Physics Flashcards
cathode
negative charge (electrons flow from)
anode
positive charge (electrons move to)
multiphase generators
reduces ripple effect of AC current (overlaps several waves)
- makes peaks = kVp (peaks)
Bremsstrahlung
- increases with accelerating voltage (kV) and anode atomic number (Z)
characteristic radiation
incoming e- ejects inner shell e-
- must exceed binding energy (only K shell is important)
- 5-10% of total usually
all targets produce characteristic radiation just below 20 keV, why?
k-edge filtering
how can you get more characteristic radiation
lower Z anode (Mo in mammo)
- get 25% characteristic (max contrast = mono-energetic spectrum)
x ray filter
stops low energy photons = reduces dose
inherent filtration
glass exit window on the tube
added filtration
preferential filtering at lower energies
- mean energy shits UP (shifts to the right) = increasing QUALITY, decrease QUANTITY (decrease dose)
K-edge filtering
filter material just about the k-edge of the target to kill higher energy keV photons
- makes spectrum narrower
- emphasize characteristic peaks
- both improve contrast
photoelectric absorption
totally absorbed
Compton scattering
lose part of energy and change direction
Photoelectric interaction
- incoming photon hits tight inner shell
- photon is totally absorbed (photoelectron ejected)
- outer shell electron fills vacancy = emits characteristic “fluorescent” x ray
kedge depends on material, what is the k edge of Iodine
33 keV
when do you get the best image contrast
when the beam spectrum matches the energy-dependent interaction inside the person
probability of photoelectric interactions
increased with increased Z
decreased with increased energy
compton scatter
photon hits loosely bound outer shell e-
- scatters in a new direction with less energy
- proportional to electron density
- falls off at higher energy
attenuation
Sum of all interactions in the patient
- went in and didn’t come out = photoelectric effect
- went in and didn’t get detected due to change in direction and energy = compton scatter
increase in energy does what to total attenuation
decrease in PE with about the same compton = decrease total attenuation
photoelectric effect dominates at what energies
lower energy
compton dominates at what energies
higher energy
linear attenuation coefficient (u)
fraction of incident photons lost from beam per distance unit
- Beer-Lambert law
half value layer
thickness of material that cuts the beam intensity in half
- increased energy -> increased HVL (need more stuff to stop beam)
- increased density of material -> decreased HVL (for same energy, need less stuff to stop it)
what is the useful equation for HVL
HVL = 0.7 / u
what is the clinical significance of HVL
shows if there is enough filtering
lower HVL = decrease in quality/energy, increased attenuation, increased dose (bad stuff)
mass attenuation coefficient
linear attenuation coefficent divided by density (u / p)
- independent of phase of matter of density
beam hardening
tissue attenuation acts like a filter (removes lower energy photons)
x ray exposure = mAs, what if you increase mAs
increased mAs = more photons, same energy distribution -> mean energy stays the same
- direct proportion to increase dose = double mAs = 2 x photons and 2 x the dose
varying energy (kVp), what happens if you increase kVp
more photons and HIGHER engery
- 50/15 rule = increase in kVp by 15% = 2x photons -> should cut mAs by 50% to get same exposure (in 60-90 kVp range)
what is meant by softer beam
decreased kVp
- decreased penetration
- increased attenuation in body tissues
- INCREASED CONTRAST (small dynamic range used in mammo due to all soft tissue imaging)
what is meant by harder beam
increased kVp
- increased penetration
- decreased attenuation
- DECREASED CONTRAST
- larger dynamic range (CXR - different tissues)
Typical kVp of different exams from lowest to highest
Mammo - 20
Extremity - 40
head, spine, hip, angiography - 70
Abdomen - 80
Barium - 115
CXR - 120
if you decrease energy (kVP) what happens
- decreased photons = AEC will increase mAs
- increased contrast
- increased dose (from increase in mAs)
lower kVp for peds reduces, dose, how?
decreased flux of x rays produced but partially limit the mA compensation (partial AEC)
- decreased dose but increased image noise
does iterative recon affect dose
no
iterative recon does what to an image
decrease noise
what does scatter do to contrast
degrades contrast
how to reduce scatter
- grid
- collimation
- air gap
Grid use
scatter is not aligned = removes a lot of scatter
- some primary is filtered as well
grid ratio
higher the grid ratio the higher the scatter rejection
Grid ratio = height of strips / distance between strips (GR = h /D)
what is bucky factor
Bucky factor = ratio of input to output flux (usually 2-6)
- need to increase exposure that much to get same exposure to the detector -> increases DOSE
Adding a grid does what if dose does not change
- decreased scatter
- INCREASED noise due to decreased through transmission
Collimation to reduce scatter
decrease FOV = DECREASED scatter, INCREASED contrast
Air Gap to reduce scatter
- scatter has a greater angle = doesn’t hit detector
- scatter has a shorter distance to travel and falls off more quickly than primary (inverse square law)
- Greater source to image distance and magnification
how much of the electron energy is converted to x rays vs heat
99% heat and 1% x rays
heat capacity of tube determined by what?
- focal spot
- anode
- housing
what is the equation for heat units
HU = kVp x mA x secs
1 HU = 0.71 J
Line focus principle
bigger area for heat dissipation
determined by anode angle theta
if you increase anode angle theta what happens
- increased heat dissipation
- increased effective focal spot
- increases FOV
- decreased sharpness
Heel effect
- decreased amount of higher energy photons on the anode side
- worse for bigger FOV or shorter source to image distance
due to heel effect which side of the tube will you place the thickest (most dense) part of the body
cathode side
Grid artifacts
- incorrect focal distance
- off center grid
- tilted grid
- inverted grid
effect of increased kVp on contrast
increased kVp = increased penetration, decreased contrast, decreased dose
Resolution due to frequency
thick about sound (violin vs bass)
- higher spatial frequency for smaller size
- lower spatial frequency for larger, thicker structures
Resolution is characterized by what function
Modulation transfer function (MTF)
- range of 0-1 (1 is perfect)
- in real life lower frequency is better preserved
resolution limits in line pairs
one line pair / mm is equal to two pixels (line pair = 1 bright and 1 dark = 2 pixels)
- 5 lp / mm = 10 pixels / mm = pixel size of 0.100 mm
Limiting resolution
Limiting resolution = frequency at 10% MTF
what are the usual limiting resolutions for different exams from high to low resolution
Mammo - 5-10 lp/mm
Radiography - 3 lp/mm
CT - 1 lp/mm
Resolution is affected by what
- motion
- focal spot size and magnification
- detector resolution and sampling
Motion unsharpness
- patient motion or tube motion
how do you decrease motion unsharpness
- shorter exposure times
- immobilize patients
how to decrease focal spot blur
use a small focal spot
what are the typical focal spots for typical radiography and for mammo
- typical - 1.2 or 0.6 mm
- mammo - 0.3 or 0.1 mm
why can you not use the smallest focal spot all the time
- reduced tube output -> longer exposure time -> more motion blur
- more heat = shorter tube life
Focal spot blur
M = SID / SOD
- Penumbra blur is directly proportional to the focal spot size and the mag
- bigger the mage or focal spot = more blur
Geometric unsharpness
small FS and small mag = less blur
big FS and big Mag = greatly increased blur
small FS and big mag = standard blur
- mag makes everything bigger including focal spot blur (use small FS to get back to standard)
Detector resolution
- detector type, material, thickness
- pixel pitch
- binning
Overall resolution is determined by each individual MTF of the system, how do you improve resolution
must know each individual MTF and address the worst one
multiply each one together to get the overall MTF of the system
Noise power spectrum (NPS)
spatial frequency content of image noise
- decrease dose = MORE noise
- increase dose = LESS noise
what causes noise
mostly = finite number of photons used in the image (quantum mottle)
Limited absorption efficiency of the x ray by the detector
Quantum mottle
Poisson distribution (variance = mean)
- relative noise = 1 / sq rt N
Signal to noise ratio
resolution in the numerator and noise is denominator (R/N)
SNR of > 5 is good
what happens to SNR if there more noise
both size and contrast of detectible lesions go down
what happens to SNR if there is more blur
size of detectible lesions goes down
Detective quantum efficiency (DQE)
dose is proportional to SNR^2
- better DQE = better machine
- DQE of 2x = same image quality for less dose (SNR same for 1/2 dose)
- higher image quality for same dose (1.4 x SNR of same dose)
what is the range of DQE
0-1
- higher = better to see small lower contrast at a lower dose
Digital imaging machines
- improved dynamic range
- post processing better quality
- PACS
Indirect flat panel
phosphor -> converts x ray to light -> CCD or photodiode coverts light to signal
- decreased signal
- increased noise
- increased blur from light spread
BUT, has higher DQE (increased absorption efficiency of materials)
Direct flat panel
photoconductor coverts x rays to signal directly
Computed radiography
storage phosphor hold x rays in latent storage -> laser stimulates light emission -> photomultiplier tube coverts light to signal
- both indirect and direct are better than CR
MTF of direct, indirect, and CR
Direct > Indirect > CR
DQE of direct, indirect and CR
Indirect > Direct > CR
Pre processing
- correct for detector defects
- convert x ray to pixel values
- display consistently with prior
Post processing
- grayscale processing
- edge enhancement
- equalization
Luminance
brightness, cs/m^2
- high >/= 350, 420 for mammo
- low = 1 or 1.2 for mammo
great luinance = increased contrast
GSDF
grayscale display function - more consistent throughout the hospital
CR twin
double exposed plates (two images on top of each other)
Delayed readout of CR
decreased density in the periphery where the fading starts
Backscatter exposure
poor collimation or long exposure = backscatter from the cassette (CR) or detector electronics (DR)
CR roller artifact
dust on the roller = horizontal bright lines that can be traced outside the patient
Grid moire pattern
interference between the grid ratio and display matrix
- can show up on monitor with to low resolution
DR lag (ghosting)
residual image of lead markers from previous image
DR incomplete grid suppression software
incomplete grid line suppression by post processing
DR detector saturation
overexposure in low/un-attenuated regions (like lung)
- can’t recover the data
Fluoro tube changes
heat capacity
- high speed anode rotation
- water or oil heat sink
Focal spot
- small focal spot for fluoro
- large focal spot during spot or cine (greater tube current)
fluoro collimation
mulitple sets of shutters
beam shaping
- decreased object glare
- decreased scatter
- decreased dose
Image intensifier
x ray - latight - electons - light - electronic signal
flat panel detector
x ray - light - electronic signal
II flux gain
e- gain kinetic energy when travelling across the high voltage
- increases light by a factor of 50-100
Minification gain
BIG input, small output (affects only brightness)
Brightness gain of II
BG = flux gain x minification gain
Frame averaging
sacrifices temporal resolution to reduce noise
- decrease dose with increased lag
Automatic brightness control
controls mA and kV automatically
magnification for II
projects only central part of the input layer onto output phosphor
- decrease field of view = increase geometric mag (increased resolution)
- less minification gain = increased exposure rate by ratio of FOV (reduce the noise)
- mostly increases kVp to minimize increase in dose (keeps air kerma and tube heat lower)
- less contrast and increase noise
digital spot mag mode
mostly increase mA
does flat panel mag increase dose?
yes, just not as much as II
Sentinel event of fluoro
anything over 15 Gy
Dose area product
DAP = emitted dose x field size (in Gy per cm^2)
- NOT patient or skin dose
- can estimate effective dose (whole body)
air kerma
total accumulated dose during procedure (mGy)
air kerma rate
air kerma rate per minute
dose area product
air kerma x beam area
peak skin dose
dose to highest area of skin (usually less than air kerma)
how to minimize dose
- keep detector close to the patient
minimize dose in peds
- remove grid
- use low dose (higher kV and lower mA)
occupational exposure comes from what
scatter from the patient (compton)
- stand on detector side
reduce exposure
- less time
- distance
- shielding
Pincushion distortion (II)
curved input phosphor to flat output phosphor
- reduced with mag use
s-distortion
external magnetic field = spatial warping
Vignetting (II)
darkening at the edges (flat panel does not get this)
in mammo on which side do you place the cathode
chest wall side (thickest portion) - heel effect
focal spot for mammo
0.3 mm for contact and 0.1 mm for mag
normal compression range for mammo
110-180 newtons
why compress breast
- decrease scatter
- decrease geometric blur
- decrease tissue overlap
- decrease dose
Grid ratio for mammo
4-5 only
- 6-16 to general radiography
because of low energy in mammo what has to be true of filtration
filtration must be low
- low Z material for tube port (beryllium, Z = 4)
K-edge filtering in mammo
use same element for target and filter = narrow spectrum = more contrast and less dose
what is the highest resolution of any imaging
mammo
- smaller focal spot
- smaller pixels
- magnification
- compression
Mag view changes from contact view
- small focus (0.1 mm) with breast closer to tube and further from detector via mag stand
- 1.5-2x mag factor
- no grid, decreased scatter due to air gap
- similar resolution due to small focus
- can use spot compression to reduce overlap
electronic mag
takes same picture and just makes the pixel bigger
geometric mag
real mag views
- better visual of calcs etc.
TOMO breast
limited angle cone beam CT
- incomplete data -> non-isotropic resolution = never sees breast from side = worse z-direction resolution
TOMO artifacts - emboss
embossing (intra-plane ringing of high contrast along the tube travel direction (think about biopsy clips)
TOMO artifact - shadow
out of plane smearing of objects (think about calcs)
what is the HVL of mammo
HVL usually 0.3 mm of Al
Mean gladular dose (for average breast)
<300 mrad (3 mGy)
- dense and thicker than 4.2 cm dose would be higher
Scoring of ACR phantom
must have:
- 4 / 6 fibers
- 3 / 5 specks
- 3 / 5 masses
Mammo ghosting
shadow from previous image
- worse on selenium
- aggravated by high dose / contrast
- aggravated by low temperature
incorrect auto technique
impants (fails to pick right kVp/mAs = underexposure = increased noise
Loss of skin line
breast periphery not fully compressed
- loss of skin can be due to post processing
Damage to filter
wrinkle or dent in filter
Grid artifact in mammo
cross hatch
detector calibration for mammo
detector line interface
TOMO truncation artifact
some breast are outside the beam path
- terrace or venetian blind artifact due to incomplete information
Most common problems for mammo
- bad compression
- motion
- bad positioning
power rating for CT power supply is determined by what
focal spot size
large focal spot size for CT is what
1-2 mm (high power rating ~65 kW)
small focal spot size for CT is what
0.5-0.6 mm (low power rating ~25kW)
what are the two collimators in CT
Pre-patient (source collimator)
- decreased dose
Post patient collimator
- decreases scatter from patient to detector
CT filters
Al or Cu
Bowtie filter for CT (usually made of Al, teflon, or graphite)
Contours beam
- more uniform beam
- decrease beam hardening artifacts
Post patient collimator on single detector machine
determines slice thickness
post patient collimator on MDCT
determines collimation width and minimum slice thickness
types of CT detectors
Gas and solid state
Gas = xenon ionization chamber (60-80% efficency)
solid state = bismuth germinate or cadmium tungstate
(scintillator -> light -> photodetector -> electric signal) = ~ 100% efficent
benefit of solid state over gas CT detector
- higher absroption effiency
- higher SNR
- less beam hardening
disadvantage of solid state over gas CT detector
less stable
difference between single and multiple detector CT
single = single detector = single slice
multi = multiple detectors = multiple slices per rotation
Axial CT
takes pictures then moves table then takes picture
Helical CT
tube and table move at the same time
what allows the tube to spin
slip ring between power and x ray tube
when might you run in axial mode
if > 128 slice detector to reduce cone beam artifacts
- look at dynamic structures that are moving (like heart)
Minimum slice thickness
MST = collimation width / number of channels
- can reconstruct images greater than, but not less than this number
difference bewteen non-isotropic voxel vs isotropic voxel
non isotropic = rectangle
isotropic = cubic
how is x and y axis determined
X and Y = DFOV (mm) / matrix size (always 512)
how is the Z axis determined
z axis = slice thickness
Line pairs
remember - take two voxels to define a line pair
same equation as the x and y axis multiplied by 2
Line pairs = (DFOV (mm) / matrix size) x 2
types of reconstruction
filtered back projection
-assumes single beam CT = not good for low radiation dose
iterative reconstruction
how do you get partial volume averaging
each voxel is assigned an attenuation coefficient (u)
- if a beam straddles two different densities you get the mean (average) of the two
- kVp dependant
what is iterative reconstruction
may reduce noise by 30-70%
reconstruction longer
types of reconstruction filters
Standard = balanced detail and noise
smooth = low detail and low noise
bone = high detail and high noise
what is the memory needed for CT images
2 bytes / voxel
in 512 x 512 matrix = 262,144 voxels = 0.5 MB per slice
average for abdomen pelvis = 75 MB
Hounsfield Unit
directly related to the linear attenuation coefficient
1 HU = 0.1% difference between the linear coefficient of tissue compared to water
Window width
number of shades of grey
- inversely proportional to image contrast
window level
center shade of grey
- adjusted to match tissue you want to look at
formula for window/level
level +/- Window / 2
typical window width and level for different tissues
ST = 400 / 40
Liver = 150 / 70
lung = 1000 / -600
Bone = 1000 / 600
Pitch
pitch = Table travel per rotation / collimation width
- pitch of 1 = no overlap (< 1 = overlap, > 1 = gaps)
relationship between helical pitch and dose
higher the pitch = lower dose
what is slice broadening
slice gets thicker with higher table pitch (ex. pitch of 1 = 20% broadening)
what is the relation of gantry rotation and dose
time for gantry to move 360 degrees
- linear to dose
- shorter the gantry time = lower the dose
radiation dose when changing kVp
exponential change
radiation dose when changing mAs
linear change
Noise is figured by…
sq rt of # of photons
Motion CT artifact
- parallel lines
Undersampling on CT
aka aliasing
- photon deficiency = streak artifacts
- white rings on part of body outside the detector FOV
Ring artifact for CT
detector malfunction or miscalibration
- looks like drop of water in a pond
Beam hardening artifacts CT
due to wide range of photon energies
- low energy absorption
Stair step artifact of CT
off axis reformations for thick slices (axial slices > 1.25 mm)
Signal to noise ratio
more photons = less noise
- SNR = # photons / voxel
SNR is affected by what
kVp higher = less noise
mA higher = less noise
Pitch higher (faster) = more noise
gantry rotation time faster = more noise
DFOV smaller = less photons (increased spatial resolution but more noise)
Radiation dose in gray vs sieverts
GrAy = Absrobed dose
SiEverts = Effective dose
CTDIvol (CT dose index for one slice)
- phantom based
- includes both direct and scatter radiation
- reflects technique, NOT total dose (mGy)
Dose length product (DLP)
basically the CTDIvol x Anatomic length
- this DOES reflect total dose in mGy cm
Dose reduction for CT
Lower kVp = expoential
Lower mA = linear
higher pitch
shorter GRT
thicker slices
Avoid small DFOV
Iterative reconstruction
Average speed of sound in the body
1540 m/sec (soft tissue)
- fat and air are slower
- muscle and bone are faster
how many micro seconds does it take per 1 cm of tissue for a sound wave to go there and back
13 micro seconds
what are the 4 ways that sound can interact with the body
- reflection (straight back or same angle)
- refraction (bent)
- scattered
- absorption (heat)
Reflection
if acoustic impedance is very different = lots of reflection
if they are similar = near zero reflection (transmitted)
Refraction
differences in speed of sound between two tissues = bent sound
Scatter
is like reflection but due to very small reflectors
- specular reflector = flat, perp to the beam -> large amount bounce back
- nonspecular reflection = scatters at angles that may not be picked up
- higher frequency = more scatter
absorption
beam turned to heat
- higher frequency = more absorption
Attenuation
all of the reflection, refraction, scatter and absorption
- higher freq = more attenuation = hard to get deep
Signal loss in dB
Signal loss = 0.5 dB/MHz/cm x distance (cm) x Frequncy (MHz)
what is the half value thickness of US
thickness of material that causes 3 dB decrease in signal
- air, high (goes short)
- water, low (goes further)
attenuation leads to through transmission
no attenuation in the fluid (brighter posterior compared to surrounding ST)
Curvilinear
good for abdomen
Linear
high spatial resolution (small parts)
Piezoelectric material
changes shape with voltage
- PZT (lead-zirconate-titanate) most common crystal
Best lateral resolution
at focal spot at end of near field
Length of near field increases with what
diameter and higher frequency
Damping degree
light damping = higher Q (narrow bandwidth) = good for doppler
heavy damping = low Q (shorter SPL) = better axial spatial resolution
Phased array
smaller number of elements that sweep at different times to steer the beam (smaller footprint)
how to make different focal zones
time difference between activating crystals on the inside vs periphery (shapes the beam)
how large is an US image
0.25 MB per image
Transmit power vs gain
transmit power = amount of signal (power of signal from radio station)
gain = how much amplification of the signal (changing the volume on the radio)
pulse repetition frequency
how often 3-5 wavelets are sent out (~ 1 kHz)
Spatial pulse length
relates to how small of physical distances can be distinguished
- higher frequency transducer -> better spatial resolution
- SPL = the number of little wavelets (usually 3)
Axial resolution
dependent on SPL (0.5 x SPL)
- NOT dependent on depth
Lateral and elevation resolution
more dependent on depth and follow shape of the beam
- lateral and elevational best in the focal zone
(elevational in plane of thickness)
put the focal zone where on a lesion?
center it on the lesion
time gain compensation
due to attenuation signal from deeper is lower -> time gain compensation turns up gain to try and correct this
B mode
brightness mode (signal more = bright, low signal = darker)
- gray scale
M mode
movement of structures plotted on time
A mode
amplitude mode (graph of signal)
Spatial compounding
compound imagines from different angles (decreased artifacts) = see around things
Doppler angle at 90 degrees
lowest signal and most error
what is the good doppler angle
60 degrees
Continuous doppler
no depth info
- no aliasing
Pulsed doppler
depth selection
aliasing possible
Used in association with 2D B mode (duplex doppler)
Aliasing
wrap around on the spectral tracing due to undersampling of the frequency shift
- increase scale (increases PRF as well)
how to stop aliasing
increase PRF (PRF/2 is the smallest that can be measured)
- lower freq transducer
- use higher angle theta
- use power doppler (no direction sense)
- might have to decrease depth
Mirror image artifact
due to a strongly reflective interface due to high difference in acoustic impedance values
how to stop mirror image artifact
different acoustic window or angle
Refraction artifact
misplaced anatomic position due to direction change in the beam at interfaces
how to stop refraction artifact
change the angle or spatial compounding
Speed displacement artifact
one path has material with different speed of sound than in adj tissues
- apparent discontinuity of the diaphragm
how to stop speed displacement artifact
change the angle
Reverberation artifact
parallel lines = parallel reflectors
- like white lines in the bladder
how to stop reverberation artifact
change angle
change window
tissue pressure
harmonic imaging
Comet tail artifact
reflectors close together and also lose signal (due to attenuation)
Ring down artifact
create vibration, commonly seen with air
Beam width artifact
looks like sludge in the GB,
extending adj tissue into a flluid structure
how to stop beam width artifact
change focal zone
different angle or window
harmonics
Shadowing
beam attenuation along one path
how to stop shadowing
compound imaging
different window
Larmor Frequency
42.6 MHz/Tesla = RF sent into the patient
where are MRI voxels
in the Gradients
Field strength of MRI per tesla
1 Tesla = 10^4 gauss
Flip angle
how much the proton orients into the transverse plane
Free induction decay
protons go out of sequence and lose signal due to field inhomogeneity
- rate depends on T2*
Acquisition MRI scan time
Time = TR x # of phase encoding steps x NEX x # of slices
- NEX = # of k spaces averaged together
what sequences are black blood
Spine echo sequences
- blood getting the 180 degree pulse didn’t get the initial 90 degree pulse
what gradient is changed with every repetition on spin echo imaging
phase-encoding gradient
what is the primary determinant of noise in MRI
receiver bandwidth
what sequences are bright blood on MRI
gradient echo sequences
