Transcript Technical Considerations in Brain DWI: A practical, image-based guide for neuroradiologists AZ Chow, MD1; JN Morelli, MD2; CM Gerdes, MD2; JD Cannell, MD2, M.

Technical Considerations in
Brain DWI:
A practical, image-based guide for neuroradiologists
AZ Chow, MD1; JN Morelli, MD2; CM Gerdes, MD2; JD Cannell, MD2,
M Saettele, MD2; VM Runge, MD1; FL Goerner, PhD1
1
UT-Medical Branch, Galveston, TX 2 Scott&White Memorial Hospital, Temple, TX
Purpose
• To provide imaged-guided instruction of technical
considerations of diffusion-weighted imaging (DWI)
for neuroradiologists.
• To give the physical basis for clinical diffusionweighted imaging and recent technological
developments and practical ways to improve image
quality
Theoretical Basis
• Diffusion principle is based on Brownian motion of water protons.
• Gradients can be applied to measure diffusion with an initial dephasing
gradient applied.
• A rephasing gradient is then applied.
• Interval movement of the protons will cause signal loss proportional to the net
movement.
RF
ADC
Gdiff
Gf
Gp
Theoretical Basis:
Cytotoxic edema versus vasogenic edema
After an ischemic event, cytotoxic edema About 6 hours after an ischemic
can develop within 15-30 minutes and is
event, capillary leak starts to result in
high signal intensity on DWI.
vasogenic edema which has high
signal intensity on T2/FLAIR.
Increased permeability
H2O
Blood Vessel
Cell
Cell
Interstitium
Cell
Cell
Cell
Why DWI?
Ischemia
- chronic
white
matter
changes
Ischemia
- hyperacute
stroke
(<6
hrs)
Multiple
sclerosis
Infection
- subdural
empyema
confounding acute stroke
DWI
T1 post contrast FLAIR
DWI
T1
post-contrast
scan
shows
low signal
intensity
contrast
enhancement.
T2 shows
many
hyperintensities
in a MS
patient
FLAIR
detects vasogenic
edema
which
takes >6collection
hours
to with
develop
post-ischemic
event
FLAIR
DWI
Restricted diffusion suggests a subdural empyema.
FLAIR
shows several
ischemic
microvascular
changes
DWI cytotoxic
hyperintensities
confirm
the acutewithin
nature
of the
lesions
DWI can detect
edemaareas
whichofdevelops
15-30
minutes
of symptoms
DWI reveals an acute infarct
et MR
al. Essentials
of Clinical
MR.2009:129.
2011:43.
Runge et al.
Runge
TheetPhysics
al. Clinical
ofRunge
Clinical
3T Magnetic
Taught
Resonance.
through
Images.
2007:68.
Why DWI?
• Ischemia
• Infection – Abscess
• MS
• CNS lymphoma
• Subacute MTX toxicity
• Acute disseminated encephalomyelitis
Diffusion Encoding:
Bipolar vs Modified ST Schemes
Stejskal-Tanner
Bipolar
Modified
ST
Diffusion encoding
Bipolar
gradient (p2)
(Parallel imaging factor)
Minimum TE (msec)
Modified
ST (p2)
RF
ADC
Bipolar gradient (p2)
95
Modified Stejskal-Tanner (p2)
77
Gdiff
Modified Stejskal-Tanner (p3)
67
Modified Stejskal-Tanner (p4)
63
Gf
Bipolar
gradient
ST (p4)
(p3)
(p2)
Modified
(p2)
Gp
-Different diffusion sequences can be performed
with varying effects as seen in this left PCAModified ST (p3)
distribution stroke with susceptibility artifact.
-The traditional bipolar gradient diffusion is
prone
to bulk susceptibility
artifact (white
The differences
in susceptibility
arrows).
Black
arrow
points to infarct.
artifact are
circled
in red.
-Modified ST schemes allow shorter TE times
for identical b values, thus increasing SNR
-With increases in SNR, parallel imaging factors
can be increased, reducing susceptibility artifact.
Modified ST (p4)
Morelli et al. Invest Radiol. 2010;45(1):29-35
K-space Sampling Strategies:
Fast Spin Echo Techniques
Fast spin echo (FSE) - modification of spin echo sequence by adding additional
180O pulses within a TR period creating additional echoes and creating an echo
train. The total number of pulses = echo train length. Total scan time is
equivalent to conventional spin echo scan time divided by the echo train length.
Can be combined with other diffusion scans such as BLADE/PROPELLER
FSE EPI
BLADE
Advantages include reduced scan time
Penalty of reduction in number of slices acquired in a single scan
Can be combined with diffusion scan techniques like BLADE/PROPELLER
Attenberger et al. Invest Radiol. 2009:659.
K-space Sampling Strategies:
Single vs Multishot (readout-segmented) EPI
Single shot echo planar imaging (ss-EPI) rapidly fills the K space in a linear fashion
Multishot (readout-segmented, rs-EPI) fills the K space with several shots with the
center of the K space sampled each time
ss-EPI
rs-EPI
ss-EPI
Fast acquisition (e.g. 1.2 minutes)
Less susceptible to motion
More prone to artifact
rs-EPI
Slower acquisition time (e.g. 3 minutes)
More susceptible to motion
Reduction in blur & susceptibility artifact
Higher effective resolution
Morelli JN, Runge VM, Porter DA, et al, ARRS 2010
K-space Sampling Strategies:
PROPELLER/BLADE and Radial Trajectories
PROPELLER/BLADE is a unique type of multishot imaging technique that fills
the K space in a radial fashion with multiple rotating echo trains
ss-EPI
PROPELLER/BLADE
ss-EPI
PROPELLER/BLADE
Susceptibility artifact
High bulk susceptibility artifact Minimal bulk susceptibility artifact
Rapid scan time
Longer scan time
Good SNR
Lower SNR
Less motion artifact
The Diffusion Tensor:
B-values and SNR
The b-value is the summary of the diffusion weighting applied to an image
A b-value of 0 generally defaults
to a T2-weighted scan; b-value of 1000
Heavy diffusion
Essentially
a loss
T2scan. Increasing b-values
correlates to a heavily diffusion-weighted
weighting
with
with
high (SNR)
cause a decrease in signal
to noise
ratio
ofscan
signal
in the
intensity (areas
signal in
ventricles
of
thediffusion)
ventricles
free
b=0
b = 300
bb==900
600
300
0
b = 600
b = 900
Runge et al. The Physics of Clinical MRI Taught through Images. 2009: 130.
The Diffusion Tensor:
Trace-Weighted Images
The trace
image
is formed
by restrain
the
White
matter
axonal
tracts
movement of protons to certain
combination
of the individual
diffusion
directions. Gradients
parallel
to the tract will cause large signal loss
tensor
images (LR + CC
+ AP) will have minimal signal loss.
while perpendicular
gradients
The trace image can incorporate color
mapping to maintain directionality data
AP
CC
LR
The genu and splenium of the
corpus callosum show loss of
signal from free diffusion of
protons in the transverse plane
CC
AP
LR
trace
Runge et al. The Physics of Clinical MR Taught through Images. 2005: 131.
The Diffusion Tensor:
Trace-Weighted Images
Clinical applications:
• Surgical planning for tumors near the major critical tracts such as the
motor, sensory, and optic tracts (Neurotherapeutics. 2007)
• Traumatic brain injury/diffuse axonal injury (Xu et al. J Neurotrauma
2007)
• Multiple sclerosis evaluation in context of normal appearing white
matter (Testaverde et al. Eur Radiol. 2012.)
The Diffusion Tensor:
Construction of ADC Maps
ADC = apparent diffusion coefficient calculated from ADC = - ln(S/S0)/b
ADC maps are constructed for the purposes of obtaining an image with only
diffusion information
0.0
- 0.2
log(S/S0)
- 0.4
small ADC
- 0.6
- 0.8
- 1.0
large ADC
- 1.2
b-value
b=0
b=1000
ADC map
Courtesy of John E Kirsch, PhD with Siemens Medical
The Diffusion Tensor:
T2 Shine Through
The need for ADC maps arise because the baseline diffusion sequence with a
b-value = 0 is essentially a T2 weighted MRI, so high intensity signals can
“shine through” and create a high intensity signal artifact on DWI
DWI
ADC
T2
DWI shows large region of restricted diffusion in the right thalamus/deep white
matter. ADC map shows high intensity signal indicating T2 shine through
artifact.
The Diffusion Tensor:
Exponential Images
Exponential images are formed by dividing the signal intensity of DWI by the B0
scan that is pure echo-planar spin-echo T2 weighted image. The resulting
signal intensity is exponentially related to the ADC and is also used to rule out
T2 shine through artifact.
DWI
EPI, b=0
Exponential
Artifacts:
Susceptibility
Differences in magnetization properties of adjacent structures/tissues can cause
susceptibility gradients and can manifest as distortion/signal changes. Within the
brain, most common near parenchyma/sinus interfaces.
High signal intensity susceptibility
Resolution improved with
multishot
BLADE/PROPELLER
is relatively
Artifactual pontine stretching
artifact near the mastoid airAnd
cellsless artifactual resistant
distortionto susceptibility artifacts
ss-EPI
multishot-EPI
BLADE/PROPELLER
Artifacts:
Chemical Shift
Chemical shift artifact arises when significant fat and water content are adjacent
to tissue. In neuroradiology, it is more commonly seen near the scalp.
A rs-EPI B0 image without fat saturation. A
chemical shift artifact is seen on the scan in the
form of a band in the posterior brain (red
arrows).
A rs-EPI B0 image with fat
sat shows resolution of
the chemical shift artifact
Artifacts:
Eddy Currents
Nearby conductive surfaces create small magnetic fields (eddy currents) during
application of diffusion gradients resulting in distortion of the main field. The eddy
currents can be minimized with active shielding around the modern coils.
Surgical hardware-related
Conductive object artifact
Considerations at 3T
Increase in magnetic field strength (e.g 1.5T -> 3T) can yield
improved SNR by allowing shortening of the TE while keeping
the b value constant.
However, susceptibility artifacts are greater at 3T (red arrows).
1.5T
ss-EPI
3T ss-EPI
1.5T
rs-EPI
3T
rs-EPI
3T BLADE
1.5T
BLADE
Additional Considerations at 3T
•Magnetic field strength requires being able to apply it fast enough to
take advantage of the increase in field strength. However, faster ramping
can cause more peripheral nerve stimulation.
•Susceptibility artifacts are generally greater at 3T with all other
variables held constant.
•Parallel imaging generally causes a loss of SNR that prevented
widespread usage with 1.5T MRIs, but with increase in SNR for 3T field
strengths, parallel imaging can decrease number of pulses needed to
complete a sequence
•Scans becomes much more sensitive to eddy currents
Summary
• The clinical utility of DWI includes evaluation of
ischemia, infection/inflammation, neoplasia, and
multiple other conditions
• Multiple techniques are available to decrease artifacts
or enhance DWI including modified ST-schemes,
parallel imaging, multishot (readout-segmented) EPI,
BLADE/PROPELLER imaging
References
1.
Alexander AL, Lee JE, Lazar M, Field AS. Diffusion tensor imaging of the brain. Neurotherapeutics. 2007 Jul;4(3):316-29.
2.
Attenberger UI, Runge VM, Stemmer A, Williams KD, Naul LG, Michaely HJ, Schoenberg SO, Reiser MF, Wintersperger BJ. Diffusion weighted
imaging: a comprehensive evaluation of a fast spin echo DWI sequence with BLADE (PROPELLER) k-space sampling at 3 T, using a 32-channel
head coil in acute brain ischemia. Invest Radiol. 2009 Oct;44(10):656-61.
3.
Hornak, JP. The Basics of MRI. 1996-2011. Rochester Institute of Technology. Accessed 8/2012. http://www.cis.rit.edu/htbooks/mri/
4.
Morelli JN, Runge VM, Feiweier T, Kirsch JE, Williams KW, Attenberger UI. Evaluation of a modified Stejskal-Tanner diffusion encoding scheme,
permitting a marked reduction in TE, in diffusion-weighted imaging of stroke patients at 3 T. Invest Radiol. 2010 Jan;45(1):29-35.
5.
Morelli JN, Runge VM, Ai F, Attenberger U, Vu L, Schmeets SH, Nitz WR, Kirsch JE. An image-based approach to understanding the physics of MR
artifacts. Radiographics. 2011 May-Jun;31(3):849-66.
6.
Runge VM, Nitz WR, Schmeets SH, Schoenberg SO. Clinical 3T Magnetic Resonance. New York: Thieme, 2007.
7.
Runge VM, Morelli JN. Essentials of Clinical MR. New York: Thieme, 2011.
8.
Runge VM, Nitz WR, Schmeets SH. The Physics of Clinical MR Taught through Images. New York: Thieme, 2009.
9.
Schaefer PW, Grant PE, Gonzalez RG. Diffusion-weighted MR imaging of the brain. Radiology. 2000 Nov;217(2):331-45. Review.
10.
Testaverde L, Caporali L, Venditti E, Grillea G, Colonnese C. Diffusion tensor imaging applications in multiple sclerosis patients using 3T magnetic
resonance: a preliminary study. Eur Radiol. 2012 May;22(5):990-7. Epub 2011 Dec 9.
11.
Xu J, Rasmussen IA, Lagopoulos J, Håberg A. Diffuse axonal injury in severe traumatic brain injury visualized using high-resolution diffusion tensor
imaging. J Neurotrauma. 2007 May;24(5):753-65.