MRI Artifacts: Susceptibility, Chemical Shift, Wraparound

MRI produces a correct image only if two things are perfect: a uniform magnetic field across the whole field of view, and a clean radiofrequency (RF) signal. In reality neither is perfect. A metal implant bends the field, a tissue–air boundary erases signal, patient motion creates ghosts, and signal processing itself produces rings. This is the MRI deep-dive companion to our general artifacts guide.

Classifying artifacts

Bushberg divides MRI artifacts into three broad groups:¹

Machine-related: magnetic field inhomogeneity (global/focal), susceptibility, gradient and RF artifacts.
Patient-related: motion (voluntary/involuntary) and flow (blood, CSF).
Signal-processing-related: chemical shift, wraparound, ringing (Gibbs).

A non-uniform field causes tissues to be mismapped in the image; this is why shimming, self-shielded magnets and regular maintenance are the foundation of MRI quality assurance.¹

Magnetic susceptibility

Magnetic susceptibility is the ratio of a tissue's induced internal magnetization to the external field. If susceptibility varies little across the FOV the field stays uniform; but a sharp change distorts it.¹ The most common change occurs at tissue–air interfaces (lungs, sinuses), causing signal loss from faster T2* dephasing; any metal (ferrous or not) also creates marked distortion in nearby tissue.¹ Interestingly, susceptibility is sometimes diagnostically helpful: the susceptibility effects of hemorrhage products (deoxyhemoglobin, methemoglobin, hemosiderin) help date a bleed as acute/subacute/chronic.¹

Metal or a tissue–air boundary bends the surrounding uniform field; faster T2* dephasing leaves a signal void (dark region) and geometric distortion. Spin-echo (180°) sequences reduce this.¹

In the clinic

Signal loss around a dental filling, an aneurysm clip, or at tissue–air boundaries (paranasal sinuses, temporal bone) can be mistaken for a real lesion or a bleed. The reverse is also true: GRE/SWI sequences deliberately use this effect to reveal microhemorrhages and calcifications. The same artifact is both a pitfall and a diagnostic tool.

Chemical shift

Protons in fat and water precess at slightly different (Larmor) frequencies. Because MRI reads position from frequency, this small frequency difference shifts fat by a few pixels relative to water along the frequency-encoding direction.¹ The result is a dark band on one side of the fat–water boundary and a bright band on the other — classically seen at kidney or vertebral margins. The shift magnitude varies with field strength and the chosen bandwidth.

Because fat protons precess at a slightly different frequency than water, the fat signal shifts along the frequency-encoding direction; a dark band forms on one side of the boundary and a bright band on the other.¹

In the clinic

The dark/bright band at a kidney–fat or vertebral boundary can be mistaken for a real margin. The same physics is turned around: in-/opposed-phase imaging uses chemical shift to detect an adrenal adenoma or hepatic steatosis.

Wraparound

Anatomy outside the field of view (FOV) but within the slice volume is not ignored; it is mapped onto the opposite side.¹ The cause is undersampling (aliasing): signals outside the FOV lie above the Nyquist limit and are interpreted as if they were low-frequency, "wrapping" to the far edge of the image. It is typically seen along the phase-encoding direction; reduced by enlarging the FOV or applying an anti-aliasing saturation pulse outside it.¹

Anatomy that does not fit in the FOV is mapped onto the opposite edge due to undersampling. Remedy: a larger FOV or an anti-aliasing saturation pulse.¹

In the clinic

In small-FOV cervical or extremity exams, anatomy outside the FOV (a nose, the opposite arm) can fold onto the region of interest and obscure or mimic pathology. A larger FOV or a saturation pulse prevents it.

Motion and ghosting

Motion is MRI's most ubiquitous and noticeable artifact: voluntary/involuntary movement and flow (blood, CSF).¹ The long acquisition time of some sequences raises the chance of motion blur and contrast loss. Motion artifacts appear mostly along the phase-encoding direction, because adjacent phase-encoding measurements in k-space are separated by a TR interval (which can be 3,000 ms or longer), so even slight motion changes the recorded phase over that time.¹ The visual result is faint copies arrayed along the phase-encode direction — ghosting. The frequency-encoding direction is less affected. The simplest remedy is to swap the phase and frequency encoding gradients (PEG/FEG) to move the ghosts away from the region of interest.¹

In the clinic

Respiratory and vascular pulsation ghosts can overlie organs along the phase-encoding direction; aortic pulsation, for instance, can create a false appearance that leads to misinterpretation. Swapping the phase and frequency encoding directions moves the ghost away from the region of interest.

Ringing (Gibbs)

Ringing (the Gibbs phenomenon) is a set of faint parallel lines near sharp boundaries.¹ Its cause is the lack of high-frequency signals in k-space needed to represent a sharp transition; so it is more pronounced at small matrix sizes (e.g. 128 instead of 256).¹ It is typically seen at the skull–brain boundary, and increasing the matrix size is the most direct fix.

In the clinic

Ringing along the spinal cord can mimic a false central-canal appearance (a pseudo-syrinx) — a classic pitfall. Increasing the matrix size resolves the lines.

RF and machine

RF interference leaves its own patterns: narrow-band noise produces a zipper pattern perpendicular to the frequency-encoding direction, while broadband noise produces a herringbone pattern.¹ The remedy is proper site planning and RF-shielding the room with a Faraday cage.¹ Because these are machine/environment artifacts, their solutions lie on the engineering and quality-assurance side.

General artifacts guide: What Are Image Artifacts? · CT artifacts: CT Artifacts · Modality principles: How Do Modalities See?

References

Bushberg JT, Seibert JA, Leidholdt EM, Boone JM. The Essential Physics of Medical Imaging, 3rd ed. Lippincott Williams & Wilkins, 2011. §13.5 MR Artifacts (s.475–486): sınıflama — makine/hasta/sinyal işleme (s.475); manyetik duyarlılık (susceptibility) — doku–hava arayüzü ve metal, T2* faz kaybı (s.475–476, Şekil 13-24); hareket artefaktları — faz kodlama yönü, akış (kan/BOS), hayalet (ghosting), PEG/FEG değişimi (s.480, Şekil 13-29); kimyasal kayma (s.480); halkalanma/Gibbs — matris boyutu (s.484, Şekil 13-37); katlanma/wraparound (s.485–486, Şekil 13-38); RF zipper/herringbone ve Faraday kafesi (s.477–478, Şekil 13-44). Sayfa numaraları bu baskıya aittir.
IAEA. Diagnostic Radiology Physics: A Handbook for Teachers and Students (STI/PUB/1564), 2014 — MR fiziği, kalite güvencesi ve artefaktlar (Bölüm 14). iaea.org
Genel artefakt çerçevesi için bkz. Görüntü Artefaktları Nedir?; MR'ın çalışma prensibi için bkz. Modaliteler Nasıl Görür?

Note: This content is for education; for clinical decisions or regulatory compliance, consult a qualified medical physicist and current regulations.

← All articles