Chapter 11 CT I Flashcards
how is power to the x-ray tube supplied?
slip ring technology
highest tube current
1000 mA
shortest rotation time for 360 degree of x-ray tube
0.3 s
typical is 1 s
focal spot sizes
1.2 mm and 0.6 mm
relative x-ray output
mAs
absolute output depends also on tube design and kV
how is the heel effect minimized?
anode-cathode axis is perpendicular to imaging plane
cost of CT x-ray tube
200 k
max power to CT x-ray tubes
100 kW
-use of 0.6 mm focal spot requires power reduced to 25 kW
heat deposition into anode
100 kJ every s
heat lost from x-ray tube
10 kJ/s, ten times less than rate of heat deposition
anode heat capacities
5 MJ, 10x higher than for normal radiographic and fluoro tubes
what can happen if CT are operated for a long time?
tube overheating
filtration on CT vs radiography and fluoro
heavier
reduces beam hardening effects
bow tie filter
minimize beam hardening differences
transmission through thinner body regions result in higher detector signals- prevent this with bow tie filter
low Z material (Teflon)
reduce scatter
reduce patient dose
antiscatter collimation
thin lamellae
between the detector elements
why do CT detectors have high data output rates (bandwidths)
multi-slice scanners
short gantry rotation times
photon starvation
small number of photons reaches detector
when are CT not quatum mottle limited?
when x-ray signals are low and electronic noise becomes significant
characteristics of CT scintillators
-low after-glow characteristics
-rapid signal decay
-scintillation detectors are coupled to light detectors
-high quantum efficiency (absorb most of the incident x-ray energy)
typical distance from x-ray tube focus to isocenter
60 cm
magnification of objects at patient center
X2
third generation CT system
-both x-ray tube and detector rotate around patient
FOV for fan beam of 50 degrees
50 cm diameter
how many detectors in fan beam?
up to 1000
~ 1.2 m from focal spot
detector offsets
-central rays acquired at 0 and 180 degrees do not overlap, but are offset by half detector width
-offset improves data sampling and spatial resolution
how many axial slices in a typical single rotation?
64
detector width of 0.625 mm
total beam width = 64 * 0.625 = 40 mm
such a scanner would cover a 32 cm long abdomen in 8 rotations
what are CT images?
2D patterns of x-ray attenuation coefficients
what do you call the x-ray transmission that each individual detector measures?
ray
what is a projection?
collection of rays for all detectors in each slice at a given tube angular projection
how many individual data points in a projection?
1000
corresponds to number of detectors (each has a ray)
how many projections are acquired for a single rotation of the x-ray tube?
1,000
what is sinogram
plot of projection vs x-ray tube angle
-each reconstructed slice requires a complete sinogram
axial scanning vs helical scanning
axial: table and patient are stationary while the tube rotates and acquires the necessary projection data
helical: patient moves through the gantry at same time as x-ray tube rotates. Reduces CT scan time
pitch
table increment distance/ (x-ray beam width)
pitch < 1 is oversapled
pitch > 1 is undersampled
interpolation algorithms for helical CT
at any given long patient axis position, helical CT scanning provides one of the 1,000 projections that are required
The 999 missing lines are obtained with interpolation algorithms with projections upsteam and downstram of the location of interest
-after interpolation, a sinogram is generated at eah z-axis location where image reconstruction is to be performed
effective mAs in helical scanning
true mAs/CT pitch
-increasing effective mAs increases patient dose and reduces image mottle
filtered backprojection
-acquired projection data are multiplied by a filter
-filtered projections undergo backprojection
-back projection allocates each ray value equally to each pixel along ray traced through the patient
ramp filter
increases with increasing frequency
generates noisy images
modified ramp filters
reduce noise but at expense of resolution
Shepp-Logan
Hanning
cosine
where would you use filter that reduces noise? improves resolution?
reduce noise- low contrast liver lesion
good resolution- inner ear
iterative reconstruction algorithms
-starts with assumed image
-computes projections from the image
-computed projections are compared with acquired projections
-improvements are based on differences between calculated and acquired projections
-process is repeated until agreement is satisfactory
pros and cons of iterative reconstruction algorithms
-better reconstruction
-higher computation time
where are iterative reconstruction algorithms always used?
PET and SPECT
-will eventually replace filtered backprojection in CT
HU values of water, air, fat, soft tissue, bone
0, -1000, -100, 50, > 1000
lesions with HU value of 10 attenuate what percent more than water?
1% (i.e. (10-0)/1000 *100%)
doe attenuation coefficients depend on photon energy? Z?
of course
HU increases as photon energy decreases
changes in HU with energy are large for high Z materials (iodine (Z=53) with HU of 400 at 80 kV would have HU of 100 at 140 kV)
soft tissue (Z~7) HU drops from 60 to 50 with tube voltage change from 80 to 140 kV
definition of FOV
diameter of body region being imaged
head CT FOV
250 mm
body CT FOV
400 mm
CT image matrix sizes
512 X 512
pixel sizes
0.5 mm head scan (250 mm/512)
0.8 mm body scan (400 mm/512)
fundamental factors that limit CT resolution
focal spot size
x-ray detector size
smaller pixels won’t improve resolution if inherent limits are reached
slice thickness
1.25 - 5 mm
-too thin will be too noisy
what is each pixel coded with?
12 bits = 2 bytes of computer memory
how many shades of grey can be used with 12 bits?
2^ 12 = 4096
storage requirement for each CT image
0.5 MB
i.e. 512X512 pixels * 2 bytes/pixel
chest/abdo/pelvis CT has 100 images - 50 MB
-acquired raw data takes even more storage because each sinogram needs 2 MB/slice
-may be more acquired thin slices than displayed thick slices
what are CT images displayed on?
2 MP monitors - display up to 8 512^2 images at full res
what happens to lesion brightness as level is increased?
-brightness is reduced
HU value never changes only brightness
what happens to image contrast as window width increases?
contrast decreases (but see larger range of HU components)
i.e. if window width is 50, all HU values outside of level +/-25 appear black
window level settings for head
window = 80
level = 40
window level setting for chest (mediastinum)
window = 400
level = 50
window level setting for chest (lung)
window = 1500
level = -500
window level setting for abdomen (liver)
window = 150
level = 60
how is CT output best quantified
CTDIvol
-with identical techniques, small phantoms record twice the dose as large phantoms
what is CTDIvol proportional to?
at constant kV it is proportional to mAs
at constant mAs, it is proportional to kV^2.6
what is dose-length product?
CTDIvol * scan length
measures total amount of radiation incident on a patient
proportional to patient stochastic radiation risk
when does increasing CT beam quality reduce patient dose?
when tube output (mAs) is adjusted so Kair at CT detectors is constant
-must be adjusted by operator in CT (not done automatically after quality change like it is in mammo, fluoro, radiography)
what determines patient transmission?
beam quality
attenuator atomic number
density
thickness
tissue HVL for typical CT
4 cm
transmission through obese patient
patient thickness for neonate is 10 cm
50 cm for obese patient
40 cm is 10 HVLs
-thus obese patients transmit 1000 less than a neonate (2^10)
when patient penetration is inadequate, increasing mAs does not improve image quality
-must improve quality to increase penetration
HVL of 80 kV CT
4.5 mm Al
HVL of 120 kV CT
9 mm Al
what does AEC do in CT?
CT doesn’t have radiation detector at image receptor (unlike fluoro, mammo, radiography)
-operator must set kV, mA, and rotation time for average patient attenuation
AEC increases mA when patient attenuation increases and vice versa
If operator changes beam quality, they must adjust mA manually (it is not done automatically like in fluoro, radiography, mammo)
angular modulation
tube currents are modulated as x-ray tube rotate around patients
-tube current for lateal projection through shoulders will be more than AP or PA projection
longitudinal modulation
tube currents are modulated as patients move through the gantry
tube current in chest will be lower than in the abdomen
commercial AEC
vendors have different programs to adjust current based on patient size etc
-includes angular and longitudinal modulation
making the lamella in anti-scatter grid longer will result in better?
contrast
-less scatter
what does smooth filter too?
-reduces noise but also hurts resolution
1 HU corresponds to an increase of what % in attenuation?
0.1%
a head CT with 20 images requires how much storage space?
one CT image has 500x500 = 250,000 pixels
2 bytes required per pixel
one image is 0.5 MB
20 images is 10 MB
are anode capacities in CT higher than radiography?
yes 10 X higher
small focal spot (0.6 mm) power
25 kW
large focal spot (1 mm) power
100 kW
what are bow tie filters made of?
Teflon
tissue-like material to minimize beam hardeningn
how much of the incident fluence do CT detectors absorb
> 90%
doubling both CT pitch and effective mAs would do what do patient dose?
double it
patient dose is proportional to effective mAs. Doubling mAs at twice the pitch value requires mAs be quadrupled
thus patient dose doubles
a head CT with 20 images requires what storage space?
one CT image has 500x500 = 250000 pixels, and requires 2 bytes for each pixel
20 images thus need 10 MB
dose from projection radiograph is what % of dose from CT chest exam?
< 5 %
what kV are dual energy CT scans acquired at?
usually highest and lowest kV available
contrast between fat and tissue
fat = -100 HU
tissue = 50 HU
difference is 150 HU
best achievable temporal resolution for dual source CT
1/4 of xray tube rotation time
directly irradiated organs receive more or less than CTDI dose?
more
-about 50% more
