CT physics Flashcards

Question

List the risk of radiation induced cancer per dose in an adult, adult \>50yrs and child.

Answer 1

**Adult**: 5% per Sv **Adult \>50yrs**: 0.5% per Sv **Child**: up to 15% per Sv

Answer 2

1. It's recommended that you reduce the mAs. 2. Reduced CT techniques are possible b/c XR penetration is greater in children. 3. Dose reduction in peds head CTs are more modest than peds belly CTs.

Answer 3

1. Do the scan at reduced mA (but the problem is that the images are terrible. 2. Use a mA modulation (adjust based on density), which is the preferred method. 3. Shield the breasts with bismuth: this will decrease dose by ~30% but will give artifact and a degraded image (beam hardening can falsely elevate HUs directly deep to the shield).

Answer 4

* It can increase HU of tissues underneath it. * Bismuth shields the pt but also prevents radn from exiting the pt, causing beam hardening, which elevates HU directly deep to the shield.

Answer 5

\<1mSV Because they don't contain any radiosensitive organs/tissue.

Answer 6

\>10% of the annual dose limit (500mrem).

Answer 7

Both, body and head CTDI_x will increase.

Answer 8

* Scan length (that would be DLP). * CTDI_vol = dose per slice as estimated with a 16 or 32cm phantom measured over 100mm length, taking into account mAs, kVp, pitch, and difference in dose at the periphery and center of the phantom. * Equation: CTDI_vol= [(1/3 x CTDI100_center + 2/3 x CTDI100_periphery) /pitch]

Answer 9

1. mA adjustment depending on the attenuation of the pt. The mA is varied as the tube rotates around the pt and along the long axis to provide best compromise b/w image quality and dose. New scanners can modulate during the rotation and along the Z-axis to decrease dose to radiosensitive organs. 2. kVP: this may be adjusted depending on body habitus, but it is not referred to as dose modulation.

Answer 10

1. The LAC describes the fraction of photons that are attenuated per unit thickness of material. It inlcudes all possible interactions, including those removed directly by absorption by the PE effect or by Compton scatter. 2. The LAC decreases w/increasing photon energy except at k-edges. 3. As Z (atomic # increases), LAC increases as more photons are absorbed.

Answer 11

1. To remove low/high energy XRs that would only serve to increase dose. 2. a) compensate for uneven filtration; b) reduce scatter; c) reduce dose. 3. 6mm copper, aluminum, Teflon (any low Z material). 4. The periphery; they are more shelled out in the middle, just like a bowtie. 5. Perpendicular to the imaging plane to reduce heel effect.

Answer 12

1. At the XR tube and at the detector. 2. To shape the XR beam (you know this as in fluoro the collimators are adjustable diaphragms); to reduce some scatter. 3. The section thickness on a single slice.

Answer 13

* Image data is sequentially collected from one object. * Fancy math takes all that noisy data and removes the noise/sharpens the image. * Then it back projects the sharpened image. * This is being replaced by iterative reconstruction, which requires higher powered computing, that allows more noise and less dose.

Answer 14

* Sharpening of the projection data prior to back projection. * Blurring is removed during a deconvolution process. * Without filtering, the images would be blurred and unsuitable for diagnostic imaging. * This filtering does not remove artifacts, e.g., beam hardening. * The application of different reconstruction kernels will determine how much sharpening occurs during the deconvolution process, resulting in less sharp (soft tissue kernel), or sharper (bone kernel) images.

Answer 15

* It goes through repeated updates of the reconstructed data by comparing the forward data projections from each iteration to the actual measured projections to adjust the reconstruction. * With each iteration, the difference b/w the estimated and real projections (the error) gets smaller. * This improves the SNR.

Answer 16

The spatial resolution in the transaxial (X-Y) plane is the same as that in the Z-plane. That is, the resolution is the same in all dimensions. This is important as it allows non-axial reconstructions (coronal/sag) without stretching pixels. It also helps to avoid artifacts: stair-step, and creates well-defined images.

Answer 17

Detector element aperture width.

Answer 18

1. Scout: they use the scout data to estimate density & auto-adjust. 2. On the fly: as the scan is happening they use continuous modulation to adjust.

Answer 19

**Ray**: a single line of data of XR attenuation from the source to a single detector. **Projection**: all the rays at a given angle of the XR tube, aka all the rays that pass through the patient at the same orientation. **Sinogram**: a collection of squiggly lines that represent the data from all of the projections of the tube angles.

Answer 20

1. Highly filtered, high kV (average energy 75 keV). 2. You increase the noise/get grainier images. (Think low-dose CTs which are super grainy.)

Answer 21

1. 512 x 512 pixels = 262,144 pixels total. 2. 4,096 (12 bit, 2¹² = 4.096). 3. Pixel size = FOV / matrix size. 4. Make the pixels smaller, so decrease FOV or increase matrix size.

Answer 22

**distance of (table movement during a single tube revolution)** --------------------------------------------------------------------------- **beam width**

Answer 23

1. Water = 0. 2. HU = 1000 x (attenuation of material - attenuation of water) / (attenuation of water) 3. When HU increases by 10 HU, XRs are attenuated by 1% (which is to say that the material stopping the XRs is denser). 4. 0.5% HU.

Answer 24

1. Since the PE dominates at lower energy, low keV will create higher HU. The closer you approach the k-edge of a given substance, the more impressive the increase in attenuation (and higher HU). Example: IV contrast will have a HU of ~100HU at 140kV but will increase markedly to 400 HU at 80kV, where the average energy will land flush on its k-edge. 2. In beam hardening, the beam is filtered to increase the quality of the beam, i.e., increasing the kV. So if you remove lower energy photons to harden the beam, that will increase your average energy and lower your HUs. This is related to "cupping artifact".

Answer 25

Air = -1,000 Water = 0 Bone = high, but variable, i.e., cortical bone can be 1,200-1,500 while the otic capsule can be \>2,000. 2. 1,000 HU

Answer 26

1. Window = the range of HUs that will be displayed, so the larger the window, the larger the range of HU #s that will be displayed. 2. Level: midpoint of the window grayscale that you've chosen. You want this to be at the level of the attenuation of the thing that you're interested in.

Answer 27

**Axial pros:** stationary table; tube takes a picture, table moves up a slice, another full picture is taken. * Better spatial resolution in Z-dimension since full image sets are taken per slice. * No partial volume effect along the Z-axis (these are noticeable in helical CT especially along curved surfaces, e.g., skull). **Helical pros:** tube is on the whole time & table moves at a constant speed. * WAY faster. * Lower probability of anatomic discontinuities b/w adjacent slices containing moving anatomy.

Answer 28

1. **Dose**: increase. 2. **Image noise**: decrease. 3. **Conspicuity of iodinated contrast**: decrease. Iodinated contrast will be more conspicuous at lower kVP (e.g., 80 kVP), as the average energy of the XR beam will be closer to the k-edge of iodine.

Answer 29

1. **Scan time**: reduced; larger coverage w/1 turn. 2. **Motion artifact**: reduced; less scan time. 3. **Partial volume**: increased; more divergent beam. * **_Partial volume defn_**: occurs when tissues of wildly different absorption (HUs) are captured in the same voxel, which shows a beam attenuation that is the average value of the 2 different tissues. 4. **Dose**: no change; mAs unchanged, and even though the scan time is less, a larger area of tissue is scanned at the same time.

Answer 30

1. It can correct for noise, so lower doses may be used. 2. Noise & streak artifacts.

Answer 31

* B/c the acquired data from a 360° revolution does not lie on one plane.

Answer 32

1. During diastole. 2. Prospective, retrospective. **Prospective**: step & shoot during the R-R interval * Pro: less dose b/c the scanner isn't on the whole time. * Con: No functional imaging. **Retrospective:** scans the whole time then back-calculates. * Pro: can do functional imaging. * Con: more dose b/c on the whole time w/low pitch. 4. Prospective is _always axial_, never helical.

Answer 33

1. Mottle = noise. 2. It is the random fluctuation in the number of photons reaching the detector at any given point. 3. **Poisson**: the events occur independently of the previous events & the probabilities don't change over time. The average # of photons reaching individual pixels on the detector = N. 4. In low-dose imaging, e.g., XR, fluoro.

Answer 34

1. XR flux, i.e., mA; 1:1 proportion, so if mA doubles, so does signal. 2. sqrt(N), so if mA doubles, noise increases by sqrt(2) = 1.4. 3. 2/1.4 = 40%. 4. 6 factors: * Higher mA * Higher kVp * Longer rotation time * Smaller pitch (pt moves through the gantry slower) * Larger slice thickness * Larger pixel

Answer 35

1. The ability to discriminate b/w structures w/similar attenuation characteristics, i.e., pixel values/colours. 2. The number of colours generated, i.e., pixel values. So CT and MR are excellent at generating colours, but XR is terrible. 3. CT is excellent. 4. B/c there is minimal scatter/noise reaching the detectors. 1) Very tight collimation (pre-patient & pre-detector); 2) windowing--a tight window maximizes contrast resolution. 5. Any factor that reduces noise: mAs, kVP, longer rotation time, decreased pitch, larger slice thickness, larger pixels.

Answer 36

1. Best: XR & fluoro; worst: CT & MR. 2. 1. **Focal spot size**: smaller = better SR; a larger focal spot means the details are spread out over several detectors which degrades/blurs the image. Focal spot determines SR in the X-Y plane, so side to side. 2. **Magnification**: more mag blurs. 3. **Detector aperture size**: like focal spot size, smaller = better; this determines SR in the Z-plane; as the detector size is reduced, the z-axis increases; the x-y axis is not affected by aperture size. 4. **# of projections**: more projections, more data, better SR. 5. **Reconstruction slice thickness**: the thinner the element aperture, the better the spatial resolution in the Z-direction.

Answer 37

1. DFOV = the space defined by the user based on the anatomy size to be displayed. 2. pixel size = DFOV / matrix size. 3. As FOV decreases, so does pixel size, which increases SR but decreases contrast resolution. 4. As FOV increases, so does pixel size, which decreases SR but increases contrast resolution.

Answer 38

1. As pitch increases, the width of the slice sensitivity profile (SSP) increases, so the slice thickness increases, which increases contrast but decreases SR. 2. Pt motion blurs, which decreases SR. 3. Sharp = better SR, but more noise. 4. Smooth = worse SR, but less noise.

Answer 39

**Beam hardening: cupping.** * As the XR beam passes through an object the lower energy photons are removed preferentially, leaving a harder beam w/an increased average energy. * The XRs passing through the middle of a uniform shape (like a head) are hardened more than those at the periphery. * The harder the beam, the slower the rate of attenuation, so the center of the image appears darker. * **Fixes:** * Modern CT scanners use filters to overcome beam hardening so it's less often seen. * **Filtration**: pre-hardening of the beam to remove lower energy photons and/or bowtie filter. * **Calibration correction**: w/a phantom. * **Software**: use iterative correction algorithms. * **Gantry tilt**: tilt the gantry or position the pt to avoid areas that can cause hardening.

Answer 40

**Beam hardening: streak/dark bands.** * As the XR beam passes through an object the lower energy photons are removed preferentially, leaving a harder beam w/an increased average energy. * Occurs in the setting of 2 dense objects. XRs that pass through one are less attenuated than those that pass through both. The result is dark bands or streaks b/w them. * Classic location: bone, or when dense contrast is used. * **Fixes:** * Modern CT scanners use filters to overcome beam hardening so it's less often seen. * **Filtration**: pre-hardening of the beam to remove lower energy photons and/or bowtie filter. * **Calibration correction**: w/a phantom. * **Software**: use iterative correction algorithms. * **Gantry tilt**: tilt the gantry or position the pt to avoid areas that can cause hardening.

Answer 41

**Partial volume averaging:** * When 2 tissues of very different HUs/absorption inhabit the same voxel, resulting in a beam attenuation/HU that is the average value of the 2 tissues. * **Classic location**: CT angiography looking for PEs, where the vessel looks like it has a contrast filling defect. * **Fixes:** * Use thin section reconstructions to decrease voxel size.

Answer 42

**Photon starvation:** * A type of streak artifact seen in high attenuation areas, particularly behind metal implants. * B/c of the high attenuation, insufficient photons reach the detector & during reconstruction, the noise is magnified leading to characteristic streaks. * **Classic location**: shoulders or prosthesis site. * **Fixes:** * Automatic tube current modulation. * Adaptive filtration to correct the attenuation profile & smooth the data. * Iterative reconstruction techniques.

Answer 43

**Ring artifact:** * A CT hardware artifact due to miscalibration or failure of one or more CT detector elements in a **3rd generation scanner--these only occur in 3rd generation scanners!** * It causes error in angular position. * **Classic location**: close to the isocenter of the scan & usually visible on multiple slices at the same location. Common in cranial CTs. * **Fixes:** * Recalibrate the scanner. * Replace the faulty detector element(s). * These cannot be corrected by the technologist at the scanner.

Answer 44

**Undersampling/aliasing:** * This is a physics-based artifact. * When sampling the scanned data, the computer undersamples which causes the computer to process an inaccurate image, resulting in aliasing or misregistration artifacts. * **Fixes:** * Acquire the largest number of projections per rotation--slow the rotation speed. * Use manufacturer-employed high resolution techniques.

Answer 45

**Stairstep artifact:** * Found in straight structures that are obliquely oriented w/respect to the table motion (z-motion). * B/c of wide collimation of non-overlapping intervals. * **Classic location**: on sagittal/coronal reformatted images, often 3Ds. * **Fixes:** * Use thinner slices. * Less severe on a helical scanner where you can decrease the pitch to get overlap.

Answer 46

1. Motion. 2. Stabilize the pt; use a fast scanner; align the scanner in the direction of motion; decrease pitch (over-scan); use gating. 3. Over-scanning = decreasing pitch to create overlap; over-ranging: scanning above & below the target to collect additional data to add to the helical scan.

Answer 47

1. High Z metals have more metallic artifact, e.g., iron, platinum. Lower Z have less, e.g., titanium. 2. Ask the pt to remove the metal; increase the kVp; use thinner slices.

Answer 48

1. When body parts hang outside the field but attenuate XRs which messes w/the computer's math. 2. W/arms beside the pt or IV contrast on/near the pt.

Answer 49

To decrease helical artifact in the axial plane: * This often occurs near the top of the skull and is worse w/higher pitch scans. * To minimize this, reduce variation in the z-direction, so decrease pitch, use thin sections.

Answer 50

* Compton scatter. * At energies used in routine CT, most of the scattered XR photos are due to Compton scatter, and a minority from Rayleigh scatter or characteristic radiation from the PE. * Scatter in the pt is not reduced by post-pt collimation. * Scattered XR photons from the primary beam contribute significantly to the overall pt dose.

Answer 51

* With all other parameters being the same, pt dose increases linearly w/scan length. * Recall, scan length is the anatomy covered in a particular scan. * DLP = CTDI_volx (scan length)

Answer 52

* It's the average dose to a slice in the central portion of an entire scan, and includes the dose from scanning that slice as well as the dose to that slice by scatter from scanning adjacent slices.

Answer 53

* Only the voxel with the highest CT # along each ray, in every view, is displayed. * That is, for each XY coordinate, only the pixel with the highest HU along the Z-axis is displayed. * MIP images show attenuation information, not depth information. * Bone and contrast-containing structures.

Answer 54

* It's the scanner-specific radiation output in this exam. * DLP = CTDIvol x scan length, so it represents scanner output x scan coverage.

Answer 55

* Iodine * Urate: gout * Bone * Water

CT physics Flashcards

(82 cards)