MRI Flashcards
MRI resolution
1-2 mm FWHM (0.3 lp/mm)
depends on acquisition parameters
MRI matrix size
256x256, 2 bytes/pixel, 1/8 MB
does T2 depend on FS?
no
what do contrast agents generally do
reduce relaxation times
as BW increases, noise..
increases
BW relation to gradient strength
BW is proportional to G
electron magnetic field vs proton
electron is 1000X stronger
paramagnetic
increase magnetic field
deoxyhemoglobin, gadolinium
diamagnetic
decrease magnetic field
-tissues and plastic
what is magnetic field strength
number of magnetic lines per unit area
gyromagnetic ratio of H
42.6 MHz/T
soft tissue T1 and T2
T1 = 500 ms
T2 = 50 ms
fluid T1 and T2
both > 1000 ms
bone T1 and T2
T1 > 1000 ms
T2 < 0.01 ms
what molecules have short T1?
those that move at larmor frequency
T2 time, FID is where?
FID decayed to 37%
net magnetization at 1 T
3 in a million
what is magnetic susceptibility?
local M changes due to electron interactions
paramagnetic contrast agents work how?
by disrupting local field homogeneity, therefore T2* decreases
what does quadrupling field strength do to T1?
doubles it
super-paramagnetic
develops magnetization when placed in a magnetic field
T2 time, where is Mz?
64% of Mz has formed
full Mz forms after 4T1
after what time is Mxy null?
after 4T2
earth’s magnetic field strength
50 uT
typical gradient strength
30 mT/m
higher gradient strength = thinner slices
how many echoes do you need to obtain 128x128 MRI image?
128 echoes and sample the echo 128 times
ferromagnetic
dramatically increases local susceptibility
homogeneity of MR magnet
should be within a few ppm
what is used to select the voxel?
slice select, frequency encode, phase encode
shimming
correct main field to improve homogeneity
type of coil for x, y, z gradients
Helmholtz= z
saddle = x and y
quench
loses superconducting properties, magnetic energy converted to heat
dielectric artifacts
-occurs at high fields because RF waves are short compared to body size
-standing waves can have constructive and destructive interference
linear coil vs quadrature coil
linear receives signal only from x or y
quadrature receives from both, increases SNR by root2
distance of 5 G line from isocenter of 1.5 T magnet
-12 m unshielded and 4 m shielded along bore
-9.5 m unshielded and 4.5 m shielded perpendicular to bore
RF shielding used
faraday cage
where is MRI access restricted?
within 5 G line
T2 times vs T1 times
T2* 10x shorter than T2 and 100X shorter than T1
BW vs slice thickness
larger BW = thicker slices
signal of surface coil
signal drops off as you go away from coil
slew rate
time required to achieve desired magnetic field amplitude
flip angle formula
gamma * B1 * time
what artifact does RF leakage into MR suite cause?
zipper artifact
active vs passive shielding
active: use coils to make magnetic fields to cancel out main field
passive: sheets of metal are permanently installed
why can fat saturation failure occur?
due to non-inhomogeneities
equation for slice thickness
transmit BW/ (gamma*gradient strength)
In STIR, what IR is used to suppress fat?
250 ms
brighter values on ADC map is more or less diffusion?
more diffusion
think it is opposite…
FLAIR
fluid attenuated IR
suppresses signal from fluids
TI = 1700 ms
3D MRI imaging time
phase encode x * phase encode y * TR
TOF
ex, GRASS, FISP
-stationary tissue becomes saturated (no Mz to flip)
-fresh blood always has Mz to flip
MR angiography
subtract GD contrast from non contrast- get image of veins
-useful for patients who can’t have iodinated contrast agent
fMRI
relies on blood O2 or flow changes with activity
BOLD
oxyhemoglobin has longer T2 than deoxyhemoglobin
-brain activity increases local venous blood oxygenation, which increases T2*
-change from diamagnetic oxyhemoglobin to paramagnetic deoxyhemoglobin that happens with brain activity yields decreased signal intensity on MRI
magnetization transfer
-bound protons(short T2) are saturated- exchange signal with free water
-free water becomes partially saturated due to coupling between the pools
-free water image free water before and after- drop in signal is related to bound proton content
what is the bright area on trace DW images in combination with high ADC value?
T2 shine through
-reflects tissues with very long T2 values
b values in DWI
the higher the b value, the stronger the diffusion effect - controls degree of diffusion weighting
trace DW image and ADC map
diffusion-weighted map
-mathematically remove T2 effects
-can obtain ADC map from here- ADC map intensities are opposite to those of DW
-lesions with low diffusion have low ADC- they appear bright on DW and dark on ADC
-onbtain 2 images: one with diffusion gradients and one without
MRS voxel sizes for 1H and 31P
1 cm3 for 1H
8 cm3 for 31P
advantage of surface coils
less noise
-closer to anatomy (more sensitive)
temporal and spatial resolution of PET vs fMRI
fMRI has better temporal and spatial resolution
echo train length
of echoes acquired per each TR in fast spin echo
chemical shift artifact type II
fat and water protons are in and out of phase- get light or dark bands
what is aliasing caused by
-caused by undersampling
-remove by increasing FOV
images required to get complete view of flow
need 3 images
what do susceptibility effects do to EPI
Degrade it
what does Fe2O3 do?
increases lesion contrast in T2 weighted image
Gibbs artifact
bands adjacent to edges
-there isn;t high enough frequency to model the edge or interface
most important determinant of MR quality
SNR
time dependence on receiver BW
as receiver BW increases, time decreases
noise dependence on receiver BW
noise increases when receiver BW increases
what does taking 4 acquisitions do to SNR and imaging time
-4 acquisitions = 4 x signal but 2x noise, therefore get 2x SNR
-4X imaging time
concerns about 10T
-can get hazardous biological effects
what does scan time depend on?
scan time = TR *Np * signal averages
effect of FOV on pixel size
-smaller FOV = smaller pixel size if matrix is unchanged
SNR dependence on Bo for body noise dominant and electronic noise dominant
SNR ~ Bo for body noise
~Bo^7/4 for electronic noise
SNR equation
~ proton density
~voxel volume
~ Bo (body noise dominant)
~square root (NSA * TR * Np)
NSA TRNp is total time
if it is 3D MRI then total time is NSA * Nz * Ny *TR
what sequence does BOLD use
EPI with T2* weighting
hearing noise level in MRI
65-120 dB
fats on T1 images
bright
max ramp speed to prevent nerve stimulation
3T/s
what is the point of using a gradient echo?
-T2* but instead of getting FID you get a centered signal
what causes ghost image
periodic patient motion in phase encode direction
Using faster sequences and respiratory-ordered phase encoding can help eliminate artifacts, as can navigator echo gating to track the motion of the diaphragm to time image acquisition
TE and TR for T1 contrast
TE = 20 ms
TR= 500 ms
TE and TR for T2 contrast
TE=100 ms
TR= 2000 ms
TE and TR for density contrast
TE= 20 ms
TR= 2000 ms
how to increase T1 weighting in GRE
increase flip angle
whole body SAR heating limits
-normal- 0.5 degrees C
1st level controlled- 1 degree C
2nd level controlled- > 1 degree C
phase encode direction for abdomen
AP
phase encode direction for head
lateral
issues with 3D MRI
longer acquisition time = more motion artifact
-gaps or overlaps between slices
rules for average SAR in head and body
3 W/kg head
4 W/kg body
fluids on T2 weighted images
bright
chemical shift artifact type I
mis-registration of fat leads to bonds at interfaces
magic angle
tendons align at 55 degrees to the main field and yield longer T2 times that can appear bright
-still has short T2 time that only appears for short TE though
4 zones of MRI
free access, interface, restricted, MR room
bright and dark areas near a foreign object is due to what artifact?
magnetic susceptibility
how to minimize chemical shift artifacts
-increase BW
-reduce matrix size (ie reduce number of pixels)
-suppress fat-frequency signal
flow artifact
Due to the nature of GRE sequences, blood flow will produce a bright signal. Using saturation bands can minimize this artifact, saturating the slice upstream so the blood will not produce a signal. Gradient moment nulling can also be applied to try to correct flow artifacts.
what does spike in k space cause
streaks in image
size and location determine angle and width of streaks
gradient field distortions
produce image distortions
eddy currents cause
geometric distortions
metal artifacts
using spin echo improves issue (removes effect of static inhomogeneities), but still have distortion
chemical shift difference btween water and fat
3.5 ppm
fat resonates at slightly lower frequency and is thus placed at lower voxel (chemical shift artifact)’
The size of the chemical shift artifact can be readily computed in advance based on two parameters selected prior to imaging: receiver bandwidth and size of the frequency-encode matrix. For example, if the total receiver bandwidth selected is 32 kHz and with 256 pixels in the frequency-encode direction, the bandwidth per pixel is 32,000/256 = 125 Hz. Since the fat-water frequency difference at 1.5 T is about 215 Hz, the size of the chemical-shift artifact will be (215 Hz) ÷ (125 Hz/pixel) = 1.7 pixels.
Reducing the bandwidth per pixel will accentuate this artifact. Narrow bandwidth techniques should therefore generally be avoided in locations where chemical-shift artifacts may obscure important interfaces (for instance, between the optic nerve and orbital fat).
number of hydrogen protons per cm3 tissue
10^22
spin density of lung, bone, fat
3%,5%,98%
this is why you cant see lung or bone on proton density weighted MRI
what materials have long T1?
liquids and solids
what are the tissue differences in proton density
10%
how do we improve resolution in MRI
stronger gradients
high SNR
large data acquisition matrix
examples of ferromagnetic materials
iron
nickel
cobalt
equation for gyromagnetic ratio
eg/2m
charge/2m
what causes SAR to increase?
-field strenght
-RF power and duty cycle
-transmitter coil type
-body size
T2 of CSF vs bone
T2 of CSF is longer (more fluid like)
what appears brightest on T1 weighted images?
fat
what type of MRI is less sensitive to motion
radial
what does SAR increase with
duty cycle
flip angle^2
Bo^2
permanent vs resistive vs superconducting magnet
0.35 T - temperature sensitive
0.5 T - uses a lot of power
up to 20 T
Permanent MRI magnets use permanently magnetized iron like a large bar magnet that has been twisted into a C-shape where the two poles are close together and parallel.Their magnetic field homogeneity is also sensitive to ambient temperature so room temperature must be controlled carefully. The initial purchase price and operating costs are low compared to superconductive magnets
Resistive (air core) MRI magnets operate at room temperature using standard conductors such as copper in the shape of a solenoid or Helmholtz pair coil. A solenoid is a cylindrical-shaped coil of wire. The uniform magnetic field is found inside the coil, especially in the center. These magnets are relatively inexpensive to make but require a large constant flow of current while magnetized and imaging. The coil has an electrical resistance that requires cooling of the magnet. The operating costs are high because of the large power requirements of the magnetic coils and the associated cooling system.
Both permanent and resistive MRI scanners are limited to low-field applications, primarily open MRI and extremity scanners. These magnets are useful for claustrophobic patients
Superconductive MRI magnets use a solenoid-shaped coil made of alloys such as niobium/titanium or niobium/tin surrounded by copper. These alloys have the property of zero resistance to electrical current when cooled down to about 10 kelvin. The coil is kept below this temperature with liquid helium. The power supply is connected on either side of a short heated segment of the coil and the current to the coil is gradually increased over several hours until the desired magnetic field is reached. The heated segment is allowed to cool to superconducting temperature and the power supply removed and taken away. The current continues in the closed-loop of the coil for years without significant decline. A resulting property is that the magnetic field is always present.
The surrounding copper acts as an insulator at low temperatures compared to the zero resistance of the alloy. The copper also protects the alloy coil from being destroyed in case of a quench of the magnet. A quench can occur if the helium levels drop too low or if a large ferromagnetic object is brought into the fringe field of the magnet. A quench results in loss of superconductivity with a large amount of heat produced by the current and rapid boiling-off of the cryogen. The gas produced is vented out of the room but can occasionally enter the scanner room with life-threatening consequences
The cost of cryogen replacement is reduced on modern magnets, which incorporate a refrigeration system called a “cold head” to condense the cryogen gas. Startup costs for the scanner can run up to about $1.5 million for a 1.5 T MRI. Site preparation can frequently run into several $100,000s including room radiofrequency (RF) shielding, possible magnetic shielding, floor reinforcement, vibration mitigation and a very reliable uninterruptible power supply (UPS).
Superconducting magnets at 1.5 T and above allow functional brain imaging, MR spectroscopy and superior SNR and/or improved time and spatial resolution. Magnets above 1.5 T have additional challenges from RF heating of the subject, and increased artifacts from susceptibility and RF penetration among others.
optimal flip agle for SSFP
cos theta= (T1-T2)/(T1+T2)
