5. Principles of CT and MRI Flashcards
What are the 2 primary advantages of CT / MRI over other modalities? 1 disadvantage?
Adv:
1) Tomographic nature
2) Increased contrast resolution (e.g.smaller differences in x-ray attenuation detectable in CT due to reduced scatter and more sensitive detectors -> fluid from ST!; MR even greater contrast resolution -> use of combined sequences)
Dis:
1) POORER SPATIAL RESOLUTION
=>CT 0.3mm and MR 1mm respectively
What is generally considered to be the limiting factor to resolution in CT / MR?
- SLICE THICKNESS -> tends to be largest voxel dimension
What are the advantages of helical scanning? When / why would you use sequential scanning?
=> produces data volume rather than single slices . Interpolation of data to reonstruct
- Reduced motion artefact
- Increased speed
- Less prone to step artefact
- May use sequential when e.g. scanning area with limited motion such as head. Less strain on tube
What are CT detectors made from?
- Ceramic solid state detectors
=> scintillation crystal -> reacts with Xray, and through amplifaction emits light, converted to digital
What does multi-detector refer to?
- DEPTH OF ACQUISITION -> multiple detectors in plane of gantry allow multiple contiguous slices to be acquired at a time
=> QUICKER ACQUISITION
e.g. 2.5mm thick slices of 30mm thorax: 64 slice <2 secs; 1 slice 60-120 secs
reduced interscan delays
=> MORE EFFICIENT USE OF RADIATION
What does temporal resolution refer to?
- Resolution improvements due to minimising motion artefacts and tissue misregistration. See with MDR scanning and helical aquisition
What is Linear Attenuation Coefficient?
Attenuation mainly depends on ELECTRON DENSITY OF A MEDIUM
LAC: Measures fraction of radiation removed in passing through a given thickness of a specific material
=> absorption probability described by LAC (µ)
Nt = N0 e-µx
Where N0 = Number of initial photons at tube exit, Nt = number of transmitted electrons meaured by detector, e = base of natural log (2.718), x = thickness of absorber, and µ is LAC present along xray path
Rearranged, LAC can be derived as N values measured in system, and other values are known.
THIS CALCULATION is performed multiple times by the machine! approx 800 transmission calculations, 1000 diffferent projection angles per image! => 800,000 measurements total
What is filtered backprojection?
- Once µ is calculated for a single ray -> mathmateical process (think sudoku) where voxel values are assigned
Values are transformed into HU / CT numbers
What is the formula for calculation of HU for different tissues?
HU of tissue = [(µtissue - µw) / µw] x1000
Where pure water µw = 0
In basic terms, how to reconstruction filters affect image?
- Can determine level of edge reinforcement in raw data:
Low pass / ST: emphasis ST, smoother but more blurry e.g. brain
Bone: more spatial resolution -> sharper but grainier (noisy)
What is the typical range of HU measureable by CT?
- 1000 - +3095
- 4096 shades of grey (12 Bits)
What do window width and window level refer to?
- ww = number of shades of grey displayed (max to min)
- wl = HU at centre of window
=> adjusted for specific tissues
List 3 uses of CT contrast
- Evaluation of perfusion characterisitcs of tissues
- Angiography -> assessment of vascular phases
- Excretory urography (superior to XR) -> estimate GFR and assess collecting system
What are the advantages / disadvantages of cone beam CT?
Cone shaped beam -> reduced patient radiation
Adv:
- High spatial resolution
Dis:
- Increased scatter
- Lower contrast resolution (Lower contrast to noise ratio)
- Lower temporal resolution of cesium iodide detectors -> inc motion artefact
- Longer reconstruction times
What is Faraday’s law of induction?
- When a current is fun through a coiled wire, a magnetic field is produced in a direction that is perpendicular to the flow of current, and that is proportional in strength to the current.
What is B0 and what is net magnetization? What features contribute to / affect net mag?
- B0 refers to the axis of the externally administered magnetic field
- Net mag: Spins align with B0, SLIGHTLY more parallel with field > antiparallel with field
- > Net mag increass proportionally to STRENGTH OF FIELD
- > Dependent on PROTON DENSITY of tissues
What is precession? And what is larmour freq?
- Precession describes the wobbling behviour of spins under the influence of B0
- Motion controlled by RF pulse
- Larmour (/resonance) freq = frequency of precession of protons in B0, and is proportional to magnet field strength
w0 = gammaB0
Where gamma = constant, or gyromagnetic ratio, for each type of nuclei
Hydrogen protons -> GM ratio = 42.6MHz/Tesla
Define flip angle, resonance, excitation and relaxation
- RF pulse application to protons at larmour freq -> energey transfer (RESONANCE) causes energetic EXCITATION of protons, with more protons adopting antiparallel orientation
=> Results in net macroscopic magnetization shift away from z axis, towards xy axis = angle of change in direction of mag = FLIP ANGLE
- Once RF removed, return of spins to normal equlibrium = RELAXATION
Define the processes of T1 and T2 relaxation
- Realignment of spins with B0 on end of RF pulse -> Relaxation in longitudinal axia, with loss of energy into lattice = T1 (longitudinal)
- Spins precess in coherent fashion as approaching XY axis. Loss of phase coherence after RF removes releases energy into lattice = T2 (transverse)
TWO DISTINCT PROCESSES OCCURING SIMULTANEOUSLY
=> tissue specific
What is T2* relaxation?
- Occurs due to inhomogeneity in the magnetic field e.g. metal, blood, calcium, air, or local variation in magnet strength
- In stead of normal dephasing of spins once RF removed, very rapid dephasing occurs -> T2* relaxation
- Tissues dont relax at specific time associated with tissue type, but much faster
Define TR and TE
SPIN ECHO SEQUENCES - designed to control for T2* relaxation
- TE: time of echo
Time from iniital 90deg pulse to echo. Echo results after second (usually 180deg) rephasing pulse -> this allows for ‘slow’ and ‘fast’ processing spins to become in synch, with resultant larrge transverse signal (echo)
- TR: Time to repetition
Time from 90 deg pulse to 90 deg pulse (or repeat of whole sequence)
NB: Basic sequences, 1 row of image data each TR, so: 256x256 matrix, 500msec TR => 128,000 m/sec (128 sec) to acquire
How do recieving coils receive signal?
- Aligned loops of wire perpendicular to transverse axis -> when spins in transverse plane, induce current in coils (proportional to transverse field strength)
Explain how TE / TR relate to T1w, T2w and PD images
In each instance, need to seperate tissues based on slow or fast relaxation (and thus differential signal intensity)
- T2w:
Reliant on TE
LONG TR, LONG TE
Allow short T2 tissues to decay, while highlighting signal from low T2 tissues (high signal at time of echo)
- T1w
Reliant on TR
SHORT TR, SHORT TE
Allow short T1 tissues to decay while highlighting signal from long T1 tissues
** OPTIMIZING ONE INHIBITS THE OTHER!**
- PD images
Both T1 and T2 effects inhibited
LONG TR and SHORT TE
Varying intensity levels for same tissue
Describe the different gradients used to localise components of the image
- Slice selection gradient:
Selects slice along B0 -> Gradient creates differential magnetif field strength, and as such allows differential slices to be flipped due to predictable variation in larmour frequency (fucntion of magnet strengthY)
- Phase encoding gradient:
Soon after 90deg pulse
Gradient across slice -> Different PHASE
- Freq encoding gradient:
During echo
gradient across row -> Differental PRECESSIONAL FREQ
=> allows individual voxel signal to be mapped
What are fast spin / turbo spin sequences?
- Spin echo sequences where multiple 180 rephasing pulse sequences are applied for a single TR
- More signals localised, speeds up acquisition
How are inversion recovery sequences produced? What is Time of inversion (TI)?
- Initial 180 pulse -> reverses proton alignment (-z)
- Relaxation towards +z, all tissues eventually cross z=0 (nulled) -> predictable
- If normal Spin echo seq started AFTER 180 pulse, can place read signal when desired tissue is nulled
=> DEPENDENT ON T1 RELAXATION TIME (LONGITUDINAL)
TI = Time from 180 inversion pulse and 90 deg RF pulse
FAT = SHORT TI as short T1 relaxation (STIR IMAGES USE THIS)
FLAIR similar, but can be T2 or T1w
How do gradient echo sequences differ from spin-echo?
- Technique:
Use smaller flip angles
No 180 deg refocusing pulse
Gradients used to dephase and rephase transverse mag -> generate echoes
Shorter TR
No compensation for T2* relaxation / inhomogeneity -> if long TE = T2* weighted
- Time:
Short TR and smaller flip -> Shorter aquisition
Reduced motion
- Utility:
Angiography (less motion)
Magnetic susceptility => haemorrhage
Describe the different types of magnetic susceptibility properties
- Paramagnetic:
Small +ve susceptilbility
E.g. magnesium, molbdenum, lithium
- Ferromagnetic
Strong +ve susceptibility
E.g. iron, nickel, cobalt
- Diamagnetic
Weak -ve susceptibility
E.g. gold, silver
What is the predominant action of gadolinium?
- Shortened T1w relaxation -> generate greater signal on T1w seq
BIt confusing, but correct. Either accept or investigate further….
How does chemical fat saturation work?
- Only in high field
- > Exploit difference in precessional freq of fat and water = 220Hz at 1.5T, becomes smaller at lower strengths
- Freq specific preparation pulse -> selectively excite lipid protons, followed by spoiling gradient that dephases fat signal
=> Generated signal only arises from non-fatty tissues
Briefly, what are DWI and PWI?
- DWI = diffusion weighted. Sensitive to BROWNIAN motion of water molecules -> THUS CYTOTOXIC OEDEMA, seen in restricted diffusion with stroke
- PWI = T2*w following bolus contrast -> semiquantifiy susceptibility-induced signal loss over time, and thus blood flow.
How are SNR and Spatial resolution related?
- Spatial resolution: improved by decreasing voxel size e.g. smaller slices, smaller FOV, larger matrix
- SNR: requires volume of tissue per voxel, so decreases with factors improving resolution. Can average signal over multiple acquisitions (NSA), but at cost of increased scan time.
- > also, some coils help. Parallel imaging = use of multichannel phased array coil. Signals from each element combined to imprve SNR without longer aquisition
List some benefits / disadvantages of low field MR (<1T)
- Adv:
Less susceptibility
Open magnet design, can help with patient access
Accidents more preventable
Dis:
Reduced SNR / spatial resolution
