborated by previous researchers, a new era of more powerful MR imaging and multidetector CT modalities has renewed the importance of a systematic consideration of methods for artifact reduction. Knowledge of the factors that contribute to artifacts, of related theories, and of artifact reduction techniques has become mandatory for radiologists. Factors that affect artifacts on MR images include the composition of the metallic hardware, the orientation of the hardware in relation to the direction of the main magnetic field, the strength of the magnetic field, the pulse sequence type, and other MR imaging parameters (mainly voxel size, which is determined by the field of view, image matrix, section thickness, and echo train length). At multidetector CT, the factors that affect artifacts include the composition of the hardware, orientation of the hardware, acquisition parameters (peak voltage, tube charge, collimation, and acquired section thickness), and reconstruction parameters (reconstructed section thickness, reconstruction algorithm used, and whether an extended CT scale was used). A comparison of images obtained with different hardware and different acquisition and reconstruction parameters facilitates an understanding of methods for reducing or overcoming artifacts related to metallic implants.

© RSNA, 2007


   Introduction
Top
 Abstract
 Introduction
High-Field-Strength MR Imaging
 Multidetector CT
 Summary
 References
 
With the continual development of new state-of-the-art imaging equipment, the clinical use of high-field-strength magnetic resonance (MR) imaging and multidetector computed tomography (CT) is increasing in the field of musculoskeletal radiology. However, in patients who have metallic orthopedic implants, artifacts due to misregistration at MR imaging and beam hardening at CT often have prevented the accurate evaluation of regions of interest near the implants (1). Moreover, the use of high magnetic field strengths at MR imaging produces more obtrusive artifacts than does the use of lower field strengths. In contrast, multidetector CT performed on a scanner with more than four channels does not produce more pronounced artifacts than CT performed on a scanner with fewer channels. For musculoskeletal radiologists, an understanding of artifact reduction techniques that may be used at MR imaging and CT is increasingly important. In this article, we survey the factors that affect metal implant–related artifacts and review the theories and techniques of artifact reduction at 3.0-T MR imaging and multidetector CT.


   High-Field-Strength MR Imaging
Top
 Abstract
 Introduction
 High-Field-Strength MR Imaging
Multidetector CT
 Summary
 References
 
Magnetic Properties of a Material
In general, when a material object is placed within a magnetic field, the magnetic forces associated with the electrons of the object are affected. In the presence of an external magnetic field, different materials react differently, with the reaction depending on factors such as the atomic and molecular structure of the material and the net magnetic field associated with the atoms of the material.

Most materials can be classified as diamagnetic, paramagnetic, superparamagnetic, or ferromagnetic (2). The magnetic properties of substances are generally considered in relation to those of water, which has extremely weak (nearly zero) diamagnetism.

Paramagnetic materials, which have unpaired electrons, concentrate local magnetic forces and thus increase the local magnetic field, an effect that is referred to as magnetic susceptibility. Superparamagnetic materials contain particles with a much stronger magnetic susceptibility than that of paramagnetic materials. Ferromagnetic materials contain large solid or crystalline aggregates of molecules with unpaired electrons and are subject to an effect known as "magnetic memory," by which a lingering magnetic field is created after their exposure to an external magnetic field. Examples of ferromagnetic metals include iron, nickel, and cobalt, all of which distort magnetic fields, thereby causing severe artifacts on MR images (1).

Mechanism of Artifact Generation
Matter in a constant magnetic induction field (B0) induces a change in the magnetic flux density. When an object is placed within a homogeneous magnet, the object produces inhomogeneities in the local magnetic field that interfere with the imaging gradient field. The resultant imaging distortions are called magnetic susceptibility effects (3,4). Artifacts on MR images obtained in patients with metallic implants are produced by the large differences between the magnetic properties of human tissues and those of the implanted metals (5). The artifacts are more marked when the differences in magnetic susceptibilities between the metallic object and the surrounding matter are substantial. In this regard, ferromagnetic materials, which have high magnetic susceptibilities, produce the largest artifacts (3). In contrast, paramagnetic and diamagnetic substances are far less likely to cause artifacts (4).

Differences in the magnetic susceptibilities of adjacent tissues and implants create local magnetic field inhomogeneities, altering the phase and frequency of local spins. Thus, the spins are subsequently mapped to an erroneous location within the image. The results are distortion of the shape of the metallic object along the axes of frequency encoding and section selection, and loss of signal within the metallic object. A rim of high signal intensity appears around the metallic object as a result of the mismapping of a disproportionate number of spins to that location (6) (Fig 1). The resultant misregistration effect is especially exaggerated in the direction of frequency encoding (7,8). Hence, at clinical MR imaging of a joint, if the frequency encoding direction is chosen optimally, away from the intraarticular structures of interest, those structures will be less obscured by artifacts (9). For example, to obtain sagittal images for optimal evaluation of the posterior cruciate ligament in a patient with a metallic screw located at the middle of the posterior tibial cortex, near the tibial plateau, the frequency encoding direction should be anterior to posterior.


Figure 1A
View larger version (27K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1a.  Schematics show the mechanism that causes metal-related artifacts at MR imaging. (a) The spins in the main magnetic field predictably rotate at the same frequency. When the frequency encoding gradient is applied, the Larmor frequencies of the spins change along the gradient direction. The spins at locations of higher magnetic field strength (upper part of the triangle) rotate at higher frequencies. When a metallic object is placed in the region of interest (arrow inside the circle), the spins near the object rotate with higher frequency than those that are distant from it. (b) The region of interest near the metallic object is encoded as if it were at a higher gradient location (upper circle) than it actually is. The signal around the region of interest is summed at a higher gradient location (+) and subtracted at a lower one (–). The result is geometric distortion of the object.

 

Figure 1B
View larger version (24K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1b.  Schematics show the mechanism that causes metal-related artifacts at MR imaging. (a) The spins in the main magnetic field predictably rotate at the same frequency. When the frequency encoding gradient is applied, the Larmor frequencies of the spins change along the gradient direction. The spins at locations of higher magnetic field strength (upper part of the triangle) rotate at higher frequencies. When a metallic object is placed in the region of interest (arrow inside the circle), the spins near the object rotate with higher frequency than those that are distant from it. (b) The region of interest near the metallic object is encoded as if it were at a higher gradient location (upper circle) than it actually is. The signal around the region of interest is summed at a higher gradient location (+) and subtracted at a lower one (–). The result is geometric distortion of the object.

 
Factors That Affect Artifacts
Factors that influence the production of metal-related artifacts at MR imaging include the composition, size, and orientation of the metallic object with regard to the direction of the external magnetic field; the type of pulse sequences applied; and the sequence parameters, including magnetic field strength, voxel size (determined by the field of view, image matrix, and section thickness), and echo train length. To demonstrate these factors, we performed MR imaging in a phantom and a patient.

The phantom consisted of three orthopedic metallic screws positioned inside a rectangular plastic container filled with 2% agarose gel. One screw was made of a titanium alloy (4.5-mm-diameter Cortical Bone Screws; Howmedica, Rutherford, NJ) and two were made of stainless steel (3.5- and 4.5-mm-diameter Cortical Bone Screws; A-O Synthes, Paoli, Pa). MR imaging of the phantom was performed with a 3.0-T scanner (Intera Achieva; Philips Medical Systems, Amsterdam, the Netherlands) by using a coil produced by the scanner manufacturer (SENSE Flex-S; Philips Medical Systems). Axial images were obtained with systematic variation of pulse sequences and screw orientations (Figs 2, 3).


Figure 2A
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2a.  Comparison of metal-related artifacts according to the physical composition and size of the metallic object and the MR pulse sequence used. (a) Schema shows the relative positions of a titanium-alloy screw (diameter, 4.5 mm) and two stainless steel screws (diameters, 3.5 and 4.5 mm) within the phantom. (b, c) Axial MR images of the phantom, obtained with a gradient-recalled echo (GRE) sequence (repetition time msec/echo time msec, 500/14; flip angle, 80°) (b) and a fast spin-echo (SE) sequence (500/14; echo train length, 16) (c), show that the titanium alloy screws produced smaller artifacts than did the stainless steel screws, regardless of the sequence used. The artifacts are smaller on c than on b. The arrow in b shows the direction of the frequency encoding gradient.

 

Figure 2B
View larger version (77K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2b.  Comparison of metal-related artifacts according to the physical composition and size of the metallic object and the MR pulse sequence used. (a) Schema shows the relative positions of a titanium-alloy screw (diameter, 4.5 mm) and two stainless steel screws (diameters, 3.5 and 4.5 mm) within the phantom. (b, c) Axial MR images of the phantom, obtained with a gradient-recalled echo (GRE) sequence (repetition time msec/echo time msec, 500/14; flip angle, 80°) (b) and a fast spin-echo (SE) sequence (500/14; echo train length, 16) (c), show that the titanium alloy screws produced smaller artifacts than did the stainless steel screws, regardless of the sequence used. The artifacts are smaller on c than on b. The arrow in b shows the direction of the frequency encoding gradient.

 

Figure 2C
View larger version (92K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2c.  Comparison of metal-related artifacts according to the physical composition and size of the metallic object and the MR pulse sequence used. (a) Schema shows the relative positions of a titanium-alloy screw (diameter, 4.5 mm) and two stainless steel screws (diameters, 3.5 and 4.5 mm) within the phantom. (b, c) Axial MR images of the phantom, obtained with a gradient-recalled echo (GRE) sequence (repetition time msec/echo time msec, 500/14; flip angle, 80°) (b) and a fast spin-echo (SE) sequence (500/14; echo train length, 16) (c), show that the titanium alloy screws produced smaller artifacts than did the stainless steel screws, regardless of the sequence used. The artifacts are smaller on c than on b. The arrow in b shows the direction of the frequency encoding gradient.

 

Figure 3A
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3a.  Effect of hardware orientation on metal-related artifacts at fast SE MR imaging. (a) Schematics demonstrate the orientation of a screw in the vertical position (0°) and at successive angles of 30°, 60°, and 90°. (b) Coronal T1-weighted fast SE images (500/17; echo train length, three), obtained in the phantom with sequential changes in the angular orientation of the screw (as specified in a), show increases in artifact size that are proportional to the increase in the angle between the long axis of the screw and the direction of the main magnetic field.

 

Figure 3B
View larger version (46K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3b.  Effect of hardware orientation on metal-related artifacts at fast SE MR imaging. (a) Schematics demonstrate the orientation of a screw in the vertical position (0°) and at successive angles of 30°, 60°, and 90°. (b) Coronal T1-weighted fast SE images (500/17; echo train length, three), obtained in the phantom with sequential changes in the angular orientation of the screw (as specified in a), show increases in artifact size that are proportional to the increase in the angle between the long axis of the screw and the direction of the main magnetic field.

 
One patient, who had a stainless steel screw in his knee as a result of reconstructive surgery of the anterior cruciate ligament, participated as a volunteer in this study. MR imaging was performed mainly with the 3.0-T scanner described earlier, but a 1.5-T scanner made by the same manufacturer (Intera Achieva; Philips Medical Systems) also was used to enable a comparison of artifacts produced with the two different magnet strengths (1.5-T and 3.0-T). A dedicated coil (SENSE knee coil; Philips Medical Systems) was used. Sagittal images were obtained with systematic variation of the parameters described earlier as factors that affect metal-related artifacts. Reduction of an artifact was defined as a reduction of distortion due to misregistration.

Composition and Size of Orthopedic Hardware
Metal-related artifacts vary according to the metallic composition of the orthopedic hardware. Implants made of titanium alloy are nonferro-magnetic and produce much less severe artifacts than do ferromagnetic implants made of stainless steel (Fig 2) (4,9,10). In addition, the artifact size is affected by the implant size, with larger implants producing more obtrusive artifacts (9,10).

Orientation of Orthopedic Hardware
The angle between the long axis of a metallic screw and the direction of the main magnetic field may affect the severity of metal-related artifacts on MR images: Placement of the screw parallel to the axis of the main magnetic field (B0) helps reduce the size of the susceptibility artifact (Fig 3) (3,8–14). The artifact size increases proportionally with an increase in the angle between the long axis of the screw and the main magnetic field direction (9). In addition, the shape of susceptibility artifacts changes with the angle of the long axis of the screw. For example, when the long axis of the screw is perpendicular to the main magnetic field, the artifact has a cloverleaf shape (Fig 2), whereas it is round or oval when the long axis of the screw parallels the direction of the main magnetic field (9). On successive images acquired along the long axis of a screw, the artifacts vary in shape and size, and those variations are obviously exaggerated with an increasing angle (Fig 3).

Imaging Parameters That Affect Artifacts
Pulse Sequence.— The severity of the image distortion produced by magnetic susceptibility effects depends also on the type of MR pulse sequence used. Magnetic susceptibility effects on spins near a metal object in a magnetic field result in two main types of image distortion: (a) distortion of the spatial geometry and (b) loss of signal in the object. The first type of distortion, spatial misregistration, is visible on both conventional SE and GRE images, whereas the second type (loss of signal intensity because of intravoxel dephasing, or the so-called T2* effect) appears only on GRE images (3,4). Unlike SE sequences, GRE sequences include no 180° refocusing pulse. As a consequence, GRE sequences allow no correction for large and fixed magnetic field inhomogeneities induced by metallic implants. Such magnetic field inhomogeneities cause marked intravoxel dephasing (T2* effect) and result in local signal loss (Fig 2). At MR imaging performed with SE sequences, a 180° refocusing pulse enables recovery of the transverse signal lost because of static magnetic field inhomogeneities and bulk susceptibility differences (2,7–9).

Substantial magnetic field inhomogeneities such as those caused by bulk ferromagnetic materials result in additional local dephasing of the spins of hydrogen protons in randomly diffusing water molecules, and this diffusion-related dephasing is not recoverable with the application of a 180° refocusing pulse. This type of dephasing, which is accentuated by imaging sequences with long echo times (eg, T2-weighted sequences), contributes directly to the overall local signal loss in the vicinity of metallic hardware (2,7–9). Fast SE sequences with short echo spacing (short time intervals between echoes) are less sensitive to magnetic susceptibility effects than are fast SE sequences with longer echo spacing or conventional SE sequences (3,8,9,15,16). However, fast SE sequences that have long echo trains may still be vulnerable to magnetic susceptibility effects.

Frequency-selective fat saturation, also referred to as spectral fat saturation, is commonly used to suppress the signal from fat at musculo-skeletal imaging. Frequency-selective fat saturation relies on the different resonance frequencies of hydrogen protons within water and fat. Fat signal suppression is achieved with the application of a narrow-bandwidth radiofrequency pulse limited to the spectral frequency of fat, and the magnetic field must be homogeneous within the imaging volume in order to obtain uniform fat suppression within the field of view. Variation of the regional magnetic field surrounding metallic devices or debris creates an inhomogeneous magnetic field, with resultant areas of suboptimal fat saturation (2). Short inversion time inversion recovery (STIR) imaging is an effective alternative method of fat signal suppression and is less dependent on the homogeneity of the main magnetic field (Fig 4) (2,5).


Figure 4A
View larger version (171K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4a.  Effect of fat signal suppression method on metal-related artifacts at MR imaging. A comparison of images obtained with a frequency-selective fat saturation sequence (3000/62) (a) and with a STIR sequence (repetition time msec/echo time msec/inversion time msec, 3000/33/210) (b) shows a smaller artifact on the STIR image.

 

Figure 4B
View larger version (185K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4b.  Effect of fat signal suppression method on metal-related artifacts at MR imaging. A comparison of images obtained with a frequency-selective fat saturation sequence (3000/62) (a) and with a STIR sequence (repetition time msec/echo time msec/inversion time msec, 3000/33/210) (b) shows a smaller artifact on the STIR image.

 
Magnetic Field Strength.— High-field-strength magnets produce larger magnetic susceptibility artifacts (Fig 5) (3,8). Misregistration artifacts seen in the frequency-encoding direction are proportional to the magnitude of the local inhomogeneity in the main magnetic field and inversely proportional to the strength of the applied frequency-encoding gradient. However, at high magnetic field strengths, broader receiver bandwidth and consequently higher gradient field strengths commonly are used. Thus, the increased distortion effects of a higher main magnetic field strength could be offset by higher gradient strengths (17).


Figure 5A
View larger version (173K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5a.  Effect of magnetic field strength on metal-related artifacts. A comparison of T2-weighted fast SE images (3600/90; echo train length, 10) obtained at 1.5 T (a) and at 3.0 T (b) shows a smaller artifact on a, the image obtained with lower magnetic field strength.

 

Figure 5B
View larger version (172K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5b.  Effect of magnetic field strength on metal-related artifacts. A comparison of T2-weighted fast SE images (3600/90; echo train length, 10) obtained at 1.5 T (a) and at 3.0 T (b) shows a smaller artifact on a, the image obtained with lower magnetic field strength.

 
Voxel Size.— The voxel size is determined by the field of view, image matrix, and section thickness selected at image acquisition. The use of a small voxel size increases the spatial definition of a signal distortion surrounding a metallic device but has little if any effect on the apparent size of the corresponding artifact. Furthermore, the use of a small voxel size helps reduce diffusion-related signal loss in the vicinity of metallic hardware (2,8,9). Hence, the use of a small field of view, high-resolution matrix (eg, 256 x 256 or 512 x 512), thin section, and high gradient strength can help reduce metal-related artifacts (Figs 6–8) (14). Nonetheless, the most important parameter for reducing such artifacts is not the voxel size but the value of FOV/Nx, a term that denotes spatial resolution along the frequency encoding axis (FOV being the size of the field of view, and Nx being the number of pixels in the frequency encoding direction).


Figure 6A
View larger version (171K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 6a.  Effect of field of view on metal-related artifacts. A comparison of T2-weighted fast SE images (3600/90; echo train length, 10) obtained on a 3.0-T system with a field of view of 14 cm (a) and 24 cm (b) shows a smaller artifact on a, the image obtained with a smaller field of view.

 

Figure 6B
View larger version (165K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 6b.  Effect of field of view on metal-related artifacts. A comparison of T2-weighted fast SE images (3600/90; echo train length, 10) obtained on a 3.0-T system with a field of view of 14 cm (a) and 24 cm (b) shows a smaller artifact on a, the image obtained with a smaller field of view.

 

Figure 7A
View larger version (128K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 7a.  Effect of image matrix on metal-related artifacts. A comparison of T2-weighted fast SE images (3600/90; echo train length, 10) obtained with matrix dimensions of 128 x 126 (a), 512 x 508 (b), and 1024 x 1016 (c) shows a smaller artifact on b and c, the images obtained with a high-resolution matrix.

 

Figure 7B
View larger version (140K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 7b.  Effect of image matrix on metal-related artifacts. A comparison of T2-weighted fast SE images (3600/90; echo train length, 10) obtained with matrix dimensions of 128 x 126 (a), 512 x 508 (b), and 1024 x 1016 (c) shows a smaller artifact on b and c, the images obtained with a high-resolution matrix.

 

Figure 7C
View larger version (159K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 7c.  Effect of image matrix on metal-related artifacts. A comparison of T2-weighted fast SE images (3600/90; echo train length, 10) obtained with matrix dimensions of 128 x 126 (a), 512 x 508 (b), and 1024 x 1016 (c) shows a smaller artifact on b and c, the images obtained with a high-resolution matrix.

 

Figure 8A
View larger version (148K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 8a.  Effect of section thickness on metal-related artifacts. A comparison of T2-weighted fast SE images (3600/90; echo train length, 10) obtained with section thicknesses of 1 mm (a), 4 mm (b), and 7 mm (c) shows reduced artifacts on a and b (a shows the least artifact), the images obtained with thinner sections.

 

Figure 8B
View larger version (142K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 8b.  Effect of section thickness on metal-related artifacts. A comparison of T2-weighted fast SE images (3600/90; echo train length, 10) obtained with section thicknesses of 1 mm (a), 4 mm (b), and 7 mm (c) shows reduced artifacts on a and b (a shows the least artifact), the images obtained with thinner sections.

 

Figure 8C
View larger version (143K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 8c.  Effect of section thickness on metal-related artifacts. A comparison of T2-weighted fast SE images (3600/90; echo train length, 10) obtained with section thicknesses of 1 mm (a), 4 mm (b), and 7 mm (c) shows reduced artifacts on a and b (a shows the least artifact), the images obtained with thinner sections.

 
Echo Train Length.— The set of echoes that follows the application of an excitation pulse is referred to as an echo train. The period during which these echoes are acquired is the echo train duration, and the number of echoes in the echo train is the echo train length. An increase in the echo train length, with a constant echo train duration, leads to a reduction of metal-related artifacts (Fig 9) (1).


Figure 9A
View larger version (174K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 9a.  Effect of echo train length on metal-related artifacts. A comparison of T2-weighted fast SE images (3600/90) obtained with echo train lengths of four (a), 10 (b), and 20 (c) shows reduced artifacts on b and c, the images obtained with increased echo train length.

 

Figure 9B
View larger version (167K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 9b.  Effect of echo train length on metal-related artifacts. A comparison of T2-weighted fast SE images (3600/90) obtained with echo train lengths of four (a), 10 (b), and 20 (c) shows reduced artifacts on b and c, the images obtained with increased echo train length.

 

Figure 9C
View larger version (168K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 9c.  Effect of echo train length on metal-related artifacts. A comparison of T2-weighted fast SE images (3600/90) obtained with echo train lengths of four (a), 10 (b), and 20 (c) shows reduced artifacts on b and c, the images obtained with increased echo train length.

 

   Multidetector CT
Top
 Abstract
 Introduction
 High-Field-Strength MR Imaging
 Multidetector CT
Summary
 References
 
Mechanism of Artifact Generation
An x-ray beam is composed of individual photons with a range of energies. As the beam passes through an object, it becomes "harder"; that is, its mean energy increases because lower-energy photons are absorbed more rapidly than are higher-energy photons. A beam hardening artifact, which appears as dark bands or streaks near a dense object, is the result (18).

Metallic hardware causes severe beam hardening and dramatically attenuates the x-ray beam. Thus, metallic hardware significantly degrades image quality to the extent that the resultant image is either incomplete or is a faulty projection of the data with consequent reconstruction artifacts (18,19).

Factors That Affect Artifacts
The factors that may contribute to or diminish the production of artifacts at multidetector CT include metallic hardware composition and orientation, image acquisition parameters (peak voltage, tube charge in milliampere-seconds, collimation, and section thickness), and image reconstruction parameters (reconstructed section thickness, reconstruction algorithm [kernel], and extended CT scale). To demonstrate and evaluate these factors, we performed a study with two phantoms.

The first phantom was constructed from a 3.5-mm-diameter stainless steel orthopedic screw (Cortical Bone Screw; A-O Synthes, Paoli, Pa) and a rectangular plastic container filled with 2% agarose gel. A second phantom consisted of a 4.5-mm-diameter titanium alloy screw and two stainless steel screws with diameters of 3.5 mm and 4.5 mm that had been placed in a pig femur. Helical CT was performed by using a 64-channel multidetector scanner (Somatom Sensation 64; Siemens, Erlangen, Germany) with variation of the scanning parameters. The maximum window width for reconstructions with the standard window was 4000 HU, and that for reconstructions with the extended CT scale was 40,000 HU. A fixed window level of 400 HU and fixed window width of 2000 HU were used for reviewing images. Decreased conspicuity and smoothing of a beam hardening artifact was considered reduction of the artifact.

Orthopedic Hardware Composition
The specific metallic content of an implant may affect the severity of artifacts on CT images (20). Titanium alloy hardware causes the least obtrusive artifact at CT imaging, whereas stainless steel implants cause significant beam attenuation and artifact (Fig 10) (18,20). Knowledge of the composition of the implanted material at the time of the CT examination may be helpful, as technical parameters then may be adjusted to minimize artifacts and to spare the patient excess radiation (18).


Figure 10A
View larger version (78K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 10a.  Effect of metal composition on metal-related artifacts at multidetector CT. (a) Scout view of an ex vivo model of a pig femur with a 4.5-mm titanium alloy screw (arrow) and a 4.5-mm stainless steel screw (arrowhead). (b) CT image obtained through the short axis of the screws, with 140 kVp, 120 mAs, 0.6-mm collimation, and 1.0-mm reconstructed section thickness, shows more severe streak artifacts from the stainless steel screw than from the titanium alloy screw.

 

Figure 10B
View larger version (62K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 10b.  Effect of metal composition on metal-related artifacts at multidetector CT. (a) Scout view of an ex vivo model of a pig femur with a 4.5-mm titanium alloy screw (arrow) and a 4.5-mm stainless steel screw (arrowhead). (b) CT image obtained through the short axis of the screws, with 140 kVp, 120 mAs, 0.6-mm collimation, and 1.0-mm reconstructed section thickness, shows more severe streak artifacts from the stainless steel screw than from the titanium alloy screw.

 
Orientation and Thickness of Orthopedic Hardware
In theory, the thickness of the hardware has a direct effect on the degree of x-ray beam attenuation. In other words, a smaller angle between the long axis of the gantry and the metallic implant can help reduce artifacts (18). The axis of a metallic implant in the human body cannot always be adjusted at imaging; however, to the extent that adjustment is possible, the axis of the metallic implant should be aligned so that the x-ray beam traverses the smallest possible cross-sectional area of the implant.

Imaging Parameters That Affect Artifacts
With the optimization of imaging and reconstruction parameters, metal-related artifacts may be reduced. However, artifact reduction that results from alteration of a particular imaging parameter is not always conspicuous; some parameters do not seem to affect artifacts as much as others do. We therefore performed an additional comparative study of a phantom with a better and a poorer combination of parameters to identify those that effect a significant difference (Fig 11). CT images were obtained through the short axis of the screw in an agarose gel phantom. It was found that artifacts were reliably minimized with the better parameters. The optimal parameters for minimizing artifacts are described in the following sections.


Figure 11A
View larger version (52K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 11a.  CT images obtained through the short axis of a stainless steel screw in an agarose gel phantom. Image obtained with optimal parameters (140 kVp, 300 mAs, 0.6-mm collimation, 3.0-mm reconstructed section thickness, a low kernel value [B30f], and the extended CT scale) (a) shows a less pronounced artifact than that in b, which was obtained with less advantageous parameters (80 kVp, 80 mAs, 1.2-mm collimation, 1.5-mm acquired section thickness, 1.5-mm reconstructed section thickness, and a high kernel value [B70f], without the extended CT scale).

 

Figure 11B
View larger version (98K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 11b.  CT images obtained through the short axis of a stainless steel screw in an agarose gel phantom. Image obtained with optimal parameters (140 kVp, 300 mAs, 0.6-mm collimation, 3.0-mm reconstructed section thickness, a low kernel value [B30f], and the extended CT scale) (a) shows a less pronounced artifact than that in b, which was obtained with less advantageous parameters (80 kVp, 80 mAs, 1.2-mm collimation, 1.5-mm acquired section thickness, 1.5-mm reconstructed section thickness, and a high kernel value [B70f], without the extended CT scale).

 
Peak Voltage.— Technical scanning factors affect the ability of the x-ray beam to penetrate metal, and therefore they influence the production of image artifacts. Increasing the peak voltage increases the likelihood that x-rays will penetrate the metallic hardware (18); thus, artifacts are expected to be diminished with higher peak voltage (Fig 12). However, higher peak voltage may only marginally improve the penetration of steel orthopedic devices by x-rays, and this remedy may not appreciably affect the imaging of large items such as hip implants and metal plates. Moreover, some authors assert that there is no additional value in using exposure settings greater than 120 kVp at 340 mAs when acquiring axial CT images with a scanner with four or fewer channels in patients with orthopedic hardware (20). More investigation of this question seems necessary with regard to multidetector CT scanners with a larger number of channels.


Figure 12A
View larger version (63K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 12a.  Effect of peak voltage on metal-related artifacts. A comparison of CT images obtained through the short axis of the screws in the pig femur model, with 140 kVp (a) and 80 kVp (b) and with other parameters the same (120 mAs, 0.6-mm collimation, and 1.0-mm reconstructed section thickness), shows a less obtrusive artifact on a, the image obtained with a higher peak voltage.

 

Figure 12B
View larger version (67K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 12b.  Effect of peak voltage on metal-related artifacts. A comparison of CT images obtained through the short axis of the screws in the pig femur model, with 140 kVp (a) and 80 kVp (b) and with other parameters the same (120 mAs, 0.6-mm collimation, and 1.0-mm reconstructed section thickness), shows a less obtrusive artifact on a, the image obtained with a higher peak voltage.

 
Tube Charge.— A potential source of severe streak artifacts is photon starvation, which may occur in areas that cause high beam attenuation, such as the shoulders. When the x-ray beam is traveling through a dense object, the attenuation is greatest, and an insufficient number of photons reaches the detector. The resultant projections are very noisy at these tube angulations, and the reconstruction process greatly magnifies the noise, which results in horizontal streaks on the image. If the tube charge (in milliampere-seconds) is increased for the duration of scanning, the problem of photon starvation is overcome and artifacts are reduced. However, the patient inevitably receives a higher radiation dose if the beam passes through a region with lower attenuation (18).

Although in theory the alteration of tube charge should produce a diminished artifact, the decrease is not always discernible (Fig 13). The automatic dose control option available on the new generation of CT scanners provides an efficient approach to tube current modulation along the z-axis and at various projection angles following the patient’s anatomy. With the use of automatic exposure control, the image quality is at least as good as that with conventional scanning, and the dose is significantly reduced.


Figure 13A
View larger version (64K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 13a.  Effect of tube charge on metal-related artifacts. CT images obtained through the short axis of the screws in the pig femur model, with tube charge of 80 mAs (a) and 120 mAs (b) and with other parameters the same (140 kVp, 0.6-mm collimation, and 1.0-mm reconstructed section thickness), show similar streak artifacts.

 

Figure 13B
View larger version (62K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 13b.  Effect of tube charge on metal-related artifacts. CT images obtained through the short axis of the screws in the pig femur model, with tube charge of 80 mAs (a) and 120 mAs (b) and with other parameters the same (140 kVp, 0.6-mm collimation, and 1.0-mm reconstructed section thickness), show similar streak artifacts.

 
Collimation.— The x-ray tube of most CT scanners may be adjusted to obtain narrow collimation (a small focal area) or wide collimation (a larger focal area). The size of the focal area, in conjunction with the image reconstruction algorithm, determines the image resolution (18). Artifacts may be reduced with the use of a narrow collimation setting (Fig 14). In general, thin-section acquisition is expected to help minimize artifacts by reducing the partial volume averaging effect (21). However, in some cases (eg, a thick patient anatomy or significant bone and metal plus tissue), narrow collimation may have limited effectiveness in reducing artifacts.


Figure 14A
View larger version (61K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 14a.  Effect of collimation on metal-related artifacts. A comparison of CT images obtained through the short axis of the screws in the pig femur model, with collimation of 0.6 mm (a) and 1.2 mm (b) and with other parameters the same (140 kVp, 120 mAs, and 1.5-mm reconstructed section thickness), shows a less obtrusive artifact on a, the image obtained with narrow collimation.

 

Figure 14B
View larger version (65K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 14b.  Effect of collimation on metal-related artifacts. A comparison of CT images obtained through the short axis of the screws in the pig femur model, with collimation of 0.6 mm (a) and 1.2 mm (b) and with other parameters the same (140 kVp, 120 mAs, and 1.5-mm reconstructed section thickness), shows a less obtrusive artifact on a, the image obtained with narrow collimation.

 
Image Reconstruction Parameters
Section Thickness.— Partial volume artifacts can best be avoided by acquiring thin sections (18,22). However, during image reconstruction, thicker sections may be generated by combining several thin sections to reduce image noise. The increased thickness of the reformatted sections dramatically affects image quality, decreasing the severity of metal-related artifacts and thereby allowing increased accuracy at image review and interpretation (18).

Reconstruction Algorithm (Kernel).— Selection of an appropriate reconstruction filter may play a critical role in the appearance of a metal-related artifact. For example, such an artifact is accentuated by the use of a bone algorithm in image reconstruction. The use of a standard or smooth reconstruction filter is preferred, particularly in the presence of dense metallic hardware and in patients with a large body habitus (Fig 15); however, the usefulness of smooth reconstruction filters is limited by a consequent reduction of spatial resolution (18).


Figure 15A
View larger version (53K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 15a.  Effect of reconstruction algorithm (kernel) on metal-related artifacts. A comparison of CT images of the short axis of the screws in the pig femur model, obtained with a standard reconstruction algorithm (B30f, medium smooth kernel) (a) and a higher kernel value (B70f, very sharp kernel) (b) and with other parameters the same (140 kVp, 120 mAs, 0.6-mm collimation, and 1.0-mm reconstructed section thickness), shows less severe artifacts but also a loss of sharpness in a.

 

Figure 15B
View larger version (67K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 15b.  Effect of reconstruction algorithm (kernel) on metal-related artifacts. A comparison of CT images of the short axis of the screws in the pig femur model, obtained with a standard reconstruction algorithm (B30f, medium smooth kernel) (a) and a higher kernel value (B70f, very sharp kernel) (b) and with other parameters the same (140 kVp, 120 mAs, 0.6-mm collimation, and 1.0-mm reconstructed section thickness), shows less severe artifacts but also a loss of sharpness in a.

 
Extended CT Scale.— The extended CT scale is an expansion of the Hounsfield scale from a standard maximum window of 4,000 HU to 40,000 HU. This technique makes use of the fact that metals have high linear attenuation coefficients that lie outside the normal range of reconstructed CT numbers. Most metallic implants are in the range of 8,000–20,000 HU, whereas the standard upper limit of CT scanners is 4,096 HU because of their 12-bit storage capacity. The cutoff of very high CT values produces blurring and distortion of the contours of metallic implants on images (22).

With a 5-mm acquired section thickness, differences in diagnostic performance with use of the standard window versus that of the extended CT scale were only moderately significant; however, differences were more significant with a 2-mm section thickness. Thus, use of the extended CT scale may be more effective with thin-section acquisitions (22). The extended CT scale is not available on every scanner, but its use when available helps reduce metal-related artifacts by allowing for a window width as large as 40,000 HU (18,22). Currently available picture archiving and communication systems offer radiologists the potential to improve the accuracy of CT image interpretation beyond the level achievable with traditional film-based display, and the unlimited CT scale available on standard monitors with these systems has reduced the usefulness of the extended CT scale (23).


   Summary
Top
 Abstract
 Introduction
 High-Field-Strength MR Imaging
 Multidetector CT
 Summary
References
 
The usefulness of MR imaging and multidetector CT to answer clinical questions about nonunion, prosthetic loosening, infection, and tumor recurrence may be limited by artifacts on images obtained in patients with metallic hardware implants near the region of interest. However, such artifacts can be minimized by positioning the patient for optimal orientation of metallic implants and by using optimal modality-specific parameters for image acquisition and reconstruction. With these methods, depiction of the region of interest near the metallic implant may be significantly improved.

To reduce metal-related artifacts at MR imaging, orthopedic hardware should be positioned to parallel as closely as possible the direction of the main magnetic field. With respect to hardware composition, a titanium alloy produces less severe artifacts than does stainless steel. The fast SE pulse sequence is the best MR imaging sequence for artifact reduction, and the GRE sequence is the least beneficial. For purposes of suppressing the signal from fat while avoiding severe metal-related artifacts, the STIR pulse sequence is preferable to frequency-selective fat saturation. Use of lower magnetic field strength is desirable; however, if a currently available clinical MR imaging system with a high-field-strength magnet is used, the imaging parameters chosen (eg, small field of view, high-resolution image matrix, thin sections, increased echo train length, and higher gradient strength for small voxel sizes) may help reduce metal-related artifacts.

At CT, orthopedic hardware composed of a titanium alloy produces a less severe artifact than does stainless steel hardware. The orthopedic implant should be positioned so that the x-ray beam traverses the metallic cross section with the smallest diameter. During image acquisition, the use of a high peak voltage (kilovolts peak), high tube charge (milliampere-seconds), narrow collimation, and thin sections helps reduce metal-related artifacts. During image reconstruction, the use of thick sections, lower kernel values (similar to the standard reconstruction algorithm), and the extended CT scale helps reduce these artifacts.

To the best of our knowledge, there is no absolute contraindication to MR imaging in a patient with an orthopedic appliance. However, MR imaging in a patient with a cardiac pacemaker is contraindicated in most cases.

In the optimization of multidetector CT protocols for artifact reduction, another factor that should be addressed is the radiation hazard; increasing the tube charge may help reduce metal-related artifacts, but this modification inevitably results in an increased radiation dose to the patient. Thus, radiologists also should take the radiation dose into account when optimizing artifact-reducing multidetector CT protocols.


   Footnotes
 

Abbreviations: GRE = gradient-recalled echo, SE = spin echo, STIR = short inversion time inversion recovery


   References
Top
 Abstract
 Introduction
 High-Field-Strength MR Imaging
 Multidetector CT
 Summary
 References
 

  1. White LM, Buckwalter KA. Technical considerations: CT and MR imaging in the postoperative orthopedic patient. Semin Musculoskelet Radiol 2002;6:5–17.[CrossRef][Medline]
  2. Mitchell DG, Cohen MS. Transverse magnetization and T2 contrast. In: Mitchell DG, Cohen MS, eds. MRI principles. 2nd ed. New York, NY: Springer-Verlag, 2000; 35–47.
  3. Guermazi A, Miaux Y, Zaim S, Peterfy CG, White D, Genant HK. Metallic artefacts in MR imaging: effects of main field orientation and strength. Clin Radiol 2003;58:322–328.[CrossRef][Medline]
  4. Eggers G, Rieker M, Kress B, Fiebach J, Dickhaus H, Hassfeld S. Artefacts in magnetic resonance imaging caused by dental material. MAGMA 2005;18:103–111.[CrossRef][Medline]
  5. Viano AM, Gronemeyer SA, Haliloglu M, Hoffer FA. Improved MR imaging for patients with metallic implants. Magn Reson Imaging 2000;18: 287–295.[CrossRef][Medline]
  6. Ludeke KM, Roschmann P, Tischler R. Susceptibility artefacts in NMR imaging. Magn Reson Imaging 1985;3:329–343.[CrossRef][Medline]
  7. Cho ZH, Ro YM. Reduction of susceptibility artifact in gradient-echo imaging. Magn Reson Med 1992;23:193–200.[Medline]
  8. White LM, Kim JK, Mehta M, et al. Complications of total hip arthroplasty: MR imaging—initial experience. Radiology 2000;215:254–262.[Abstract/Free Full Text]
  9. Suh JS, Jeong EK, Shin KH, et al. Minimizing artifacts caused by metallic implants at MR imaging: experimental and clinical studies. AJR Am J Roentgenol 1998;171:1207–1213.[Abstract/Free Full Text]
  10. Ganapathi M, Joseph G, Savage R, Jones AR, Timms B, Lyons K. MRI susceptibility artefacts related to scaphoid screws: the effect of screw type, screw orientation and imaging parameters. J Hand Surg [Br] 2002;27:165–170.[CrossRef][Medline]
  11. Williams DF. Titanium as a metal for implantation. II. Biological properties and clinical applications. J Med Eng Technol 1977;1:266–270.[Medline]
  12. Ladd ME, Erhart P, Debatin JF, Romanowski BJ, Boesiger P, McKinnon GC. Biopsy needle susceptibility artifacts. Magn Reson Med 1996;36:646–651.[CrossRef][Medline]
  13. Camacho CR, Plewes DB, Henkelman RM. Non-susceptibility artifacts due to metallic objects in MR imaging. J Magn Reson Imaging 1995;5:75–88.[Medline]
  14. Tormanen J, Tervonen O, Koivula A, Junila J, Suramo I. Image technique optimization in MR imaging of a titanium alloy joint prosthesis. J Magn Reson Imaging 1996;6:805–811.[Medline]
  15. Tartaglino LM, Flanders AE, Vinitski S, Friedman DP. Metallic artifacts on MR images of the postoperative spine: reduction with fast spin-echo techniques. Radiology 1994;190:565–569.[Abstract/Free Full Text]
  16. Petersilge CA, Lewin JS, Duerk JL, Yoo JU, Ghaneyem AJ. Optimizing imaging parameters for MR evaluation of the spine with titanium pedicle screws. AJR Am J Roentgenol 1996;166:1213–1218.[Abstract/Free Full Text]
  17. Alanen A, Bondestam S, Komu M. Artifacts in MR imaging caused by small quantities of powdered iron. Acta Radiol 1995;36:92–95.[Medline]
  18. Barrett JF, Keat N. Artifacts in CT: recognition and avoidance. RadioGraphics 2004;24:1679–1691.[Abstract/Free Full Text]
  19. Yazdi M, Gingras L, Beaulieu L. An adaptive approach to metal artifact reduction in helical computed tomography for radiation therapy treatment planning: experimental and clinical studies. Int J Radiat Oncol Biol Phys 2005;62:1224–1231.[CrossRef][Medline]
  20. Haramati N, Staron RB, Mazel-Sperling K, et al. CT scans through metal scanning technique versus hardware composition. Comput Med Imaging Graph 1994;18:429–434.[CrossRef][Medline]
  21. Wang G, Frei T, Vannier MW. Fast iterative algorithm for metal artifact reduction in x-ray CT. Acad Radiol 2000;7:607–614.[CrossRef][Medline]
  22. Link TM, Berning W, Scherf S, et al. CT of metal implants: reduction of artifacts using an extended CT scale technique. J Comput Assist Tomogr 2000;24:165–172.[CrossRef][Medline]
  23. Reiner BI, Siegel EL, Hooper FJ. Accuracy of interpretation of CT scans: comparing PACS monitor displays and hard-copy images. AJR Am J Roentgenol 2002;179:1407–1410.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Am. J. Roentgenol.Home page
J. M. Bestic, J. J. Peterson, J. K. DeOrio, L. W. Bancroft, T. H. Berquist, and M. J. Kransdorf
Postoperative Evaluation of the Total Ankle Arthroplasty
Am. J. Roentgenol., April 1, 2008; 190(4): 1112 - 1123.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract