Transcript CT - 8 Lecture Notes Page

▪
▪
▪
History
Equipment
Image Production/Manipulation
Dateline
1895 - Roetgen discovers x-rays
▪ 1917 - Radon develops recontruction formulas
▪ 1963 - Cormack develops mathematics for xray absoprtion in tissue
▪ 1972 - Housfield demonstrates CT
▪
▪
▪
▪
▪
▪
1975 - first whole body CT
1979 - Housfield and Cormack win Nobel prize
1983 - EBCT
1989 - spiral CT
1991 - multi-slice CT
▪
Original idea was to move the patient not the beam.
▸The intent was to produce a homogeneous or monoenergetic
beam.
▪
Original scanner used a radioisotope instead of a tube.
To date there have been four accepted
generations with some consideration as EBCT
to be the fourth.
▪ The first fourth generation scanner was
unveiled in 1978 four years after the first
scanner.
▪
Translate/rotate
Pencil thin beam - highly
collimated
▪ Single radiation detector
▪ 180 translations at 1 degree
of rotation
▪ One image projection per
translation
▪ 5 minutes of scan time per
image
▪ Heads only
▪
Translate/rotate
Fan shaped beam
▪ Multiple detectors - a
detector array
▪ 18 translations with 10
degrees between them.
▪ Multiple image
projections per
translation
▪ 30 second scan time
per image
▪ Head and body imager
▪
Rotate/rotate
Fan beam that covers the
entire width of the patient
▪ Several hundred detectors
in a curvilinear detector array
▪ Both the source and the
detector array move
▪ Hundreds of projections are
obtained during each rotation,
thereby producing better
spatial and contrast
resolution.
▪ Scan time is reduced to one
second or less per image
▪
Rotate/stationary
Still a fan beam
▪ Thousands of detectors
are now used
▪ Thousands of
projections are acquired
producing better image
quality
▪ Sub-second scan times
▪ Various arcs of scanning
are possible increasing
functionality
▪
Intended for rapid
imaging
▪ Scan time less than 100
msec
▪ No tube, instead
tungsten rings are used
▪ Four rings allow four
slices to be acquired
simultaneously
▪ No moving parts
▪
Third or fourth generation scanners with
constant patient movement
▪ Use slip ring technology
▪ Can cover a lot of anatomy in a short period of
time
▪
first
spiral
<1 s
scan time 300 s
80x80 1024x1024
matrix
slice th
13 mm
spatial res 3 lp/cm
1 mm
15 lp/cm
CT image circa 1971
▪
▪
▪
▪
X-ray source
Detector array
Collimator
High voltage generator
10,000 rpm anodes
▪ 8 MHU
▪ Tube is parallel the patient to reduce anode
heel effect
▪ 200 - 800 mA
▪
Bow tie filters are used to ‘even out’ the beam
intensity at the detectors
▪ Primary purpose is to harden the beam
▪
▸Reduces artifacts
CT uses a high kVp to minimize
photoelectric effect
▪ High kVp allows the maximum number of
photons to get to the dectector array
▪ All current scanners use high frequency
generators
▪
▸High frequency generators are much smaller than
three phase units allowing for a smaller footprint
and less voltage fluctuation
Early scanners used scintillation crystal
photomultiplier detectors as a single element
▪ Currently two types of detector arrays
▪
▸Gas filled
▸Solid state
▪
▪
▪
▪
Filled with high pressure xenon
Fast response time with no afterglow or lag
50% dectection efficiency
Can be tightly packed
▸Less interspacing, fewer lost photons
Ion chambers are approximately 1 mm wide
▪ Geometric efficiency is 90% for the entire
array
▪ Total detector efficiency = geometric efficiency
x intrinsic efficiency
▪
▪
Cadmium tungstate
▸Scintillator
▪
▪
▪
Material is optically coupled with a photodiode
Nearly 100 % efficiency
Due to design they cannot be tightly packed
▪
▪
▪
▪
▪
80 % total detector efficiency
Automatically recalibrate
Reduced noise
Reduced patient dose
More expensive than gas filled
Located between the detector array and the computer
▪
▪
▪
Amplifies the signal
Converts the analog signal to digital(ADC)
Transmits the signal to the computer
Multiple detector arrays allow for multiple
slices to be acquired simultaneously
▪
▸Pre-patient
▪ Controls patient dose
▪ Determines dose profile
▸Post-patient
▸Controls slice thickness
▪
▪
▪
▪
Most common process is filtered back projection
Fourier transformation
Analytic
Iterative
▪
▪
Data acquisition
Preprocessing
▸Reformatting and convolution
▪
▪
▪
Image reconstruction
Image display
Post-processing activities
Suppress low spatial frequencies resulting in images
with high spatial resolution
▪
▸Bone
▸Inner ear
▸High-res chest
▪
▪
▪
Suppress high spatial frequencies
Most commonly used filters
Images appear smoother
▸Less noisy
▪
▪
Images are displayed on a matrix
Today most are 512 x 512 or 1024 x 1024
▸The original matrix was 80 x 80
▪
▪
The matrix consists of pixels
Pixels represent voxels
The diameter of the reconstructed image is
the FoV
▪
Generally, pixel size is the limiting factor in
spatial resolution.
▪ The smaller the pixel the higher the spatial
resolution.
▪ Pixel size (spatial resolution) is determined by
matrix size and FoV.
▪
Post-processing does not increase the amount
of information available. It presents the original
information in a different format
▪
This is numerical value assigned to each pixel.
▪ CT numbers are derived from the attenuation
coefficient of the tissue in the voxel.
▪ CT numbers are also called Hounsfield units
▪
tissue
bone
muscle
white matter
gray matter
blood
CSF
water
fat
lung
air
CT number
1000
50
45
40
20
15
0
-100
-200
-1000
Att Coeff
0.46
0.231
0.187
0.184
0.182
0.181
0.18
0.162
0.094
o.0003
▪
▪
▪
Atomic number
Tissue density
Beam energy
Attenuation
▪
▪
I=Ioe-µx
Based on a homogenous beam
▪
The higher the CT number the brighter the pixel
▪
▪
▪
Calculation
Positive and Negative
Numbers for various anatomical structures
▪
Water is 0.206
µT - µi
µI
X
1000
▪
▪
▪
▪
▪
▪
▪
▪
▪
▪
Air = -1000
Lungs = -200
Fat = -50 to – 100
Water = 0
CSF = 15
Blood = 42-50
Gray matter = 40
White matter = 45
Muscle = 50
Bone = >500
▪
▪
This is the range of CT numbers displayed.
The wider the width the lower the contrast.
▸Think scale of contrast, a long scale (wide width)
has low contrast.
Level is the center number of the width.
▪ Usually, this represents the anatomy of
interest.
▪ You can see by the similarities between CT
numbers that the level doesn’t change much.
▪
▪
▪
▪
Increase pixels increase resolution
Decrease voxel size increase resolution
Typically need to increase technique with higher res
The most common is maximum intensity
projection (MIP)
▪ Also, volume rendering is used to provide an
image with depth. Used to be called shadedsurface display (SSD).
▪ Quantitative CT uses a phantom to establish a
bone mineral density exam.
▪
This is the basis for CT angiography.
▪ Voxels are selected for their intensity along a
proscribed axis of reconstruction.
▪ MIP images are volume rendered
▪
▪
▪
ROI
Measurement
▸Linear
▸Volume
▪
Magnification
Spiral scanners greatly improved sagittal and
coronal reconstructions because they limited
movement.
▪ Multi-slice scanners are even better because
they have smaller slice thicknesses and
isotropic voxels.
▪
Conventional CT
Axial image
Spiral
▪
▪
▪
Source moves, detectors probably not
Source stops and starts
Patients moves between exposures
▪
▪
Source moves, detectors may move
Patient moves during exposure
Couch movement per rotation divided by
slice thickness
▪ Contigous spiral: pitch = 1, 10mm of
movement with a slice thickness of 10mm
▪ Extended spiral: pitch = 2, 20mm of
movement with a slice thickness of 10mm.
▪ Overlapping spiral: pitch = ½
▪
The lower the pitch the better the z-axis
resolution.
▪ The narrower the collimation the better the zaxis resolution.
▪ Increase pitch, decrease dose
▪ When pitch exceeds 1, interpolation filters
must be applied
▪
Spiral scanners don’t acquire true axial
images so interpolation becomes necessary at
larger pitches.
▪ So data is interpolated and then back filtered.
▪
Image noise is higher for spiral CT than
conventional CT regardless of the scanning
parameters.
▪
▪
▪
▪
▪
▪
▪
Faster image acquisition
Contrast can be followed better
Reduced patient dose at pitches > 1
Physiologic imaging
Improved 3d and reconstructions
Less partial volume
▪
▪
▪
▪
Fewer motion artifacts
No misregistration
Increased throughput
Real time biopsy