DoseSave
DoseSave · Topics · CT Artifacts: Hardening, Metal, Motion and Helical
CT · Artifacts

CT Artifacts: Hardening, Metal, Motion and Helical

CT has the greatest variety of artifacts: the polychromatic beam, reconstruction and fast helical geometry produce many of them together. Following the IAEA classification, we split artifacts into acquisition-, reconstruction- and patient-related, covering each one's physics, appearance and reduction with animations and page citations. This is the CT deep-dive companion to our general artifacts guide.

Advanced · radiologists, technologists & physicists· ~17 min read· Sourced

In CT, not every dark band, bright line or "lesion" belongs to the patient. The polychromatic x-ray beam, the detector array, the reconstruction algorithm and the rotating–translating helical geometry can each produce their own artifact. This is the CT deep-dive companion to our general artifacts guide: it organizes artifacts by the IAEA classification, explains each one's physics, and shows how to reduce it.

Classifying artifacts

The IAEA divides CT artifacts by origin into three groups:2

This split is not just academic: knowing an artifact's origin tells you how to fix it — calibration, parameter choice, or patient management.

Beam hardening

Standard filtered backprojection does not fully account for the polychromatic spectrum; it treats thickness-dependent attenuation with a single "effective" linear attenuation coefficient.1 In a beam crossing dense tissue (compact bone, calcification, metal), low-energy photons are absorbed first, so the beam's average energy rises — it "hardens."1 There are two typical results: a uniform object's value dropping toward its center (cupping), and a dark wedge/webbing band between two dense structures (e.g. bilateral hip prostheses).1 CT scanners pre-harden the beam with added filtration (e.g. ~10 mm Al equivalent); reconstruction also applies a look-up-table (LUT) correction and water-phantom calibration.3

prosthesisprosthesisdark wedge band (webbing)Beam hardening · between two dense objects
Rays between two metallic prostheses are excessively hardened, so a dark band that isn't really there appears between them. The same physics shows up as "cupping" toward the center of a uniform object.1
In the clinic
A dark band running between the two petrous bones at the skull base (the Hounsfield artifact) can give the false impression of a posterior-fossa hypodensity — even an infarct or hemorrhage. Recognizing it as an artifact avoids unnecessary further work-up.

Metal artifact

Streaking occurs when a region's attenuation exceeds the detector's dynamic range.1 When a metal prosthesis absorbs the beam almost completely, this reaches its most severe form — the metal-specific star-burst: bright/dark lines radiating outward from the metal.23 Classic sources are metallic dental fillings, pacemaker/neurostimulator implants, bullet fragments and orthopedic prostheses. Motion of the high-density object (jaw movement, swallowing) markedly amplifies the streaks.1 Reconstruction can partly correct the abnormally high absorption values by thresholding the peak values; modern systems add metal artifact reduction (MAR) algorithms.3

slicemetal prosthesisMetal → star-burstThe prosthesis absorbs almost allof the beam → dynamic range exceeded →radiating false streaks.
Metal exceeds the surrounding tissue's attenuation values many times over; reconstruction turns this into "star" lines radiating outward. Peak thresholding and MAR algorithms reduce the effect.3
In the clinic
Streaks radiating from a hip prosthesis or dental fillings obscure adjacent structures — the soft tissue around the implant, or head-and-neck tumor margins — making them hard to assess. Metal artifact reduction (MAR) algorithms and a suitable gantry angle reduce the effect.

Partial volume effect

With a thick slice, a voxel can span tissues of different density along the z axis; the voxel value becomes the average of their linear attenuation coefficients.2 As a result a small, high-density object can appear larger and lower-density — for example when looking at cortical bone in thick slices.2 Multi-slice CT (MDCT) with thin sections (e.g. ~1.25 mm) has largely reduced this; imaging the same region in a different plane (axial + coronal) also helps separate partial volume from a real finding.14

Partial volume · voxel averagingsoft tissuebonevoxel spanning the boundary→ average of bone + tissue (gray)→ small object looks larger/fainter
A voxel spanning the bone–soft-tissue boundary averages the two and produces an intermediate gray value. A thin section shrinks the voxel and reduces this averaging.2
In the clinic
In a thick slice a small lung nodule can be averaged with the surrounding air and look lower-density than it is — and be read as benign; a thin section reveals the true density (e.g. the high HU of a malignant lesion).1

Motion artifacts

Patient movement, swallowing or breathing produces streaking and blur, most pronounced next to high-density objects.1 Instructing the patient correctly and holding the breath (especially for trunk scans) prevents most of it; but cardiac motion and vessel pulsation cannot be avoided.2 So coronary or aortic scans must be optimized for the best temporal resolution. A clinically critical caution: aortic pulsation can produce an artifact that mimics an aortic dissection that isn't there; if not recognized as an artifact, it can have serious consequences for the patient.2

In the clinic
Aortic pulsation can create a motion-induced false line / double contour (a pseudo-flap) in the ascending aorta and mimic an aortic dissection. ECG-gated acquisition and recognizing it as an artifact prevent a wrong diagnosis.2

Ring artifact

When one or more detector elements respond in a faulty or imbalanced way, the image shows a concentric ring around the isocenter (or a halo around objects).23 In third-generation (rotating-detector) geometry, a single bad element's error falls at the same radius at every angle, so it traces a circle. Imbalance in the detectors' energy dependence creates similar rings.4 The fix is hardware-side: detector calibration, warm-up/equilibration time, and repair of the faulty element.3

image (isocenter at center)Ring artifactdetector arrayone faulty element →same radius at every angle → ring
In rotating-detector geometry a faulty element falls at the same radius in every projection, so it appears as a fixed ring around the isocenter in the image.2

Helical and cone-beam

Helical (spiral) CT — a rotating tube plus a moving table, introduced in 1989 — shortened scan time and enabled volumetric imaging, but brought its own artifacts.2 The best known is the windmill artifact: because of undersampling along z and helical interpolation, a rotating pinwheel pattern appears near high-contrast structures as you scroll through slices.2 At large multi-slice cone angles, z is undersampled and a cone-beam artifact (false connections between dense structures) forms; the fix is a more complete data set / a smaller cone angle.1 These artifacts are managed through pitch and reconstruction choices.

Reducing artifacts

Once the origin is known, the fix follows:

The common principle: recognizing an artifact matters as much as reducing it. A mimic of aortic dissection, or a calcification "enlarged" by partial volume, is told from a real finding safely only through physicist–radiologist collaboration.

Related articles
General artifacts guide: What Are Image Artifacts? · CT parameters: CT Imaging Parameters · Reconstruction: FBP, IR and Deep Learning · CT dose: Dose in CT (CTDI, DLP, SSDE)

References

  1. Bushberg JT, Seibert JA, Leidholdt EM, Boone JM. The Essential Physics of Medical Imaging, 3rd ed. Lippincott Williams & Wilkins, 2011. §10.6 CT Image Artifacts (s.367–370): demet sertleşmesi (Şekil 10-64; çift kalça implantı Şekil 10-65), çizgilenme/metal (Şekil 10-66), görünüm örtüşmesi (Şekil 10-67), kısmi hacim (Şekil 10-68), koni-ışın (Şekil 10-69). Sayfa numaraları bu baskıya aittir.
  2. IAEA. Diagnostic Radiology Physics: A Handbook for Teachers and Students (STI/PUB/1564), 2014. Bölüm 11 (Computed Tomography): artefakt sınıflaması — edinim/rekonstrüksiyon/hasta kaynaklı (s.288); halka artefaktı ve bozuk dedektör elemanları (Şekil 11.25, s.288); kısmi hacim (z ekseni ortalaması, s.288); demet sertleşmesi ve metal artefaktı (s.288–289); sarmal BT ve windmill artefaktı (§11.5.3, s.274–275); hareket — aort pulsasyonunun diseksiyonu taklit etmesi (s.289). iaea.org
  3. Dowsett DJ, Kenny PA, Johnston RE. The Physics of Diagnostic Imaging, 2nd ed. Hodder Arnold, 2006. Bölüm 14 (CT): metal artefakt girişimi — yıldız patlaması (star-burst) ve tepe-değer eşikleme düzeltmesi (s.399, Şekil 14.18c); dedektör düzgünsüzlüğü → halka/halo (s.399); demet sertleşmesi ve LUT düzeltmesi, su fantomuyla kalibrasyon (s.395).
  4. Hendee WR, Ritenour ER. Medical Imaging Physics, 4th ed. Wiley-Liss, 2002. Dedektör dengesizliği/enerji bağımlılığına bağlı halka artefaktları ve ince kesitle azalan kısmi hacim etkisi (çapraz doğrulama).
Note: This content is for education; for clinical decisions or regulatory compliance, consult a qualified medical physicist and current regulations.

← All articles