Recognizing and Minimizing Artifacts at CT, MRI, US, and Molecular Imaging - PowerPoint Presentation

Slide1

Recognizing and Minimizing Artifacts at CT, MRI, US, and Molecular Imaging

Benjamin L. Triche, MD

•

John T. Nelson Jr, MD

•

Noah S. McGill, MD

•

Kristin K. Porter, MD, PhD

•

Rupan Sanyal, MD

•

Franklin N. Tessler, MD

•

Jonathan E. McConathy, MD , PhD

•

David M. Gauntt, PhD

•

Michael V. Yester, PhD

•

Satinder P. Singh, MD

An earlier incorrect version of this presentation appeared online. This presentation

was corrected on March 9, 2022.

Slide2

Authors’ Affiliation:

Department of Radiology

University of Alabama at Birmingham619 19th Street SouthBirmingham, AL 35233Address correspondence to:BLT (e-mail: btriche1@tulane.edu)Current address: Department of Radiology,Tulane University School of Medicine,1430 Tulane Avenue, 8654 New Orleans, LA 70112.Presented as an Educational Poster at the 2017 RSNA Annual Meeting Acknowledgment: The authors thank Patricia D. Moore for her assistance in the design of the RSNA poster.

Disclosures of Conflicts of Interest.—K.K.P.

Activities related to the present article:

disclosed no relevant relationships.

Activities not related to the present article:

stockholder in Pfizer.

Other activities:

disclosed no relevant relationships.

F.N.T.

Activities related to the present article:

disclosed no relevant relationships.

Activities not related to the present article:

legal consultation fees from Lamb, Goldman, and Cullaro, PA and a speaker’s fee from the American College

of Radiology,

North

Carolina chapter

.

Other activities:

disclosed no relevant relationships.

J.E.M.

Activities related to the present article:

disclosed no relevant relationships.

Activities not related to the present article:

research consultant for Eli Lily, Blue Earth Diagnostics Ltd, Siemens, and General Electric.

Other activities:

disclosed no relevant relationships.

D.M.G.

Activities related to the present article:

disclosed no relevant relationships.

Activities not related to the present article:

patent agreement with Radcal Corporation.

Other activities:

disclosed no relevant relationships.

Slide3

Recognize common artifacts at CT, MRI, US, and molecular imaging.

Review the physical principles that result in these artifacts and their implications for image interpretation.

Explain how to resolve or accentuate artifacts to improve diagnostic quality and accuracy.Evaluate this material in a quiz or review format.

Slide4

Introduction

In this presentation, we illustrate artifacts that are commonly encountered in

nonradiographic modalities—CT, MRI, US, and molecular imaging and therapeutics—that are widely used across a broad range of practices. We describe how artifacts are generated, how to recognize them, and how to mitigate them as images are acquired.

Slide5

CT

(

slides 6-16) Beam Hardening Related to MetalBeam Hardening in the Posterior Fossa Volume AveragingPhoton StarvationRespiratory MotionCardiac MotionTruncationStair StepRing

BloomingWindmill

Motion

Pulsation

Zipper

Susceptibility

Truncation/Gibbs

Chemical Shift (First Kind)Chemical Shift (Second Kind)Bounce PointAliasingMagic AngleDielectric EffectMoiré FringesReverberationComet Tail and Ring DownMirror ImageAcoustic ShadowingIncreased Through TransmissionIncorrect Dynamic Range (DR)Partial Volume AveragingTwinklingAliasingSpectral MirroringFaulty Angle CorrectionStar PatternOff PeakRing/NonuniformityNonlinearityPhotomultiplier Tubes (PMT) MalfunctionDamaged Scintillation CrystalMetallic Objects and Contrast MaterialTruncationIncorrect Data EntryContaminationPatient Preparation

MRI(slides 17-28)

US(slides 29-39)

Molecular Imaging and Therapeutics

(

slides 40-50

)

Slide6

Appearance: Dark and bright streaks radiating from and between high-density objects, such as dental amalgam (arrow), as depicted on the CT axial image.

Cause

: The filtered back-projection algorithm assumes that all projections have the same beam hardness. Lower-energy photons are absorbed and/or attenuated in high-density tissue, leaving higher-energy photons to continue on, hardening the beam. A similar artifact (nonlinear partial volume artifact) is caused when high-density objects cover part of a single-detector channel.Resolution: Increase the voltage to better penetrate dense objects. Use a dual-energy acquisition technique. Turn on the metal artifact reconstruction algorithm (eg, iterative reconstruction with beam-hardening correction). Acquire data using narrow detector channels (eg, 0.625 mm) to reduce nonlinear partial volume averaging.Back to HomeCTBeam-hardening Artifact due to MetalReferences 1, 2.

Slide7

Appearance: Alternating dark and bright streaks in a narrow band extending across the posterior fossa (arrow) are depicted on the axial CT image. The position of the band moves vertically from one plane to another.

Cause

: Beam-hardening streaks from the dense skull base follow the acquisition plane. When using a nontilted gantry, these beams can pass through the posterior fossa. When the reconstructed images are tilted, the streaks appear in multiple image planes.Resolution: Tilt the gantry to move the artifact from the posterior fossa to the neck. Tilt the patient's head forward. Adjust the parameters as one would to reduce beam hardening due to metal.Back to HomeCTBeam-hardening Artifact in the Posterior FossaReferences 1, 2.

Slide8

Appearance: The left upper lobe segmental arterial branch appears to have a low-attenuating filling defect (arrow) on the CT image obtained with 3-mm section thickness (

a

), which resolves on the CT image obtained with 1.5-mm section thickness (b).Cause: The wide array of tissues occupying a voxel leads to the average attenuation of all the materials contained in the voxel being expressed in a single pixel. This is most pronounced at differing tissue density interfaces. Resolution: Acquire data with narrow detector channels (eg, 0.625 mm) to reduce nonlinear partial volume averaging. Reconstruct thinner images (eg, 1.25 mm instead of 2.5 mm).Back to HomeCTVolume Averaging Artifact3 mm1.5 mm

a

b

References 2, 3.

Slide9

Appearance: Sagittal CT image (a

) of the cervical spine shows significant noise and decreased contrast discrimination (oval) through the lower cervical and upper thoracic vertebral bodies. A similar appearance is depicted on the axial CT image (

b) of the chest, secondary to lateral body soft-tissue thickness.Cause: This artifact is secondary to photons traversing a dense structure, limiting the number of photons that strike the detector, which results in increased noise.Resolution: (1) In automatic exposure control mode, increase the milliampere-second (mAs) setting or reduce the noise index setting. (2) Increase the maximum mAs setting (eg, increase the maximum tube current, increase rotation time, or reduce the pitch). (3) In manual milliampere (mA) mode, increase technique affect (effective mAs) by increasing the tube current or rotation time or reducing pitch. (4) If there is variation among the thickness of the body tissues being imaged (eg, as in a thorax to neck examination), turn on the longitudinal tube current modulation. (5) Use an iterative reconstruction algorithm.Back to HomeCT

Photon Starvation Artifact

a

b

Reference 2.

Slide10

Appearance:

An apparent coronary artery filling defect and/or occlusion (arrow in a; oval in b) is depicted in these prospectively gated cardiac CT angiograms. The defect is related to respiratory motion, which is best appreciated on CT images obtained in the lung window (c).Cause: Distortion of anatomy by dark and/or bright bands and loss of image resolution secondary to respiratory motion between axial acquisitions causes this artifact.Resolution: Prevent voluntary movement with proper patient preparation and breath-hold coaching prior to the examination initiation.Acquire the images during a single breath-hold by reducing rotation time and increasing the detector coverage. Use an overscan acquisition mode (specific to axial acquisitions; rotation angle greater than 360°).Back to Home

CT

Respiratory Motion Artifact

References 2, 4.

a

b

c

Slide11

Appearance: Apparent coronary artery filling defect and/or occlusion (arrow in a

) resolved on the CT image (

b) after the vendor-provided intelligence correction was applied.Cause: Distortion of anatomy by dark and/or bright bands and loss of image resolution secondary to misregistration related to motion Resolution: (1) Use cardiac gating. (2) If retrospective gating is used, look at images obtained from other phases of the cardiac cycle. (3) Use overscan acquisition (specific to axial acquisitions; rotation angle greater than 360°). (4) Reduce rotation time and increase the detector coverage. (5) Use a vendor-provided intelligence correction. Back to HomeCTCardiac Motion ArtifactReferences 4, 5.a

b

Slide12

Appearance: Axial CT image through the abdomen shows increased curvilinear attenuation (arrows) along the edges of the image.

Cause:

Occurs when the anatomy of interest is outside the scan field of view (FOV) (eg, a large patient, or patient’s arms resting at their side), which results in the disruption of the accurate attenuation measurement along image edges. The selection of the display field of view does not affect this artifact.Resolution: Use proper patient positioning, if possible (eg, arms placed above head at imaging). Use an extended FOV reconstruction, if available.Back to HomeCTTruncation ArtifactReferences 2, 6.

Slide13

Appearance: (a) Jagged margins along high-attenuating borders of the right axillary vessels (oval) and right hemidiaphragm (arrow) on a reconstructed coronal image obtained with 5-mm section thickness axial acquisitions.

(b)

This resolved on image b, which was reconstructed from 2-mm section thickness axial acquisitions.Cause: Arises when axial images with a large image interval (thick nonoverlapping images) are used to create coronal or sagittal reformatted images.Resolution: Generate reformatted images from thin overlapped axial images. The interval of the axial images should be half of the axial image thickness, if possible. For example, use 1.25 × 0.625–mm axial images for generating soft-tissue reformatted images and 0.625 × 0.625–mm axial images for generating bone reformatted images. Back to HomeCTStair-step Artifact

Reference 7.

a

b

Slide14

Appearance:

Concentric circular structures expanding from the scanner isocenter (arrows) are depicted on axial CT images.

Cause: Related to miscalibrated detector (a), insufficient dose (b), or contrast material on the polyester film window that covers the detector (c). As the assembly rotates, the detector gives an incorrect value at each location, causing the ring artifact. Resolution: (1) Recalibrate the faulty detector. (2) Perform routine preventative maintenance on the detector. (3) Acquire images with sufficient radiation doses. (4) Evaluate for the presence of contrast material on the detector window cover.Back to HomeCTRing Artifact

References 1, 2, 5

.

a

b

c

Slide15

Appearance: Small highly dense structures such as calcifications and stents appear larger than they truly are, as depicted on the accompanying CT images. The circumferential coronary artery calcification (arrow) appears progressively smaller from left to right, with the changing window width (W) and level (L).

Cause

: Very high CT numbers of the structure cause pixel saturation when using typical lookup table (LUT) windows, causing the structure to appear larger than it is. Also, using a smoothing filter kernel makes small bright objects appear larger.Resolution: Use a LUT window that is wide enough so that the displayed pixels are not saturated. Use a standard or sharper filter kernel with an iterative noise reduction algorithm. Reduce the display FOV to improve the spatial resolution. Use a high kiloelectron-volt (keV) monoenergetic reconstruction (dual-energy acquisitions only). Use small section thicknesses. W/L: 800/350 HUW/L: 1000/500 HU

W/L: 1650/750 HU

Back to Home

CT

Blooming Artifact

Reference 5.

Slide16

Appearance: Regular equidistant streaks (arrow) radiating from high-density foci scattered throughout the soft tissues are depicted on the axial CT image. The streak pattern rotates and continues from one image to another.

Cause

: This is caused by inadequate sampling in the z direction (aliasing). This artifact occurs when multiple rows of detectors intersect the plane of reconstruction during a rotation. The number of pairs in the windmill correspond to the number of detectors intersecting the reconstructed image plane.Resolution: Decrease the pitch. Use axial acquisition.Back to HomeCTWindmill ArtifactReference 2.

Slide17

Appearance:

Multiple image ghosts (arrow) are depicted on the axial MR image, with blurring in the phase-encoding direction.Cause: Signal-producing voxels change position between the radiofrequency (RF) pulses. Random motion produces smearing in the phase-encoding direction, while periodic motion produces discrete ghosts in the phase-encoding direction. Resolution: Use pulse sequences that shorten acquisition time. Educate the patient about the imaging procedure or consider patient sedation. Back to HomeMRIMotion ArtifactReferences 8–10.

Slide18

Appearance:  On the axial MR image, a round isointense structure (arrow) is depicted in the lateral hepatic segment in the phase-encoding direction, representing a pulsation artifact from the abdominal aorta. 

Cause:

This is a form of motion artifact in which the signal-producing voxels change position during the application of the RF pulse. This always occurs in the phase-encoding direction.Resolution: Use pulse sequences with shorter acquisition times. Change the phase-encoding direction. Use spatial presaturation bands. Increase the number of signal averages.Back to HomePulsation ArtifactMRIReferences 8–10.

Slide19

Appearance: Alternating dark and bright bands (arrows) are seen extending across the length of the MR image.

Cause

: An extrinsic RF signal creates this artifact. It occurs in the phase-encoding direction. Resolution: Examine the Faraday cage, especially the door seal. Turn off any unnecessary electronics in the examination room.  Back to HomeZipper ArtifactReference 8.MRI

Slide20

Appearance: Axial MR image shows markedly low signal intensity in the pineal region, secondary to susceptibility artifact related to blood products within a highly vascular tumor (arrow). 

Cause

: Materials with markedly different susceptibility values generate local field inhomogeneities, resulting in faster spin dephasing and thus signal loss. The artifact is also referred to as blooming artifact.Benefit: Can be used beneficially to detect materials that differ significantly in susceptibility values compared with those of the remainder of the field, such as the blood products within the lesion depicted on the axial MR image.Resolution: Use spin-echo imaging.Reduce echo time and increase the acquisition matrix. Back to HomeSusceptibility ArtifactReferences 9, 10.MRI

Slide21

Appearance:

Sagittal MR image shows a linear hyperintense signal (arrow) depicted at the cerebrospinal fluid–spinal cord interface, mimicking a syrinx in the cervical spinal cord.Cause: The Fourier transform analysis corresponds to an infinite number of frequencies at high-contrast boundaries, but as sampling is finite, a series of lines of alternating high and low signal intensities will form at high-contrast tissue interfaces. This artifact is seen more frequently in the phase-encoding direction because the number of phase-encoding steps is usually smaller (and thus the high-contrast interface is undersampled) to reduce scan time. Resolution: Increase the matrix size. Use imaging with fat suppression if fat is a boundary.Apply smoothing filters.Change the phase-encoding direction. Back to HomeTruncation/Gibbs ArtifactReferences 9, 10.MRI

Slide22

Appearance:

Axial MR image obtained at the level of the kidneys shows dark bands (arrows) with low signal intensity at the right medial and left lateral renal borders, with a concurrent high-signal-intensity bright band with high signal intensity (arrowheads) at the right lateral and left medial renal borders. Cause: Differences in the precessional frequencies of protons in water and fat result in spatial misregistration of the MR signal at fat–soft-tissue interfaces. This spatial MR signal misregistration results in a dark band on one side, and a bright band on the other side of the fat/tissue boundary. This occurs in the frequency-encoding direction, which is right to left in the accompanying image. The signal from protons in fat is recorded as arising from water protons from a different, lower frequency location. The bright band will appear in the higher frequencies, while the dark band will appear in the lower frequencies.Resolution: Increase the receiver bandwidth. Decrease the FOV. Use imaging with fat saturation. 

Back to Home

Chemical Shift of the First Kind

References 8–11.

MRI

Slide23

Appearance:

Distinct dark band ("India ink" artifact, arrow) is depicted along the fat–soft-tissue interfaces on this out-of-phase gradient-echo MR image.Cause: This artifact is depicted on out-of-phase gradient-echo images in which signals from fat and water are out-of-phase and partially cancel each other, reducing signal intensity, and thus creating a peripheral black line at the fat–soft-tissue interface. This artifact occurs independent from the frequency and phase directions of the RF pulse. Benefit: This can help in the detection of intravoxel and/or microscopic fat.Resolution: Adjust the echo time. Use spin-echo imaging.Back to HomeChemical Shift of the Second KindReferences 8–11.MRI

Slide24

Appearance:

Null signal (India ink artifact) (arrow) is depicted at tissue interfaces containing both fluid and soft tissue in the same voxel on this axial inversion-recovery MR image. Note that motion artifact is also seen through the nasal region.Cause: This occurs by phase cancellation of the MR signal in boundary pixels with inversion-recovery sequences when two tissues of different T1 values recover to the same magnitude but opposite signs. Resolution: Change the inversion time or repetition time.Phase-sensitive reconstruction inversion-recovery techniques.Back to HomeBounce Point ArtifactReference 8.MRI

Slide25

Appearance:

The soft tissues of the jaw were not included within the sampling FOV on this axial MR image. The excluded soft tissues are displayed on the posterior aspect of the image, which have "wrapped around" the image on the displayed FOV (arrow).Cause: This occurs when the acquisition FOV is smaller than the object being imaged and always occurs in the phase-encoding direction. Resolution: Increase sampling bandwidth or the FOV.Back to HomeAliasing at MRIReferences 9, 10.MRI

Slide26

Appearance

:

Apparent intermediate signal intensity (arrow in a) is depicted within the supraspinatous tendon on the intermediate-weighted MR image (a), which resolves (arrow in b) on the T2-weighted MR image (b). Note that motion artifact is also seen on b.Cause: Lengthening of T2 times in tightly bound collagen molecules occurs when molecules lie at 54.74 degrees from the main magnetic field on sequences with a short echo time (TE), resulting in high signal intensity.Resolution: Use longer TE sequences versus shorter TE sequences. Compare images with shorter TE values to those with longer TE values. Back to HomeMagic Angle Artifact

Reference 8, 12.

a

b

MRI

Slide27

Appearance: Dark shaded ovoid area with peripheral bright areas is depicted on the center of these T1- and T2-weighted MR images (arrows) in a patient with cirrhosis and ascites.

Cause:

As the magnetic field strength increases from 1.5 T to 3 T, the RF wavelength decreases from 52 cm to 26 cm, which approximates the diameter of patients with large body habitus or ascites. This leads to interference patterns from opposing wavelengths entering from opposite sides of the patient, with subsequent darkening with peripheral brightness in the center of the image. Resolution: Perform imaging with 1.5 T instead of 3 T or use dielectric pads. Back to HomeDielectric EffectReference 10.MRI

Slide28

Appearance:

Alternating curvilinear bright and dark bands (arrows) are depicted throughout the accompanying balanced steady-state free-precession gradient-echo MR image. Cause: Interference patterns, predominantly depicted with gradient sequences, may be generated from the inhomogeneity of the main magnetic field, a defect in shielding, aliasing, or differing excitation modes (ie, echo times). This results in the superimposition of signals of different phases, which add and cancel each other.Resolution: Decrease the magnetic field inhomogeneity by using a shimming process. Use spin-echo sequences.Use a surface coil. Back to HomeMoiré FringesReference 8.

MRI

Slide29

Appearance: (a) US image shows low-medium level echoes (arrow) immediately deep relative to the abdominal wall in a patient with ascites.

(b)

Gray-scale US image shows multiple equidistant linear echoes posterior to the biopsy needle (arrow), secondary to multipath reflections between the anterior and posterior needle walls. Cause: Ultrasound pulses that undergo additional reflections on the way back to the transducer take longer to return than those that follow an unimpeded path. They are interpreted by the US system as originating from a deeper location than they actually are. In a, the pulses bounce back and forth between the fluid–body wall interface and overlying structures. In b, the pulses reflect between the needle walls one or more times. No echoes are depicted deep relative to the needle bevel (arrowhead in b), as there is effectively only one wall.Resolution: Lower the gain setting. Change the angle of insonation. Use harmonic imaging.Back to HomeUS

Reverberation Artifact

Reference 13.

a

b

Slide30

Appearance: US images show a thyroid nodule with colloid (arrow in a

) and a scrotum with soft-tissue gas (arrow in

b).Cause: Multipath reflections within small structures:Comet tail: Most commonly depicted in cases of colloid thyroid nodules; V-shaped echo train deep relative to echogenic fociRing down: Multiple reflections within fluid-filled tetrahedral spaces between air bubbles; often change dynamically at real-time imagingBenefit: These are helpful to identify colloid and gas.Resolution: Adjust the insonation angle.Back to HomeUSComet Tail and Ring-down Artifacts

Reference 13.

b

a

Slide31

Appearance:

Color Doppler US image shows a patent transjugular intrahepatic portosystemic shunt (TIPS) in the liver and a duplicate (arrow) above the diaphragm. Cause: The ultrasound beam is reflected by a highly reflective surface (eg, the diaphragm) and encounters another structure (eg, the TIPS) on its return to the transducer, reflects back to the diaphragm, and ultimately back to the transducer. This can be recognized by the duplication of a structure in an atypical location.Resolution: Manipulate the angle of insonation.Back to HomeUSMirror-Image ArtifactReference 13.

Slide32

Appearance: US images show shadowing artifacts (arrows) from gallstones and medullary

nephrocalcinosis

.Cause: Structure reflects the most sound back to the transducer, leaving a much lower amplitude to interrogate deeper tissues. This is depicted as dark regions deep relative to the highly reflective structures.Benefit: This helps diagnose gallstones and similar abnormalities. Resolution: Resolve by scanning from another vantage point. However, eliminating the artifact may make it more difficult to characterize the structure that caused it.Back to HomeUSAcoustic Shadowing Artifact

Reference 13.

a

b

Slide33

Appearance: Gray-scale US image obtained through the right upper quadrant shows increased echogenicity (arrow) deep relative to the gallbladder.

Cause:

Fluid-filled structure attenuates the ultrasound beam less than the adjacent tissues, resulting in stronger echoes deep relative to the beam.Benefit: Useful to identify fluid or characterize homogeneously hypoechoic solid lesionsResolution: Decrease the gain setting. Manipulate the focal zone if the area of artifact is too bright to see detail.Back to HomeUSIncreased Through-TransmissionReference 13.

Slide34

Appearance: US image (a

) (DR = 40 dB) shows an apparent patent vessel. US image (

b) (DR = 68 dB) shows a nonocclusive thrombus (arrow).Cause: DR is the ratio of the strongest to weakest echoes processed and displayed. Images with a low DR have high contrast, which may hide a thrombus or other low-level echoes.Resolution: Set the DR to an intermediate level that does not obscure low-level echoes.Back to HomeUSIncorrect Dynamic Range (DR)a

b

Slide35

Appearance: Artifactual echoes within an anechoic structure, such as the carotid artery (arrow) as seen in this US image.

Cause:

The ultrasound beam, which has finite thickness, partly interrogates adjacent tissues.Resolution: Be aware of the location of the scanning plane relative to the margin of the vessel. Confirm this by scanning in an orthogonal plane.Back to HomeUSPartial–Volume Averaging Artifact

Slide36

Appearance: US image shows shadowing calculus in the distal ureter (arrow in a

), with mixed signal (arrow in

b) on the color Doppler US image (b).Cause: This may be caused by broadening of bandwidth of inherent machine noise interacting with a rough surface.Benefit: This helps increase the specificity of urolithiasis detection. Resolution: The artifact varies by adjusting the scanner settings, including the focal zone and insonation angle. Back to HomeUSTwinkling ArtifactReference 14.

a

b

Slide37

Appearance: In this US image with a spectral Doppler US waveform, the internal carotid artery “wraps around” (arrow) to the opposite side of the velocity (ie, frequency) display.

Cause:

The sampling rate is less than the Nyquist limit (one-half the pulse repetition frequency), predominately a phenomenon of spectral Doppler US. A similar phenomenon occurs in the accompanying color Doppler US image. Flow that exceeds the velocity limit on the color scale bar wraps around to the color bar in the opposite direction.Resolution: Use a lower-frequency transducer.Increase the Doppler angle. Sample at a shallower depth. Increase the Doppler scale. Lower the Doppler baseline.Back to HomeUSAliasing at USReference 15.

Slide38

Appearance: Duplication (arrows) of the Doppler US waveform is depicted above and below the baseline on the accompanying US image and waveform.

Cause:

This can be caused by using an insonation angle close to 90 degrees and/or a gain setting that is too high. This can be related to pulsation thrill.Resolution: Decrease the angle of insonation. Reduce the Doppler gain setting.Back to HomeUSSpectral Mirroring ArtifactReference 15.

Slide39

Appearance: (a)

US Doppler image and waveform of the common carotid artery shows that the Doppler angle correction cursor (arrow) is incorrectly set. (b) US Doppler image and waveform show the appropriate angle correction cursor (arrow), which demonstrates higher velocity.Cause: The machine uses an angle correction set by the operator to calculate velocity. If incorrectly set, the velocity measurement may be incorrect.Resolution: Set the angle correction parallel to the flow stream, as depicted on b.Back to HomeUSFaulty Angle Correction

Reference 15.

a

b

Slide40

Appearance: Iodine 131 (131I

) whole-body scans obtained post-treatment show a six-pointed star pattern (arrow) over the lower neck and upper thorax, indicating residual thyroid tissue.

Cause: This is related to high radiotracer concentration and/or activity and septal penetration of hexagonal collimator holes. Resolution: Use a high-energy collimator. However, due to other high-energy photons in 131I decay, star pattern artifacts may still occur.Back to HomeMolecular Imaging andTherapeutics Star Pattern ArtifactReference 16.

Slide41

Appearance: Thallium 201 (

201

Th) scintigraphic images obtained with the same camera show decreased resolution, increased scatter, and poor contrast discrimination between off-peak images (a) and corrected images (b).Cause: Incorrect energy peak and/or peak drift; typically results in low-resolution images owing to increased acceptance of Compton scattering (eg, lower-energy photons being imaged). If cobalt 57 (Co-57) flood source is used for daily uniformity, peaking at 122 keV may not produce the correct peak at 140 keV for technetium-99m (99mTc).Resolution: Repeat the study after properly setting the energy peak for the radionuclide that is being used at imaging. Perform energy peaking daily.Back to HomeOff-Peak ArtifactReferences 17, 18.a

b

Molecular Imaging and

Therapeutics

Slide42

Appearance: SPECT images show concentric rings (arrow) projecting over the liver.

Cause:

This is caused by variability in detector response and/or voltage assignment at various points along the scintillation crystal owing to either ineffective light transfer or poorly balanced PMTs or an incorrect pulse height analyzer setting. A nonuniformity of 5% is typically allowed at planar imaging and 1%–3% at SPECT. It is similar in appearance to that of the CT ring artifact.Resolution: Perform flood testing daily. If the correction cannot be made within 5%, the manufacturer needs to be notified. Back to HomeRing/Nonuniformity ArtifactReference 18.Molecular Imaging andTherapeutics

Slide43

Appearance: Image of a bar phantom shows subtle curves along the bars in the lower left quadrant (rectangle in a

), which resolves after a new linearity map is acquired (

b).Cause: This is related to PMT balance relative to that of the neighboring tubes and/or a linearity correction issue. It produces bowing and distortion of straight lines and margins.Resolution: Service reacquires linearity map. Monitor with spatial resolution and/or linearity testing weekly with a bar phantom.Back to HomeNonlinearity ArtifactReference 18.ab

Molecular Imaging and

Therapeutics

Slide44

Appearance: Quality control image (a) shows a large circular defect (arrow) from PMT decoupling. Off-peak image (

b

) with circumferential distortion (arrow) around the decoupled PMT shows the influence of a single PMT.Cause: PMTs convert light produced by the scintillation camera into proportionate measureable electrical signals. Damaged or improperly tuned PMTs could produce cold or hot areas within the imaging field. Resolution: Perform flood testing daily. If unable to correct to within 5%, manufacturer needs to be notified. Back to HomePMT MalfunctionReference 17.a

b

Molecular Imaging and

Therapeutics

Slide45

Appearance: Large irregular photopenic area on the accompanying quality control images show

a cracked crystal defect (arrows).

Cause: Focal distortion of the image is related to a damaged or cracked scintillation crystal, the result of either mechanical or thermal shock.Resolution: Replace the scintillation crystal.Back to HomeDamaged ScintillationCrystalReference 17.Molecular Imaging andTherapeutics

Slide46

Appearance: Falsely elevated standardized uptake value (SUV) surrounding the right hip (arrow in

b

) is depicted on the CT attenuation-corrected PET image (b), related to the presence of a right hip prosthesis. However, the CT attenuation-noncorrected image shows a focal increase of hypermetabolic activity, suggesting infection and/or inflammation (arrowhead on d). Cause: High-attenuating objects at CT result in overcorrection on CT attenuation-corrected PET images, leading to falsely interpreted elevated SUVs.Resolution: Evaluate the CT and non-attenuation–corrected PET images and correlate that the findings are in the same region to determine if there is overcorrection from high-attenuating objects.Back to HomeMetallic Objects andContrast MaterialReference 19.

a

b

c

d

Coronal CT image (

a

), CT attenuation-corrected PET image (b), coronal PET/CT image (c), and non-attenuation–corrected PET image (d) show increased activity around the right hip that is a mixture of artifact and increased fluorodeoxyglucose (FDG) uptake. Molecular Imaging andTherapeutics

Slide47

Appearance: Axial FDG PET (

a

) and CT (b) images show falsely decreased activity of the lateral flanks, with apparent increased curvilinear activity (arrows) along the medial flanks secondary to truncation at CT. As depicted in the axial PET (c) and CT (d) images obtained following a 70-cm FOV reconstruction at CT, the falsely increased and decreased areas of activity resolved. Cause: The standard FOV for PET and CT is 50 cm. If portions of the patient’s body are outside of the CT FOV, this causes high attenuation at the margins of the CT images, leading to PET activity overcorrection in these regions and activity underestimation outside of these regions. Resolution: Use the truncation correction techniques available on PET/CT systems. Reconstruct PET and CT images to 70 cm using interpolation. Back to HomeTruncation Artifact

Reference 19.

a

b

d

c

Molecular Imaging and

Therapeutics

Slide48

Appearance: Abnormal mean SUV of the liver related to a patient dose recording error.

Whole-body FDG PET

images that were similarly scaled show an incorrect mean SUV for the liver in a (20 mCi) and a correct mean SUV in b after the correct dose was entered (9 mCi).Cause: This is caused by under- or overestimation of object SUV related to incorrect patient and/or study data, including patient body mass, tissue radioactivity at initial time point, or injected dose activity. A quick way to assess this is to look at the SUV in the liver, which should be about 2.0–3.5.Resolution: Ensure proper data entry by the technologist.Back to HomeIncorrect Data EntryabMolecular Imaging andTherapeutics

Slide49

Appearance:

Whole-body FDG maximum intensity projection (MIP) image (a) shows multiple superficial hypermetabolic foci, a finding consistent with external contamination (ovals). Whole-body FDG MIP image obtained during reimaging (b) through the chest and abdomen shows significant improvement after cleaning and changing clothes (right).Cause: This artifact is caused by radiotracer contamination on skin or clothing.Resolution: Repeat the study after cleaning the skin or removing the contaminated clothes.Back to HomeContamination Artifactab

Molecular Imaging andTherapeutics

Slide50

Appearance:

Diffuse skeletal and cardiac muscular uptake of FDG on an MIP image (a) in a patient who consumed coffee with sugar prior to undergoing imaging. Repeat MIP image (b) obtained after the patient followed specific dietary restrictions shows typical FDG distribution.Cause: Insulin stimulation from either the lack of fasting or administration of insulin prior to injection of FDG leads to an increase in glucose transporter type 4 (GLUT4) receptors and to glucose transport into skeletal and cardiac muscle.Resolution: Ask all patients about caloric intake and insulin use before FDG injection. Repeat the study with specific dietary instructions for the patient.Back to HomeArtifacts fromIncorrect Patient Preparationab

Molecular Imaging and

Therapeutics

Slide51

References

Boas F, Fleischmann D. CT artifacts: causes and reduction techniques. Imaging Med 2012;4(2):229–240.

Barrett JF, Keat N. Artifacts in CT: recognition and avoidance. RadioGraphics 2004;24(6):1679–1691.Al-Shakhrah I, Al-Obaidi T. Common artifacts in computerized tomography: a review. Appl Radiol 2003;32(8):25–30.Ginat DT, Gupta R. Computed tomography. In: Peh WCG, ed. Pitfalls in diagnostic radiology. Heidelberg, Germany: Springer, 2015; 23–38.Kalisz K, Buethe J, Saboo SS, Abbara S, Halliburton S, Rajiah P. Artifacts at cardiac CT: physics and solutions. RadioGraphics 2016;36(7):2064–2083.Mawlawi O, Erasmus JJ, Pan T, et al. Truncation artifact on PET/CT: impact on measurements of activity concentration and assessment of a correction algorithm. AJR Am J Roentgenol 2006;186(5):1458–1467.Wang G, Vannier MW. Stair-step artifacts in three-dimensional helical CT: an experimental study. Radiology 1994;191(1):79–83.Zhuo J, Gullapalli RP. AAPM/RSNA physics tutorial for residents: MR artifacts, safety, and quality control. RadioGraphics 2006;26(1):275–297.Arena L, Morehouse HT, Safir J. MR imaging artifacts that simulate disease: how to recognize and eliminate them. RadioGraphics 1995;15(6):1373–1394.Morelli JN, Runge VM, Ai F, et al. An image-based approach to understanding the physics of MR artifacts. RadioGraphics 2011;31(3):849–866. Hood MN, Ho VB, Smirniotopoulos JG, Szumowski J. Chemical shift: the artifact and clinical tool revisited. RadioGraphics 1999;19(2):357–371.

Bydder M, Rahal A, Fullerton G, Bydder G. The magic angle effect: a source of artifact, determinant of image contrast, and technique for imaging. J Magn Reson Imaging. 2007;25:290–300.

Feldman MK, Katyal S, Blackwood MS. US artifacts. RadioGraphics 2009;29(4):1179–1189.

Dillman JR, Kappil M, Weadock WJ, et al. Sonographic twinkling artifact for renal calculus detection: correlation with CT. Radiology 2011;259(3):911–916.

Pozniak MA, Zagzebski JA, Scanlan KA. Spectral and color Doppler artifacts. RadioGraphics 1992;12(1):35–44.

Bruyant PP, Sau J, Mallet JJ. Streak artifact reduction in filtered backprojection using a level line-based interpolation method. J Nucl Med 2000;41(11):1913–1919.

Gentili A, Miron SD, Adler LP. Review of some common artifacts in nuclear medicine. Clin Nucl Med 1994;19(2):138–143.

White SL. Quality assurance testing of gamma camera and SPECT systems. Presented at the 51st American Association of Physicists in Medicine Annual Meeting (AAPM), AAPM Virtual Library, Anaheim, California, July 26–30, 2009.Pettinato C, Nanni C, Farsad M, et al. Artefacts of PET/CT images. Biomed Imaging Interv J 2006;2(4):e60.

Slide52

Al-Shakhrah I, Al-Obaidi T. Common artifacts in computerized tomography: a review. Appl Radiol 2003;32(8):25–30. Barrett JF, Keat N. Artifacts in CT: recognition and avoidance. RadioGraphics 2004;24(6):1679–169.

Gentili A, Miron SD, Adler LP. Review of some common artifacts in nuclear medicine. Clin Nucl Med 1994;19(2):138–143.

Hearshen D, Ellis J, Carson P, Shreve P, Aisen A. Boundary effects from opposed magnetization artifact in IR images. Radiology 1986;160:543–547.Morelli JN, Runge VM, Ai F, et al. An image-based approach to understanding the physics of MR artifacts. RadioGraphics 2011;31(3):849–866. Pettinato C, Nanni C, Farsad M, et al. Artefacts of PET/CT images. Biomed Imaging Interv J 2006;2(4):e60.Pozniak MA, Zagzebski JA, Scanlan KA. Spectral and color Doppler artifacts. RadioGraphics 1992;12(1):35–44.