Computed Tomography
The radiation dose from 15 consecutive (no overlap) CT slices is ___x greater than the dose from a single slice
1.3. Doses from adjacent slices do not add together b/c different tissue is being exposed. The definition for absorbed dose is energy deposited per g of tissue; total amount of tissue irradiated is not assessed w/ this. The only additional dose is ~20-40% increase from scattered radiation
How much overscanning is necessary to acquire a helical image?
1/2nT in either direction. For example, for a scan where nT = 20 mm and a desired scan length of 200 mm, there would be 220 mm of acquisition, 10% wasted dose
How is 1/r blurring expressed in the Fourier domain?
1/f
For the 120 kV-140 kV spectra, Compton scatter interactions are ___x more likely than the photoelectric effect in soft tissue, and Rayleigh scattering is even lower
10
Image acquisition time window required for cardiac CT
100 ms or less (heart rate of 60 bpm or 1 cycle per s)
Typical number of protocols per CT scanner
100-300
The dose for a single CT slice (Averaged through the abdomen) is about 3.0 cGy. A helical study is performed w/ double mAs and effective pitch of 0.5, covering 20 cm. The CTDI vol would be expected to be about ____
16 cGy. CTDI is about 1.2-1.4x the single slice radiation dose. Effective pitch, double mass = 4x.
How does CTDIw vary with mAs at 120 kVp?
16 cm phantom: CTDIw ~ 0.2 mGy/mAs 32 cm phantom: CTDIw ~ 0.1 mGy/mAs
Max HU value on most CT scanners (bone/iodine)
3095
Typical true cone beam specifications
30x40cm flat panel detector, SSD 90 cm, 25 deg fan angle and 19 degree cone angle
For a typical CT scanner system with about a 50-cm source to isocenter distance, the circle that defines the x-ray tube trajectory is about 3140 mm (2pir) in circumference. Therefore, using a 2-kHz detector sampling rate, the x-ray focal spot moves about ____ along the circumference of the circle per sample
3140/2000 = 1.5 mm
CT ring artifact
3rd gen CT uses source fan geometry where the signal from 2 different detectors is used to make the 2 x-ray intensity measurements Ir and Ij necessary to compute a projection value. If one detector drifts in terms of its calibration w.r.t. the other detector, the error in the projection measurement Pj will be propagated throughout the backprojection process, leading to a ring artifact
Typical CT head perfusion protocol
40 images, each using a 1-s acquisition, resulting in ~40 s of scanning in the same area of tissue.
Effective energy of 80 kV CT spectrum
40 keV
Typical CT slice thickness reconstruction selections
5 mm axial slices, 0.5/0.625 slices saves for coronal or sagittal reconstruction
What kind of filtration is used in CT?
5-10 mmAl
The effective dose for an abdomen w/ MSAD 16 cGy covering 20 cm CT would be expected to be about...
50 mSv. Either multiply by length to get DLP and then by 0.15 weighting factor for abdomen, or recognize that effective dose for abd is usually about 1/3 of CTDIvol
SSP/ACR Slice Thickness Measurement Image
6 wires visible, multiply by 0.5 mm, ~3 mm slice thickness
Approximate fan angle of most CT scanners
60 degrees
Effective energy of 140 kV CT spectrum
60 keV
For a typical 64 to 128 detector array CT scanner, the fan angle is approximately ___ degrees, while the full cone angle is about ___ degrees
60, 2.4
For a CT scan performed with iodinated contrast, what kVp can be used to minimize the dose while maintaining similar CNR?
80 kVp, or 100 for larger pts. to reduce noise
Typical CT kVp
80, 100, 120, 140
If the CTDIw for an adult abdomen CT is about 3.0 cGy, the CTDIw for the 6mo baby using the same CT scanner settings would be expected to be about ___
9 cGy. For the same scan settings, CTDI for small babies is 3-5x greater in the center and 2-3x greater at surface due to small size
The 1.5 mm circumferential displacement of the CT x-ray source over the time it takes to acquire one projection can lead to motion blurring and a loss of spatial resolution because the patient is stationary. To compensate for this, some scanners use ____
A magnetic steering system for the electrons as they leave the cathode and strike the anode. With clockwise rotation of the x-ray tube in the gantry, the electron beam striking the anode is steered counterclockwise in synchrony with the detector acquisition interval and this effectively stops the motion of the x-ray source and helps to preserve spatial resolution.
Noise Power Spectrum
A spatial frequency-dependent function that describes how noise propagates through an imaging system. Describes both overall noise level and the noise texture. Describes how the noise at one point in the image is correlated to the noise in surrounding points
ACR Required CT QC Tests
A. Participation in Review of Clinical Protocols with the CT Protocol Review and Management Team B. Scout Prescription and Alignment Light Accuracy C. Table Travel Accuracy D. Radiation Beam Width E. Low-Contrast Performance F. Spatial Resolution G. CT Number Accuracy H. Artifact Evaluation I. CT Number Uniformity J. Dosimetry K. CT Scanner Display Calibration No more slice thickness
None/Some/Most/All high-end modern CT scanners use cone-beam acquisition
All
Typical CT gantry rotation speed
Approaching 5 rotation per second (0.20 s rotation time)
How is the noise power spectrum in 3 dimensions compressed to 2 in CT, and what does the formula look like?
Assume the NPS is rotationally symmetric
Why is a roll-off function needed for a ramp filter?
Because a simple ramp filter has a high amplitude at high frequencies, noise would dominate
Why is mA modulation important?
Because the body is not a perfect circle, and when the beam is traveling through a less thick region of the patient the extra dose is essentially wasted b/c the noise variance in the proj. data propagates into the recon. CT image in a manner more optimal when the pro. data variance levels are nearly matched. When mA modulation is used, lower radiation dose can be used to achieve the same image quality
What sort of preprocessing occurs in CT projection data?
Besides air calibration, scatter correction algorithms (before log. of the data applied), adaptive noise filtrations (eg algorithms to identify low signal areas and apply smoothing). After preprocessing, the projection data undergo logarithmic transformation and normalization, in order to correct for the exponential atten. characteristics of x-ray interaction
What are beta blockers used for in cardiac CT?
Beta blockers help restore regular heartbeats to patients with irregular heartbeats so that cardiac gating is easier
Digitally, what determines the highest CT number?
Bit depth (lowest # always -1000)
Why is an arc-shaped detector beneficial in CT?
By mounting the detector modules on a support structure that is aligned along the radius of curvature emanating from the x-ray source, there is very little difference in fluence to the detectors due to the inverse square law. Very importantly, the primary x-rays strike the detector elements in a nearly normal manner, and so there are effectively no lateral positioning errors due to x0ray beam parallax
CT Angiography
CT has higher contrast resolution than traditional angiography and consequently requires only venous access as opposed to more invasive arterial access for fluoroscopy-based angiography. CTA is really just contrast-enhanced CT where the images have been processed to highlight the vascular anatomy using image processing techniques. In general maximum intensity projection procedures produce excellent CTA images (may use pseudocolor as well)
Why is CTDIw weighted by 2/3 and 1/3?
CTDIw represents average dose over entire cross-sectional area. The weighting factor is to approximate the relative areas represented by the central and peripheral values
CT Motion Artifacts
Caused by pt motion, especially in ped exams. Fix w/: • Using immobilization device or sedation • High pitch scan mode ("flash" mode)
CT Aliasing Artifact
Caused by sharp bone edges or metal implant. To reduce, increase sampling rate (flying focal spot) or use low cut-off reconstruction kernel (smooth kernel)
CT Smudge Artifact
Caused by tube focal spot thermal drift or detector gain drift. Fix w/ tube replacement or detector gain re-calibration
CT Helical (Windmill) Artifacts
Caused by under sampling in z-direction. Reduce by increasing the sampling rate by using z-flying focal spot
Cone-beam artifacts
Caused by undersampling in the cone angle dimension (z) and in Fourier space, can be reduced with smaller cone angle
What is the purpose of air calibrations in CT?
Characterize the influence of the bowtie filter and the differences in individual detector response, which may be due to difference sin photodiode or amplifier gain. The measured projection data for a given CT scan to be reconstructed are normalized by the calibration scans, and this procedure corrects for previously identified inhomogeneities in the field. Dead pixels may be identified and replaced with interpolated data
How is the 3D Noise Power Spectrum calculated?
Computing the 3D Fourier transform of data acquired from a VOI
Features of true cone beam systems
Cone angle nearly as great as the fan angle. Planar flat detector with a 2D bank of detectors. Correction methods needed for different distances from source to detector ,giving rise to inverse square and hell effect-based differences in fluence to deach dexel. Some parallax.
Roll-off filter examples
Shepp-Logan, Hamming, Cosine. Vary by specific application (soft tissue, bone, standard or H47, B30)
How can the reduction of spatial resolution by the reconstruction in CT be reduced?
Create multiple reconstructions
What is forward projection?
Creating projections from central data (what the CT scanner does physically)
Shuttle Mode CT
Shuttle mode allows a CT to repeatedly image a volume of tissue that is wider than the detector array (nT). The table rocks back and forth a prescribed distance, during the temporal acquisition procedure. This mode allows CT perfusion analysis to be performed over a larger section of tissue. The temp. res. is lower in shuttle mode, but adequate physiological info can be derived. Due to the change in direction of the table, shuttle mode results in image data being acquired during table acceleration and deceleration, which implies that the reconstruction algorithm needs to accommodate for the changing pitch. Hybrid btwn. perfusion and helical imaging
Cone beam reconstruction
Similar to fan beam reconstruction algorithms, but they keep track of the beam divergence in the z-direction (Cone angle). Basic cone beam reconstruction process is Feldkampt algorithm (FDK). Cone beam recon. violates mathematical requirements for sampling Fourier space, and for extreme cone beam used in flat panels or other very wide CBCT systems, this can lean to cone beam artifacts
Advantages/drawbacks to Dual-Source DECT
Smaller Tube B FOV, which limits FOV of dual-energy reconstruction. Also b/c both imaging chains are active during acq., scatter from one imaging chains impacts the other and requires correction techniques, which ultimate result in more noise. However, more filtration can be added to the higher kV source, which increases energy separation and improves the energy subtraction.
How can detector size and sampling be optimized for spatial resolution in CT?
Smaller detector dimensions and oversampling (Nyquist) methods
Why does adipose tissue appear darker than soft tissue in CT?
Soft tissue components: H, N, O, C Typical Z/A ratios for NOC: 1/2 Typical Z/A ratio for H = 1 However, range of fluctuation for H is small, and lower mass density dominates. Adipose tissue has rho of 0.94 compared to soft tissue of 1
Caveat for reducing pitch
Some scanners may adjust other parameters besides pitch to maintain constant CTDIvol
Caveat for shortening tube rotation time to reduce dose
Some scanners may compensate for changes in rotation time with adjustments to pitch or mA in order to keep the mA per rotation constant • CTDI may not change in the expected manner • Varies by manufacturer and software version
Do CT scanners use 2D antiscatter grids?
Sometimes
What is one advantage of reconstructing CT slices into the coronal or sagittal planes?
Spatial resolution in the z-direction is not affected the reconstruction kernel
What factors can influence low contrast resolution in CT?
Technique factors, slice thickness, reconstruction filter, reconstruction method
Why is filtration necessary in backprojection?
The backprojection produces a characteristic 1/r blurring due to the geometry
What is "flying focal spot" in CT?
The e- beam is steered using magnetic fields provided by steering coils, and due to the steep anode angle, modulating the electron beam landing location on the focal track causes an apparent shift in the source position along the z-axis of the scanner. The result is oversampling in the z-dimension, improving the spatial resolution in z. Using n actual detector arrays in z, with 2 focal spot positions (about 1/2 a detector width apart) a total of 2N detector channels are formed
Why is CT especially susceptible to focal spot blurring?
The high mA used can increase the size of the focus (blooming)
Some uses of true cone beam imaging
Dental/maxillofacial CT cystems, breast CT systems, radiation therapy
Describe the ACR Phantom's Low-Contrast Section
Designed to have 6 HU (0.6%) greater density than background, but vary in diameter
How would you increase the dose by a factor 8 in a CT acquisition with a 0.5-s scan time, pitch = 1, 120 kVp, 300 mA?
Double the mAs to 600, double the scan time to 1 s, and decrease the pitch to 0.5
CT Truncation Artifacts
Due to patient position outside 50cm scan field of view. To fix, re-position pat inside SFOV (pt may be off-centered up/down/L/R w.r.t. isocenter)
What is the fastest kind of CT scanner?
EBCT
Prospective gating in cardiac CT
ECG triggers the x-ray tube such that projection data are acquired only during the end-diastolic period of the cardiac cycle - this is the point in the cycle where the heart is the most stationary
How can gantry motion be minimized in CT?
Focal spot rastering
Factors that influence CT resolution
Focal spot size/distribution, detector dimensions, the mag factor, whether or not gantry motion is compensated for, patient motion, recon algorithm/mathematical filters
Modern CT Scan Time Example
For a 48-cm abdomen-pelvis CT scan, with a nominal beam width of 40 mm, a pitch of 1.0, and a gantry rotation time of 0.5 s, the scan can be completed in 6 s, well within a breath-hold for most patients
Explain why overbeaming is necessary on an MDCT
For a scanner w/ a pitch near 1, eg, essentially all detector arrays contribute to the raw data acquisition for each image. Because of this, the beam profile for all detectors needs to be approx. the same, otherwise there will be artifacts. Therefore the beam penumbra must be located outside of the active detector array. Because single-detector CD slice width is determined by collimation, this is not necessary
What is the material used in CT detectors?
Gd2O2S
Ramp filter equation
H(f) = a x f where a is just a scaling factor
Standard adult head/torso sizes
Head: 17 cm Torso: 24-45+ cm, typically between 26-30 cm
Retrospective gating in cardiac CT
Heart imaged continuously while ECG data is recorder. Because the heart is longer in the z-dimension than what many MDCT scanners can image (nT), the procedure for imaging the heart requires data acquisition at several table positions, so the entire length of the heart is imaged. As long as the cardiac cycle is not in complete synchrony with the scanner's gantry rotation speed, after several cardiac cycles, enough data would be acquired over different phases to reconstruct the heart (using ECG data)
Describe the process of sintering, used in CT detector crystals
Heating up the crystals to just below their melting point for relatively long periods of time; after other processes occur, densification occurs and the scintillating powder is converted into a high-density ceramic which is then scored with a saw/laser to create individual detector elements in a module (eg 64x64). An opaque filler is pressed into the space between detector elements to reduce optical cross talk between detectors. The fabrication includes a photodiode in contact with the ceramic detector
Why do bone/iodine have greater mass attenuation coefficients in the CT energy range?
Higher Z means more photoelectric interactions = higher mass attenuation coefficient, as well as higher density = higher linear attenuation coefficients
The estimated biologic risk of developing cancer for an individual who received radiation as a fetus during an abdominal CT is about ____ cases per million exams of pregnant women
If the CTDIw is 3.0 cGy for a body scan, the surface CTDI would be around 3.6 cGy and the center would be about 1.8 cGy. Multiply center by increased risk for infants (3-4x greater than adults, or 15% per Sv). Answer: ~2700 potential cases
Why is Fourier-based reconstruction preferred?
It is easy to implement in a computer
Why is retrospective gating unfavorable in cardiac CT?
It's a high dose procedure since the CT beam is on for multiple entire cycles. Most of the interest in cardiac CT imaging is in identifying blockages in the coronary arteries (greatest risk to pt.). For evaluation of coronary stenoses, only the end-diastole images are necessary
Why is rapid-kV switching DECT more challenging than dual-source?
It's more difficult to switch the tube current (mA) with high temp. freq. since the mA is is determined by x-ray tube filament temperature, which cannot be modulated at kilohertz rates due to the thermal inertia of the filament. However, optimization studies of DECT show low (80) kV needs higher output levels than the high (140 kV) setting. However at the same mA, the tube is far more efficient at x-ray productions at higher voltages. The output air kerma at the isocenter at the same mA is about 3.3x higher at the highest kV and the lowest. To overcome this, single tube CT scanners increase the temporal pulse width at the lower kV setting to increase the relative signal levels
If the (minimum) CT slice thickness is quoted on a scanner, the actual detector width is _______
Larger by magnification factor M = B/A. That is, for a 0.5 mm min slice thickness, and B = 95 and A = 50, the actual detector width is 0.95 mm. That is why most dimensions in CT are related to the plane of the isocenter
In CT scans, increased kVp w/ the same mAs tends to produce...
Less noise (higher output)
How can CT contrast be increased?
Lower kVp, greater mAs (less noise), thicker slices, smaller matrices, smoother filtering?
Magnification factor at the CT scanner's isocenter
M - B/A where B is the SID and A is the source to isocenter distance
What is a "dynamic focal spot" in CT?
Magnetic e- beam steering around the anode, leading to oversampling in the z-dimension for helical scans, which leads to Nyquist-appropriate sampling in z
How does noise correlation in the z-direction occur in CT?
Multiple detector arrays (including cross-talk), combined with helical and cone beam reconstruction
How do you correct for a function with a 1/f dependency in convolution?
Multiplying it by a function with an f dependency
Types of CT detectors
NaI plus PMTs, Xenon ionization, CSI plus photodiodes, direct charge sensors
Why do bowtie filters improve patient dose without compromising image quality?
Noise is dominated by the central x-rays, and the extra dose from the peripheries is wasted. Reducing the dose to the periphery improves patient dose with no loss in image quality
CT Kernel Noise Power Spectrum Example Image
Noise power spectrum measures not only the magnitude but also the spatial correlation of noise properties ("texture")
How do different CT manufacturers regulate dose?
One lets you set the overall noise level. Lower noise = higher dose. Another lets you set a nominal mA value. Higher mA = higher dose.
How is mA modulation implemented?
One manufacturer uses the attenuation of mA data measured by the detectors near the central ray in the scanner to adjust the target mA for 180 deg later in the scan. Manufacturers also use attenuation data measured during the acq. of the CT radiograph to plan the mA modulation scheme for each patient.
The spatial resolution for a large-FoV Scan, such as abdomen, depends on...
Pixel size (FoVs 25-30 cm and greater)
CT Detector Element Electronics
Provide power to the detector array and receive electronic signals from each photodiode. Has gain channels for each detector in module and also contains analog-to-digital converter, which coverts the amplified signal to a digital number. Detector array modules are mounted onto an aluminum frame on the mechanical gantry assembly. The frame forms an approx. radius of curvature from the position of the x-ray source and each module can be removed and exchanged on the gantry for rapid repair
CT Perfusion Imaging Method
Scan starts prior to the arrival of contrast bolus to the organ. This is necessary b/c the mathematical models such as Pulak that are sued to compute the parametric maps require an input fnct. (HU vs time), typically the artery supply blood to the organ. From this data set of temporal CT scans, and w/ physician demarcation of the location of the input fnct., parametric maps are compute w/ high res. In addition to the raw gray scale CT images, most commercial software compute a number of diff. parametric 3D volume data sets of pseudocolored images, showing for example blood volume, tissue perfusion, and time to peak enhancement.
CT Detector Composition
Scintillator, such as GOS, LUMEX. Rare earth oxide powder is "baked" into scintillating batches. Batches are cut into "wafer-like" modules by laser beam
How is "noise" in CT computed?
(Root Mean Square method in homogeneous material)
CT effective mAs
(mA*rotation time)/pitch. For a fixed eff.mAs, pitch (+), rotation time (-), mA (++)
Typical HU values for tissue parenchyma
+30 to +220
Definition of HU for air
-1000 (u for air is bout 0)
Typical HU values for adipose tissues
-80 to -30
Typical HU value for urine
0
Definition of HU for water
0 (u = uw)
The detector array sampling time in effect becomes the acquisition interval for each CT projection that is acquired, and the sampling dwell times typically run between ______ ms, meaning that between _______ projections are acquired per 360 degree rotation for a 0.5s gantry rotation period
0.2 and 0.5, 1000 and 3000
Modern CT scanners can visualize contrast differences as low as ___%
0.5%. CT noise is about 5-10 HU. Much better than 7-8% contrast limit on screen-film.
Typical spatial resolutions for CT
0.6 or 0.9 lp/mm = 9 lp/cm
Typical CT pitch range and uses
0.75-1.5. High pitch used for thoracic/pediatric where speed is important. Low pitch used for cardiac, or to maximize dose for a large patient
How can DECT images be reconstructed?
1) Ram dual energy image data sets for each proj. angle can be subjected to logarithmic weighted subtraction techniques, similar to proj. radiography or 2) low and high kV data sets can be reconstructed separately and subtraction techniques or other digital image manipulation can be applied directly to the reconstructed CT images (most widely used) Note that HU values in CT are already linear w.r.t. the linear attenuation coefficient, so no additional logs are necessary. Ultimately the dual-energy image is a linear combination of low and high energy CT images
Typical CT focal spot size
1-2 mm
Given a 40-mm CT detector width in z, a 0.5-s gantry rotation time, and 80 mm/s table movement, what is the pitch?
1. nT= 40, F = 40 mm/rotation
The maximum FOV of a CT is defined by ______
The physical extent of the curved detector arrays (the fan angle)
What is the primary cause of noise correlation in an image?
The point spread function of the imaging system - the PSF has tails that extend from one voxel to others. Noise in the central region of the PSF is likely to propagate to surrounding areas where the PSF extends
CT Adaptive Beam Collimation
The process of removing wasted beam with collimation when overscanning for helical image acquisition
Why might you prefer to scan a 20 cm section of abdomen with 40 5-mm slices rather than 320 0.625-mm slices on a MDCT?
The thinner slices are 8 times thinner, meaning the noise is increased by a factor of sqrt(8) = 2.8
Why does a CT scanner with no bowtie filter tend to distribute dose more on the periphery of the patient?
The x-rays are attenuated less on the periphery of the patient, leading to high exit doses
Why is the plane of the anode disk parallel to the plane of gantry rotation?
To reduce gyroscopic effects that would add significant torque to the rotating anode if mounted otherwise. Also, the anode-cathode axis (heel effect) runs parallel to the z-axis of the scatter, eliminating heel effect-induced spectral changes along the typically wide fan angle. The fan beam is much more limited in the anode-cathode angle
View aliasing artifact
Too few projection images acquired to reconstruct high-frequency objects
What basic mathematical step does CT preprocessing accomplish?
Transform logarithmic attenuation into linear projections
In CT, the resolution in the z-direction can often exceed the resolution in the xy plane. True/False
True
Dual-source CT cardiac imaging
True CT reconstruction requires only 180 degrees plus the fan angle (which is small for cardiac). Full 360-deg rotation is 0.33 s so a full 180-deg rotation is only half as long (166 ms). With 2 redundant x-ray systems (1 w/ smaller FOV) configured approx 90 deg w/ respect to e/o, a quarter rotation of the gantry results in a full 180-deg acq. set when the data from both imaging chains are combined (only 83 ms)
True/False: You can always reconstruct a CT image to whatever slice thickness you want on a top of the line CT scanner
True. The data is always there because all channels are always used. For example, a 64-slice scanner with 0.5 mm slices (at isocenter) could reconstruct to 64 x 0.5 mm slices, 32 x 1.0 mm, 16 x 2.0 mm etc. However, this is not true of older scanners
How much filtration is often used in CT and why?
Up to 10 mm of Al, to pre-harden beam and reduce beam hardening artifacts
Sometimes in prospective-gated cardiac CT, if the pt.'s heartbeat is irregular during the scan, the acquisition may lead to an incomplete CT data set, which may not be recoverable. What is an alternative?
Use high mA during the end-diastolic period but low mA during the rest of the cycle, so you can reconstruct the entire cycle if necessary
CT Perfusion
Used to evaluate vascular perfusion and other physiological parameters related to blood flow to a specific organ. Most often used to evaluate stroke or vasospasm int he brain, but also used for abdominal organ imaging, especially in the oncologic setting. A high radiation dose study that acquires images of an organ repeatedly in real time to quantify the flow of iodine-based contrast agent through that organ. Uses wide coll. beam width (20-160 mm) on 64+ data channel scanners
Why is it harder to create a Noise Power Spectrum for CT reconstructed iteratively?
While filtered backprojection can be approximated as a linear system, iterative reconstruction may not be
Benefits/trade-offs of CT imaging with a wide cone angle (~10 deg, 16 cm Z-axis coverage)
Whole organ imaging can be achieved without table motion, such as CT angiography, perfusion studies (head, kidneys, heart, etc). There is also increased x-ray scatter and cone beam artifacts
Does increasing CT slice thickness increase contrast resolution?
Yes
The effective dose for CT scanners is greatest for: (cervical spinal, routine head, high-res chest, abdomen/pelvis, knee)
abdomen/pelvis
How do HU values change with energy spectrum?
air and water stay the same, other values may shift, especially bone/iodine
Most modern CT scanners use a (continuous/pulsed) x-ray source
continuous
The spatial resolution for a small-FoV CT scan, such as head or targeted spine, depends on...
focal spot size, geometry, detector size, recon. algorithm (NOT pixel size)
What bowtie filter is typically used for pediatric CT imaging?
head
All modern multidetector array CT scanners use ____ solid state detectors
indirect (scintillating)
Detector dimensions in CT are generally referred to their size as projected to the _______
isocenter
Typical settings adjusted in a protocol
kV, mA, rotation time, slice thickness, pitch, reconstruction kernel(s) used, timing/volume/injection rate of contrast injection scheme (if used), default window/level settings of recon. images, what computer to send the images to
In older CT scanners, with 4-32 detector arrays, the number of data acquisition channels is more than/less than/the same as the number of detector arrays
less than. For example, a 16-detector scanner may have only 4 data channels. In this setting, the desired slice thickness must be selected prior to the scan. For a 1.25-mm detector width, the user could acquire 4 x 1.25 mm for a total collimated beam width of 5.0 mm. The signals from adjacent detector arrays can be combined, and so by "binning" together signals from two adjacent arrays, the scanner could acquire 4 x 2.5 mm data (10.0-mm collimation) or by binning 3 and 4 adjacent detector signals, you can get 4 x 3.75 (15-mm collimation) or 4 x 5.0 (20-mm collimation).
Effective photons per slice
mA + rotation time + pitch
For a 64-slice CT scanner, in which the detector thickness is T = 0.625 mm and the tube rotation time is 1 s, how many seconds would it take to scan a 20-cm section of abdomen? How many images would you acquire?
n = 64, so the collimated beam width is nT = 40 mm. Therefore, you could scan 200 mm in 5 s, acquiring 200/0.625 = 320 images
The low-contrast discrimination of CT scanners is affected by...
noise
Definition of CT Pitch
pitch = F/nT, where F is the table feed distance per 360 degree rotation of the gantry and nT is the nominal collimated beam width
Sinogram
plot of raw ct data showing rays horizontally and angles vertically
Spatial res in CT results from the fundamental res properties of the image acquisition, as well as the _____ of the reconstruction filter
resolution characteristics
What is the largest source of noise correlation in the xy plane in CT?
the filtered backprojection kernel, due to the "smearing"
Slice sensitivity profile
the shape of the system response to a point input in the z-dimension; the LSF along the z-dimension in the CT scanner
Compton Scatter Coefficient Proportionality
where rho is the mass density of the tissue in a voxel, N is Avogadro's number, Z is the atomic number, and A is the atomic mass.
Definition of Houndsfield unit:
where u is the average linear attn. coefficient for a voxel at location xyz, and uw is the linear attn. coefficient for water for the spectrum used
CT FOV equation
with fan angle alpha, and source to isocenter distance S. Typical FOV is 50 cm, alpha is about 60 degrees, and S is about 50 cm
As the scan setting of a CT scatter is increased from 120-140 kVp, the CS is changed by about ____%
~-5%
Typical geometric efficiency of a modern MDCT
~95%. Wide beams mean the penumbra is a smaller part of the overall beam
CT Tube Arcing Artifacts
• Tube replacement or seasoning is needed • Software correction by interpolation
