Artefacts Flashcards
Give an example of the source of the PARTIAL VOLUME artefact, and provide a remedy for this and any possible trade-offs
-SOURCE-
- Choice of slice thickness can directly affect the amount of detail seen in the image.
- With THICK SLICES, pixel intensity is averaged over multiple tissues, and structures smaller than the slice thickness may not be visible
-REMEDY/TRADE-OFF-
- Slice thickness can be reduced, allowing both large and small structures to be seen with correct contrast
- Reducing slice thickness will reduce the SNR as there are fewer spins in the imaged volume
Give an example of the source of the GEOMETRIC DISTORTION/IMAGE WARPING artefact, and provide a remedy for this and any possible trade-offs
-SOURCE-
- E.g., trapped air bubble in DWI phantom. This resulted in an air-tissue interface of materials with DIFFERENT MAGNETIC SUSCEPTIBILITY
- These local field inhomogeneities result in spatial displacements and signal drop-out in PE direction
-REMEDY/TRADE-OFF-
- Can be compensated for by using:
1. Shorter TEs
- less time for dephasing, but reduce
contrast
2. Using FSE sequences over GE sequences
- Less susceptible to B inhomogeneities
but require longer scan times and
increased SAR
3. Increasing gradient strength
- Restricts FOV for same BW
4. Avoiding narrow BWs
- Reduce SNR
Give an example of the source of the GIBBS RINGING artefact, and provide a remedy for this and any possible trade-offs
-SOURCE-
- Truncation artefact due to non ideal sampling of the edges of k-space which result in ‘ripples’ before and after the edge of a high contrast interface, resulting in a ring down effect due to the truncation of the signal prior processing the FT
-REMEDY/TRADE-OFF-
- Smoothing filters can be used to reduce the effect of this artefact, but will result in a loss of overall detail in the final image
- Increasing the number of PE steps will allow for a better approximation of the sinc function when truncating for the FT, but this will increase the overall scan time
Give an example of the source of the WRAP AROUND artefact, and provide a remedy for this and any possible trade-offs
-SOURCE-
- Anatomy that extends out width the FOV in the phase encoding direction appears on the other side, overlapping with anatomy within the FOV. - - Frequencies out width the bandwidth of the FOV are miss-registered as frequencies within the bandwidth
-REMEDY/TRADE-OFF-
- Remedies for this artefact include;
1. Using surface coils which allow to only
cover the FOV and improves SNR, but
requires additional equipment and scanner
processing.
2. The FOV can be increased to include the
whole anatomy, but this requires the need to
increase the matrix size to maintain the same
resolution which will increase the scan time.
3. Oversampling can be used, but this
increases the scan time.
4. Saturation pulses to saturate signals
outside of the FOV, but this increases the
total pulse train in the sequence and can
increase total SAR.
Give an example of the source of the BULK MOVEMENT/GHOSTING artefact, and provide a remedy for this and any possible trade-offs
-SOURCE-
- Ghosts are commonly associated with bulk movement and causes artefacts in the phase encoding direction. Ghosts superimpose over the normal anatomy
-REMEDY/TRADE-OFF-
Possible solutions are:
- To use pads and restraints to minimise
movement, but if the patient is claustrophobic
this may increase their anxiety - We could encourage the patient to stay still,
but this can prove difficult for young,
ill/physically impaired patients - Faster sequences or sequences with
accelerated acquisition strategies can be used
Give an example of the source of the SUSCEPTIBILITY artefact, and provide a remedy for this and any possible trade-offs
-SOURCE-
- MRI assumes underlying magnetic fields are
completely uniform and the gradient is
perfectly linear
- However, this is not the case near metallic
implants which cause distortions in the
frequency encoding direction
-REMEDY/TRADE-OFF-
- Artefacts can be minimised using SE
sequences with short TE (image) as this
minimises the signal loss due to T2*, but this
can affect the image contrast, or increase
scanning times, SAR
- Increasing receiver bandwidth can also
minimise this artefact but this will reduce SNR
Give an example of the source of the CHEMICAL SHIFT artefact, and provide a remedy for this and any possible trade-offs
-SOURCE-
- Differences between resonance frequencies of
fat and water results in mis-registration of fat
signals due to positional processing being
normalised to the resonance frequency of
water
- Since fat processes slower than water, the
chemical shift is shifted to lower frequencies
-REMEDY/TRADE-OFF-
- Can be minimised by using high bandwidth
sequences, or lower field strength scanners,
but this reduces total SNR
- Alternatively fat suppression (e.g. STIR) can be
used, but this can increase the SAR due to
addition of 180 pulse - Longer TE times can be used as this allows for
more dephasing of the fat signal, reducing its
signal, but this will affect the total signal as well
as affect the total contrast
Give an example of the source of the RESONANCE OFFSET artefact, and provide a remedy for this and any possible trade-offs
-SOURCE-
- The useable bandwidth of conventional steady
state free procession (SSFP) is less than 1/TR
and so signal nulls are experienced every 1/TR
in resonance frequency, producing well known
banding artefacts in the images as the SSFP is
very sensitive to inhomogeneities in the field,
resulting in variations in the S-S magnetisation
- Can arise from poor shimming
-REMEDY/TRADE-OFF-
- Can be improved by improving the
shimming/re-shimming, and ensuring the area
being imaged is at isocentre where the main
field is most homogenous
- The use of phase-cycling, as this allows for
multiple steady states to be reached, but can
increase the scan time by 2x - Shortening the TR will also reduce the banding,
but will reduce the maximum readout duration
and therefore will reduce the spatial
resolution & SNR, preventing use of EPI
Give an example of the source of the NYQUIST GHOSTING artefact, and provide a remedy for this and any possible trade-offs
-SOURCE-
- EPI sequences involve a train of echoes, where
each echo gathers data from one line of k-
space and every adjacent line is acquired in an
alternate direction. If forward and backward
echoes are not perfect mirrors of each other,
then artefacts are introduced in the image
processing where signal intensity is misplaced
in the phase-encoding direction halfway across
the image
-REMEDY/TRADE-OFF-
- Nyquist ghosts can result from poor shimming,
gradient coil heating, patient motion and
reconstruction errors. The most common is
eddy current generation in response to rapidly
changing gradient pulses
* can be reduced by reducing the echo train
length, but will reduce resolution in the PE
direction, but alternatively PI acceleration
can improve this but requires more
advanced equipment and software
