Chụp cắt lớp vi tính hay chụp cắt lớp CT

Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Basic Principles - Conventional Radiograph

Computed Tomography – Chapter 13

Kalpana Kanal, Ph.D., DABR
Lecturer, Diagnostic Physics
Dept. of Radiology
UW Medicine
a copy of this lecture may be found at:
/>c.f. Bushberg, et al. The Essential Physics of Medical
Imaging, 2nd ed., p. 328.
Kanal

2

Basic data acquisition in CT

Historical Development of CT

X-ray Tube
X-ray Beam

CT Table

Detectors

Figure from Dr. Mahesh, John
Hopkins, MD, AAPM Handout.

¬

First Generation CT :
Rotate/Translate, Pencil Beam

¬

In 1972, the EMI scanner was
the first CT scanner introduced
into clinical practice

¬

5 minutes to generate a pair of
images

c.f. /> />
c.f. Wolbarst, The Physics of Radiology,
Radiology, 2nd Edition, 2005
Kanal

Kalpana M. Kanal, Ph.D.

3

Kanal

4

1


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Historical Development of CT

Historical Development of CT
¬

¬

¬

Second Generation CT:
Rotate/Translate, Narrow Fan
Beam

¬

Third Generation CT:
Rotate/Rotate, Wide Fan
Beam
0.5 to 2 seconds to acquire an
image

1 minute to generate a single
image

c.f. Bushberg, et al. The Essential Physics of Medical
Imaging, 2nd ed., p. 328.

c.f. Wolbarst, The Physics of Radiology,
Radiology, 2nd ed., p. 409.
Kanal

5

6

Historical Development of CT

¬

Fourth Generation CT:
Rotate/Stationary

¬

0.5 to 2 seconds to acquire an
image

Historical Development of CT

¬

c.f. Wolbarst, The Physics of Radiology,
Radiology, 2nd ed., p. 410.
Kanal

Kalpana M. Kanal, Ph.D.

c.f. Wolbarst, The Physics of Radiology,
Radiology, 2nd ed., p. 409.

Kanal

7

Kanal

Fifth Generation CT: Stationary/Stationary
X Electron beam scanner
X Primarily for cardiologists, 50 msec scan times
X Uses tungsten target and highhigh-energy electron beam
c.f. Bushberg, et al. The Essential Physics of Medical
Imaging, 2nd ed., p. 337.
8

2


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course

10 March – 7 April 2005

Formation of a CT image - Tomographic Acquisition

Historical Development of CT
¬

Sixth Generation CT: Helical
X SlipSlip-ring technology
developed (allows gantry to
rotate continuously without
wires)
X Helical CT scanners
acquires data while table is
moving
c.f. Seeram, Computed Tomography,
Tomography, 2nd Ed., pg. 79.

¬

Seventh Generation CT
X Multiple Detector Array
¬

A single transmission measurement through the patient made by a
single detector at a given moment in time is called a ray

c.f. www.impactscan.org
Kanal

9

Formation of a CT image - Tomographic Acquisition

¬
¬

Kalpana M. Kanal, Ph.D.

c.f. www.sprawls.org, computed tomography lecture

10

Formation of a CT image - Tomographic Acquisition

A series of rays that pass through the patient at the same orientation
orientation
is called a projection or view
All modern CT scanners incorporate fan beam geometry

Kanal

c.f. Macovski, Medical Imaging Systems,
Systems, pg. 114.

Kanal

¬

11

Kanal

You could have approx. 800 rays taken at 1,000 different projection
angles giving 800,000 transmission measurements
c.f. www.sprawls.org, computed tomography lecture

12

3


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Formation of a CT image – Reconstruction
Formation of a CT image - Tomographic Acquisition

¬

¬

¬

Before the axial acquisition of the next slice, the table that the
the
patient is lying on is moved slightly in the cranialcranial-caudal direction or
the zz-axis of the scanner

This positions a different slice of tissue in the path of the xx-ray
beam for the acquisition of the next image
High kV of 120 to 140, mA ranging from 10 – 440 and scan times of
0.4 – 2 seconds
c.f. www.sprawls.org, computed tomography lecture

Kanal

13

c.f. Seeram, Computed Tomography,
Tomography, 2nd Ed., pg.107.

Kanal

Tomographic Reconstruction
Preprocessing & Raw Data
¬

¬
¬
¬
¬
¬
¬

¬

Tomographic Reconstruction

Each ray is a transmission measurement through the object along a
line, where the detector measures an xx-ray intensity, It
I0 = unattenuated intensity of xx-ray beam
It = I0 e-µt
t = thickness of patient along the ray
µ = the average linear attenuation coefficient along the ray
ln (I0 / It) = µt for each ray
This is the preprocessing step performed before image
reconstruction
Image primarily depends on the patient’s anatomic characteristics
characteristics

Kanal

Kalpana M. Kanal, Ph.D.

14

¬

¬
¬

¬

¬

15

Kanal

After preprocessing the raw data, a CT reconstruction algorithm is
used to produce the CT image (attenuation coefficient map)
Filtered back projection is most widely used in clinical CT scanners
scanners
The backprojection method builds up the CT image in the computer
by essentially reversing the acquisition steps
During backprojection, the µ value for each ray is in essence
smeared along the same path in the image of the patient
Areas of high attenuation reinforce each other and areas of low
attenuation reinforce each other building up the image in the
computer

16

4


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Formation of a CT image – Filtered back projection

¬

¬

¬

Image Reconstruction

However, simple
backprojection produces an
image that is somewhat
blurred
Raw data must first be filtered
using a mathematical filter, or
kernel
This process is known as the
convolution technique

c.f. Bushberg, et al. The Essential Physics of Medical
Imaging, 2nd ed., p. 352.
Kanal

17

Kanal

Smooth versus Sharp Filter

¬

Bone algorithm - fine detail but
with increased noise

¬

Soft tissue filters - smoothing,
which decreases image noise
but also decreases spatial
resolution

¬

The choice of the best filter to
use with the reconstruction
algorithm depends on the
clinical task

c.f. Seeram, Computed Tomography,
Tomography, 2nd Ed., pg.108.

CT Image
¬

¬

¬

¬

Kalpana M. Kanal, Ph.D.

A pixel (picture element) is the
basic 2D element of the digital
image
Each pixel displays brightness

information concerning the
patient’s anatomy that is
located in the corresponding
voxel (volume element)
The pixel width and height are
equal to the voxel width and
height
The voxel has a third
dimension that represents the
slice thickness of the CT scan

c.f.
/>
c.f. />Kanal

18

19

Kanal

20

5


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

What Is Being Measured?

CT Image
¬

¬

¬

¬

¬
¬

¬

The rows and columns
comprise a matrix
Matrix sizes are 512 x 512,
1024 x 1024 etc
The technologist selects the
field of view (FOV).
Pixel size = FOV/matrix size
The gray scale range for each
pixel is 1212-bits (0(0-4095)
Spatial resolution improves
with a larger matrix (smaller
pixels) or smaller FOV

¬

¬

¬

c.f.
/>
Kanal

21

22

What do CT numbers correspond to physically in
the patient?
¬

The CT(x,y) number in each pixel, (x,y) of the image is derived from:

⎡ µ ( x, y ) − µ water ⎤
CT ( x, y ) = 1000 ⎢
⎥
µ water
⎣
⎦
¬

¬

¬

µ(x, y) is the attenuation coefficient for the voxel, µwater is the
attenuation coefficient of water and CT (x,y) is the CT number (or
(or
Hounsfield unit) that comprises the final clinical CT image
Air = -1000, soft tissues range from -300 (lung) to -90 (fat), water = 0,
white matter = 30, gray matter = 40, muscle = 50, dense bone and
areas filled with high contrast agent range up to +3000

Kanal

Kalpana M. Kanal, Ph.D.

c.f. />
Kanal

CT Numbers or Hounsfield Units
¬

The CT reconstruction process
results in a 2D matrix of floating
point numbers in the computer
which range from near 0.0 up to
value equal to 1.0
These numbers correspond to the
average linear attenuation
coefficient of the tissue contained
in each voxel
The CT images are normalized

and truncated to integer values
that encompass 4096 values,
between -1000 and 3095
(typically)
CT numbers are rescaled linear
attenuation coefficients

23

Kanal

CT numbers and hence CT images derive their contrast mainly from
the physical properties of tissue that influence Compton scatter
¬ The linear attenuation coefficient tracks linearly with density of
tissue and plays the dominant role in forming contrast in medical
medical
CT
CT numbers are quantitative,
¬ pulmonary nodules that are calcified are typically benign, and
amount of calcification can be determined from the mean CT
number of the nodule
¬ CT is also quantitative in terms of linear dimensions and can
be used to accurately access tumor volume or lesion diameter

24

6


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

CT Timeline

Digital Image Display: Window/Level
¬

¬

¬

¬

¬

P1 = L – ½ W
P2 = L + ½ W

CT image voxels utilize a 1212-bit
graygray-scale (212=4096 shades)
Computer monitors and laser
imagers for printing have about 8
bits of display fidelity (28=256)
The 1212-bit CT images must be
reduced to 8 bits to
accommodate most image
display hardware
The window width (W)

determines the contrast of the
image, with narrower windows
resulting in greater contrast
The level (L) is the CT number at
the center of the window

64 slice
scanners
2005

Figure from Dr. Mahesh, John
Hopkins, MD, AAPM Handout.

c.f. Bushberg, et al. The Essential Physics of Medical
Imaging, 2nd ed., p. 359.

Kanal

25

Kanal

26

Technological Advances That Led To:
Helical (Spiral) Acquisition

Helical/Spiral CT

¬

Patient is transported
continuously through gantry
while data are acquired
continuously during several
360360-deg rotations

¬

SlipSlip-ring technology

¬

HighHigh-power xx-ray tubes

¬

Interpolation algorithms

c.f. Seeram, Computed Tomography,
Tomography, 2nd Ed., pg. 82.

c.f. Kalender WA, et.al. Radiology, 176(1):181176(1):181-3, 1990

Kanal

Kalpana M. Kanal, Ph.D.

27

Kanal

28

7


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Interpolation (Helical)
¬

¬

¬

¬

Helical CT scanning produces
a helical trajectory around the
patient
CT reconstruction algorithms
assume that the xx-ray source
has negotiated a circular, not a
helical path around the patient
To compensate for these
differences, the helical data set

is interpolated into a series of
planar image data sets
This allows the production of
additional overlapping images
with no additional dose to the
patient (interleaved
reconstruction)

Helical CT - Pitch

Pitch =

Table increment per rotation (mm)
Beam collimation (mm)

¬

Pitch is a parameter than comes to play when helical scan protocols
protocols
are used
Typical Pitch Ratio - 0.5, 1.0, 1.5, 2.0

¬

Pitch <1 implies overlapping and higher patient dose

¬

Pitch >1 implies extended imaging and reduced patient dose

¬

c.f. Seeram, Computed Tomography,
Tomography, 2nd Ed., pg. 218.

Kanal

29

Kanal

30

Pitch
Slice Thickness: Single Detector Array Scanners
Slice Sensitivity Profile
¬

¬

c.f. Bushberg, et al. The
Essential Physics of
Medical Imaging, 1st ed.,
p. 261.

¬

The slice thickness in single
slice CT is determined by the
physical collimation of the

incident xx-ray beam with two
lead jaws
The SSP describes how thick
a section is imaged and to
what extent details within the
section contribute to the signal
When a small object is placed
in the center of the CT slice
X

X

Kanal

Kalpana M. Kanal, Ph.D.

31

Kanal

it produces greater contrast
from background (greater
difference in CT number) than
when the same object is
positioned near the edge of
the slice volume

c.f. Bushberg, et al. The
Essential Physics of
Medical Imaging, 2nd ed., p.

343.
32

8


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005
Detectors and Detector Arrays

Slice Sensitivity Profile for Conventional and Helical CT
Most modern CT systems use
either Xenon detectors (old
technology) or solidsolid-state
scintillator detectors

¬

c.f. Bushberg, et al. The Essential
Physics of Medical Imaging, 2nd ed., p.
339 & 340.

Kanal

33

Kanal

34

Take Home Points
¬
¬
¬
¬
¬
¬
¬

Take Home Points
CT image voxels utilize a 1212-bit graygray-scale (4096 shades)
The window width (W) determines the contrast of the image, with
narrower windows resulting in greater contrast
¬ The level (L) is the CT number at the center of the window
¬

7 generations of CT scanner
Filtered back projection is most widely used in clinical CT scanners
scanners
A pixel (picture element) is the basic 2D element of the digital
image
The pixel width and height are equal to the voxel width and height.
height.
The voxel has a third dimension that represents the slice thickness
thickness
of the CT scan
Matrix sizes are 512 x 512, 1024 x 1024 etc.
Pixel size = FOV/matrix size

¬

Table increment per rotation (mm)
¬

⎡ µ ( x, y ) − µ water ⎤
CT ( x, y ) = 1000 ⎢
⎥
µ water
⎣
⎦

Kanal

Kalpana M. Kanal, Ph.D.

35

Kanal

Pitch =

Beam collimation (mm)

¬

Pitch <1 implies overlapping and higher patient dose

¬

Pitch >1 implies extended imaging and reduced patient dose

36

9


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Raphex 2001 Diagnostic Questions
¬

¬
¬
¬
¬
¬

¬

Raphex 2000 Diagnostic Questions

D35. The CT reconstruction kernel or algorithm chosen by the
operator affects:
A. Only pixel noise.
B. Only spatial resolution.

C. Only patient radiation dose.
D. Pixel noise and spatial resolution.
E. Pixel noise and spatial resolution and radiation dose.

D36. A small lung nodule (8 mm in diameter) is noticed in a lung CT.
It has a CT number of 60. If the slice thickness is reduced from 10
mm to 2 mm, the CT number of the nodule would probably:

¬

A. Increase.
B. Decrease.
C. Remain the same.

¬
¬

¬

D. The reconstruction algorithm affects both noise and resolution,
but has no affect on dose. A smooth algorithm decreases noise and
and
resolution. A sharp algorithm increases noise and resolution.

Kanal

¬

37

Kanal

38

Raphex 2003 Diagnostic Questions
¬

¬
¬
¬
¬
¬

¬

Raphex 2003 Diagnostic Questions

D45. If a CT bone window is set at a width of 1000 with the center at
500, the range of CT numbers that will be displayed as black is
_______.
A. Greater than 500
B. Less than 500
C. Less than -500
D. Less than 0
E. Less than 1000

¬

D. The center is at 500, and the width is 1000, i.e., 0 -1000. CT
numbers below 0 are outside the window, and are displayed as

black.

¬

Kanal

Kalpana M. Kanal, Ph.D.

A. If properly centered in the slice, the CT number will go up
because of less volume averaging with air in adjacent lung.

¬
¬
¬
¬
¬

39

Kanal

D43. The CT value of white matter is 40 HU, and that of gray matter
is 45 HU. The approximate subject contrast between white and gray
gray
matter is _______.
A. 0.12
B. 0.5
C. 1.2
D. 5.0
E. 12.0

B
For low CT numbers (-200 to 200) the percent contrast can be
approximated by: % contrast = (CT number difference)/10 = (4540)/10 = 0.5.

40

10


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Single Slice vs. Multislice CT

SDCT versus MDCT

c.f. Seeram.
Computed
Tomography, 2nd
ed., p. 258.

*Rydberg et. al., Radiographics 2000, 20:1787
Kanal

41

Kanal

Slice Thickness: Multiple Detector Array Scanners

42

Slice Thickness: Multiple Detector Array Scanners

¬

The slice thickness of multiple
detector array CT scanners is
determined by the width of the
detectors in the slice thickness
dimension

¬

It is typical to adjust the
collimation so that the
penumbra falls outside the
edge detectors

¬

The width of the detectors is
changed by binning different
number of individual elements
together – that is, the
electronic signals generated by
adjacent detector elements are
electronically summed

¬

This causes the radiation dose
to be higher, specially for small
slice widths in multislice
scanners

c.f. Bushberg, et al. The Essential Physics
of Medical Imaging, 2nd ed., p. 344.

Kanal

Kalpana M. Kanal, Ph.D.

c.f. Bushberg, et al. The Essential Physics
of Medical Imaging, 2nd ed., p. 344.

43

Kanal

44

11


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course

10 March – 7 April 2005

Data Acquisition System in Multislice CT

Multiple Detector Arrays

c.f. Bushberg, et al. The
Essential Physics of
Medical Imaging, 2nd ed., p.
341.

Kanal

45

c.f. Seeram.
Computed
Tomography,
2nd ed., p. 260.

Kanal

Multislice CT

Kanal

Kalpana M. Kanal, Ph.D.

46

Slice Width Selection

c.f. Seeram. Computed
Tomography, 2nd ed., p. 259.

c.f. />47

Kanal

48

12


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Isotropic Resolution

6 to 16 Systems
Fixed

Mixed

Improved Spatial
Resolution

Adaptive

Isotropic Resolution -All
sides of the voxel have
equal dimensions

c.f. Seeram. Computed Tomography, 2nd ed., p. 264.

c.f. />Kanal

49

Kanal

50

Flexible Image Reconstruction
Pitch in MultiMulti-slice CT

Kanal

Kalpana M. Kanal, Ph.D.

c.f. />
51

Kanal

¬

Two pitch definitions seen in MSCT

¬

Pitchx = table travel per rotation
X-ray beam width

¬

Pitchd = table travel per rotation
detector width

c.f. />
52

13


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

MSCT - Advantages

Pitch in MultiMulti-slice CT
¬

¬

Two pitch definitions seen in

MSCT

¬

Pitchx = 20 = 1 = Collimator pitch
20

¬

Pitchd = 20 = 4 = Detector pitch
5

¬

Collimator pitch = Detector pitch
N

c.f. />
Kanal

53

c.f. Seeram. Computed Tomography, 2nd ed., p. 263.

Kanal

¬

¬

Multiplanar reformatting (MPR) is a method for generating coronal,
coronal,
sagittal, or oblique images from the original axial image data

¬

The inin-plane pixel dimensions approximate the xx-y axis resolution,
but the slice thickness limits the zz-axis resolution (better with MSCT)

Kanal

Kalpana M. Kanal, Ph.D.

54

Multiplanar Reformatting
Oblique Reformatting

Multiplanar Reformatting

¬

MSCT speed can be used for fast imaging for larger volumes of tissue
tissue
with wide sections
¬ Same acquisition in shorter times (fewer motion artifacts)
¬ Thinner slices for better zz-axis resolution
¬ Efficient use of xx-ray beam, reduction of radiation dose (?)
¬ Reconstruction in different slice widths
¬ Possibility of Isotropic Imaging (better MPRs and 3D images with

reduced image artifacts)

55

Kanal

Oblique reformatting is quite similar to sagittal or coronal
reformatting, except that the CT voxels in the stack are sampled
along an axis that is tilted from either the x or y planes.
Several organ systems in the human body are not especially well
visualized with routine sagittal and coronal planes, and oblique
reformatting can be useful in these instances.
c.f. Cody. Radiographics, 22:125522:1255-1268, 2002

56

14


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Oblique Reformatting

Oblique Reformatting

(a) Oblique reformatted view was
obtained by using axial images

(reconstructed from a helical
acquisition) obtained at 2.5 mm
section thickness and a pitch of
1.5. A portion of a stent (large
arrow) and a tumor (small arrows)
are seen adjacent to the superior
mesenteric artery (arrowhead).

c.f. Cody. Radiographics, 22:125522:1255-1268, 2002

Kanal

(b) On another oblique reformatted
view obtained from 1.25 mm slice
thickness helical acquisition with
pitch = 1.5, the stent (large arrow)
and the tumor (small arrows) are
much better visualized, as is the
involvement of the superior
mesenteric artery (arrowhead).

57

¬

Advantages
¬ Enables visualization of specific structures such as the optic
nerves and lesions in relation to surrounding structures
¬ To determine the true extent of lesions or fractures and to help
localize lesions and intraarticular bone fragments or foreign

bodies

¬

In surface rendering, the
voxels located on the edge of a
structure are identified, usually
by intensity thresholding, and
these voxels are displayed

¬

The remaining voxels in the
structure are usually invisible

Disadvantages
¬ Image detail is not as good as in transaxial images
¬ The plane thickness affects image detail and thus thick planes
result in blurring and loss of structural detail

¬

This approach is useful for
examining tubular structures
(virtual bronchoscopy), the
colon (virtual colonoscopy) and
blood vessels

c.f. Cody. Radiographics, 22:125522:1255-1268, 2002
Kanal

Kalpana M. Kanal, Ph.D.

58

ThreeThree-Dimensional Image Display
Surface Rendering

Multiplanar Reformatting
Advantages & Disadvantages
¬

c.f. Cody. Radiographics, 22:125522:1255-1268, 2002

Kanal

59

Kanal

¬

¬

Left: a 2D sagittal reformatted view from a CT
colonography examination shows a polypoid
projection into the colonic lumen (red circle)
that did not extend through the colonic wall.
Right: a 3D endoluminal surface-rendered
image from the “fly-through” sequence

provides an enlarged view of the lesion, which
later proved to be an adenomatous polyp at
endoscopic biopsy.
60

15


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

ThreeThree-Dimensional Image Display
Surface Rendering

ThreeThree-Dimensional Image Display
Surface Rendering

c.f. Cody. Radiographics, 22:125522:1255-1268, 2002

¬

This thresholding process is critical and sometimes difficult to
reproduce
¬ If too aggressive, actual protruding structures can be lost from
view because of partial volume effects
¬

(a) SurfaceSurface-rendered view from a

multiplemultiple-row detector CT series
demonstrates an aneurysm (arrow)
arising at the origin of the posterior
inferior cerebellar artery (arrowhead)
from the vertebral artery (circle).

(b) On a view from the interior of the
artery, the left vertebral artery (L
(L Vert),
Vert),
posterior inferior cerebellar artery
(PICA),
PICA), and neck of the aneurysm are
all easily recognized, and their relative
locations are readily appreciated.

Kanal

61

Kanal

¬

Volume rendering has replaced most of the surface rendering
applications

¬

In volume rendering, the CT numbers that make up the image are

assigned to be either visible or invisible, to be displayed with varying
colors, and often to be displayed with varying opacity levels
(transparency)

¬

3D image display for vascular anatomy provides excellent anatomic
anatomic
information for surgical planning

Kalpana M. Kanal, Ph.D.

62

ThreeThree-Dimensional Image Display
Volume Rendering

ThreeThree-Dimensional Image Display
Volume Rendering

Kanal

If too lax, non tissue materials such as fluids can be rendered as
if they were tissue, which can obscure protruding structures

(a) Frontal volumevolume-rendered image of
a patient with horseshoe kidney
who was scheduled to undergo
surgery for a mass in the right
posteriorposterior-inferior aspect of the

fused kidney (arrowheads).
The aorta is readily apparent (arrow).
63

Kanal

c.f. Cody. Radiographics, 22:125522:1255-1268, 2002

(b) Another view is somewhat closer and more
superior than the view seen in (a), and the kidney
tissue has been rendered slightly more
transparent.
The unusual vasculature pattern in this case is
readily apparent, with two upper pole arteries
(open arrows) and two lower pole arteries
(arrowheads) originating below the inferior
mesenteric artery (solid arrow).
64

16


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

ThreeThree-Dimensional Image Display
Volume Rendering

(a) VolumeVolume-rendered image for which the color
map function was adjusted to help enhance
the visibility of the tumors (arrows) on the
cortex of the kidney.
The kidney is shown as orange, the renal lesions
are reddish, and the bones are white.
Observing this case in motion (rotating right to
left) helped the viewer appreciate the depth
and interconnections of the anatomic
structures.

¬

(b) Another frame of the movie loop at
a different angle as (a) from this case
demonstrates the renal artery
structure.

¬

¬
c.f. Cody. Radiographics, 22:125522:1255-1268, 2002

Kanal

Reformatting Techniques
Maximum and Minimum Intensity Projections

65

Kanal

Reformatting Techniques
Maximum and Minimum Intensity Projections

(a) Oblique MIP of a slab (cropped volume)
of the hepatic and pancreatic regions clearly
demonstrates the enhancing regional
vasculature and orally administered
gastrointestinal contrast material.
The image noticeably shows vascular
stenosis (arrows) caused by a surrounding
tumor, but the tumor margins are not well
seen.
Kanal

Kalpana M. Kanal, Ph.D.

c.f. Cody. Radiographics, 22:125522:1255-1268, 2002

66

Curved Reformatting

¬

(b) MinimumMinimum-intensity projection of
the same oblique view emphasizes
the hepatic biliary tree.
The tumor margins (arrows) are more

clearly delineated with this technique.
A slightly different slab view volume
was used.
c.f. Cody. Radiographics, 22:125522:1255-1268, 2002

With MIP, viewing rays are traced from the expected position of the
operator through the object to the display screen, and only the
relative maximum value detected along each ray path is retained by
the computer.
This method tends to display bone and contrast material–
material–filled
structures preferentially, and other lowerlower-attenuation structures are
not well visualized.
Also, there are Minimum Intensity Projection techniques.

¬
¬
¬

67

Kanal

It is also possible to sample a 3D stack of CT images along a curved
curved
plane.
This technique can be especially useful when narrowing of tortuous
tortuous
tubular structures (e.g., blood vessels) is suspected.
The desired curve is generally defined along some anatomic structure,

structure,
such as the curve of the mandible as in the panoramic view above.
above.
Other typical curves used include the midline of a vessel in 3D space,
along an implanted stent,
stent, or along the spinal cord.
c.f. Cody. Radiographics, 22:125522:1255-1268, 2002

68

17


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Digital Image Display
Stack Mode Viewing

Curved Reformatting - Jaw
¬

¬

¬

¬

(a) Lateral transparent volumevolumerendered image reformatted from
the axial CT images demonstrates
an impacted molar (arrow).
(b, c) Axial images were used to
locate the alveolar nerve canal
(arrows) and define the formatting
curve (gray line in c).
(d) Curved reformatted image
demonstrates the alveolar nerve
canal in its entirety (gray line).
(e) Image of the plane just lingual
to the alveolar nerve canal reveals
the impacted molar (arrow).

c.f. Cody. Radiographics, 22:125522:1255-1268, 2002

Kanal

¬

¬

¬

¬

69

Kanal

70

CT Fluoroscopy
¬

¬

¬

¬

¬
¬

Take Home Points
¬

In CT Fluoroscopy, the scanner provides pseudopseudo-real time
tomographic images that are most commonly used for guidance
during biopsies
The CT image is constantly updated to include the latest projection
projection
data
Images are typically updated at the rate of six per second, which
which
provides excellent temporal resolution
Any motion at the image level can then be followed in nearly real
real
time by observing the updated reconstruction
Images are displayed on a monitor in the cine mode

Low tube currents (20 to 50 mA) used to minimize dose

Kanal

Kalpana M. Kanal, Ph.D.

In stack mode, a single CT study is displayed and the radiologist
radiologist
selects the image in the study, which is displayed by moving the
computer mouse
Possible to simultaneously display 2 images at the same cut plane,
plane,
such as precontrast and postcontrast images, images at different
window and level settings, or images from the previous and the
current CT exams
Advantage is that it is interactive: The radiologist interacts with
with the
computer in real time to visualize the image data as he or she
interprets the case, following diagnostic clues from slice to slice
slice
Example, follow arteries from image to image

¬
¬
¬

¬

71

Kanal

Single slice versus multislice CT
Certain slice widths selection possible
Pitch definitions different for single and multislice CT scanners
scanners
Multislice CT
¬ Same acquisition in shorter times (fewer motion artifacts)
¬ Thinner slices for better zz-axis resolution
¬ Efficient use of xx-ray beam, reduction of radiation dose (?)
¬ Reconstruction in different slice widths
¬ Possibility of Isotropic Imaging (better MPRs and 3D images with
reduced image artifacts)
Multiplanar Reconstruction better with multislice CT due to thinner
thinner
slices in the zz-axis direction

72

18


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Davis Notes
¬

¬
¬
¬
¬
¬

Raphex 2001 Question

For a 25 cm field of view and a pixel matrix of 500 pixels across
across the
field of view, the limiting spatial resolution is:
A. 0.5 line pairs per mm
B. 2 line pairs/mm
C. 10 line pairs/mm
D. 0.33 line pairs/mm
E. 10 line pairs per cm

¬

¬
¬
¬
¬
¬

¬ E: The pixel size in this example is 250 mm / 500 pixels = 0.5mm
per pixel. The limiting spatial resolution (lp/mm) is = 1/(2x), = 1/(2 x
0.5) = 1.0 lp/mm = 10 lp/cm.

Kanal

¬

73

¬
¬
¬
¬
¬

¬

D. Helical
Helical slice width is greater than axial for pitch 1 due to the
interpolation algorithm which averages in data from more than 1
collimation width.

Kanal

74

Raphex 2001 Question
¬

D31. A CT in helical mode with collimation of 5 mm and pitch of 1 is
compared to axial mode with 5 mm collimation. The measured slice
width will be:
A. No larger than 5 mm in either mode.
B. The same for both modes.

C. Smaller in the helical mode.
D. Larger in the helical mode.
E. Not measurable in helical mode.

Huda Question

D34. For the same collimation, at which pitch would helical mode CT
have a higher patient dose than the axial mode:
A. 3
B. 1.5
C. 1
D. 0.8
E. None of the above.

4. The measured xx-ray transmissions from a single CT fan beam
through a patient is called a:
¬ A. Filter
¬ B. BackBack-projection algorithm
¬ C. Tomographic slice
¬ D. Primary beam
¬ E. Projection

D. The dose will be higher for a pitch of less than 1, since this
represents an overlapped xx-ray beam on the patient surface.

¬

Kanal

Kalpana M. Kanal, Ph.D.

75

Kanal

E. A projection is a profile of transmitted xx-ray intensities through
the patient at any given location of the xx-ray tube, with up to 1,000
projections acquired and used to reconstruct the CT image.

76

19


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Signal to Noise Ratio (SNR)

Huda Question
¬

The main advantage of helical CT over conventional (axial) CT is
improved:
A. Spatial resolution
B. Low contrast detection
C. Data acquisition rate
D. Patient dose

E. Image reconstruction time
¬

¬

¬

¬

¬

¬

¬

C. Fast patient data acquisition is the major benefit of helical CT.

¬

Kanal

77

Kanal

78

Radiation Dose in CT

Radiation Dose in CT

¬

Signal, N = mean photons used to produce the image/unit area
Noise injects a random or stochastic component into an image –
many sources
Quantum noise is the statistical fluctuation in the photons detected,
detected,
and is given by = √N
Can adjust the noise ( ) in an image by adjusting the mean number
of photons used to produce the image
Relative noise = coefficient of variation = /N = 1/√
1/√N (decreases
with increase in N)
SNR = signal/noise = N/ = N/√
N/√N = √N (increased with increased N)
Quantum noise and structure noise both affect the conspicuity of a
target

Three aspects of radiation dose in CT that are unique in comparison
comparison
to xx-ray projection imaging:
X Single CT image obtained in a highly collimated manner
X

Even distribution of dose due to rotational acquisition compared
to a chest radiography

X

CT acquisition requires a high SNR to achieve high contrast

resolution and therefore the dose to the slice volume is higher
because the techniques used are higher
X PA Chest xx-ray – 120 kVp, 2 mAs
X Chest CT – 120 kVp, 200 mAs

Kanal

Kalpana M. Kanal, Ph.D.

79

Kanal

McNittMcNitt-Gray. Multislice CT
Workshop

80

20


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Radiation Dose in CT

Radiation Dose

¬

Compton scattering is the
principal interaction
mechanism in CT, so the
radiation dose attributable to
scattered radiation is
considerable

¬

The acquisition of a CT slice
delivers a considerable dose
from scatter to adjacent
tissues, outside the primary
beam
c.f. Bushberg, et al. The
Essential Physics of Medical
Imaging, 2nd ed., p. 363.

McNittMcNitt-Gray. Multislice CT
Workshop

Kanal

81

Kanal

82

Radiation Dose
Multiple Scan Average Dose (MSAD)

Kanal

Kalpana M. Kanal, Ph.D.

¬

The MSAD is the dose to tissue
that includes the dose
attributable to scattered
radiation emanating from all
adjacent slices

¬

The MSAD is defined as the
average dose, at a particular
depth from the surface, resulting
from a large series of CT slices

McNittMcNitt-Gray. Multislice CT
Workshop

Radiation Dose
Dose Measurement - CTDI

83

Kanal

¬

As estimate of the MSAD can
be accomplished with a single
scan by measuring the CT
dose index (CTDI)

¬

The CTDI can be measured
using a pencilpencil-type ionization
chamber in phantoms that
simulate heads (16 cm
diameter acrylic) and bodies
(32 cm diameter acrylic)

84

21


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Radiation Dose

Dose Measurement - CTDI
¬

¬

¬

Quality Assurance
Radiation Dose

CTDIFDA is defined by the FDA as the radiation dose to any point in
the patient including the scattered radiation contribution from 7 CT
slices in both directions, for a total of 14 slices
Doses at the patient surface may be higher than the dose at the
center of the patient
¬ In head scans, the surfacesurface-toto-center ratio is approximately 1:1
¬ In body scans, the surfacesurface-toto-center ratio is approximately 2:1
CTDI measurements are made at the surface, CTDIperipheral and at
the center, CTDIcenter of the phantom and combined to give CTDIw
(2/3 CTDIperipheral + 1/3 CTDIcenter)

Kanal

85

Kanal

86

Radiation Dose

Dose Measurement - CTDI
¬

¬

¬

Radiation Dose
kVp

CTDI values for body scans are lower than those for head scans
because of the greater attenuation of xx-rays in body
CTDI does not quantify the patient risk because it takes no account
account
of the number of sections scanned, or the radiosensitivity of
irradiated organs

¬

CTDI increases with tube voltage

¬

kVp not only controls the
image contrast but also
controls the amount of
penetration that the xx-ray
beam will have as it
traverses the patient

Parameter

80 kV

120 kV

140 kV

Image
Contrast

Best

Intermediate

Poor

Noise

Most

Average

Least

Penetration

Least

Average

Most

Patient Dose
per mAs

Lowest

Intermediate

Highest

decreasing kVp will reduce dose with other factors constant

Kanal

Kalpana M. Kanal, Ph.D.

87

Kanal

88

22


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course

10 March – 7 April 2005

Dose Considerations in Helical Scanning
Dose Considerations in Helical Scanning
Dose (helical) = Dose (axial) ×
Dose (helical) = Dose (axial) ×

¬

¬

¬

1
collimator pitch

Helical scanning with a collimator pitch of 1.0 is physically similar
similar to
performing a conventional (nonhelical
(nonhelical)) axial scan with contiguous
slices
Collimator pitch of 1.5, dose is 67% (33% less) that of the
conventional CT dose (assuming mAs remains the same)
Collimator pitch of 0.75, dose is 133% (i.e., 33% greater) that of the
conventional CT dose (assuming mAs remains the same)

Kanal

89

Kanal

90

Factors that influence dose –
Tube Current, mA and Time (sec), mAs

Factors that influence dose - Pitch

CTDIvol = CTDIw / pitch

¬

1
collimator pitch

pitch = 0.75

133% of dose at pitch = 1

pitch = 1.5

67% of dose at pitch = 1

pitch = 2.0

50% of dose at pitch = 1

mAs

Pitch – increasing pitch holding all other factors constant reduces
dose

Kanal

Kalpana M. Kanal, Ph.D.

¬

91

Kanal

CTDIw - Head

CTDIw - Body
5.7 mGy

100

13 mGy

200

26 mGy

12 mGy

300

40 mGy

18 mGy

400

53 mGy

23 mGy

mA and time – Dose increases LINEARLY with mAs

92

23


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Factors that influence dose - Collimation
mm

CTDIw - Head CTDIw - Body

1

35.8 mGy

15.9 mGy

3

35.4 mGy

15.7 mGy

5

34.8 mGy

15.4 mGy

7

34.4 mGy

15.3 mGy

10

34.2 mGy

15.2 mGy
mm

¬

Factors that influence dose – Patient Size

MultiMulti-Slice Detector
(Other factors constant
at 140 kVp, 200 mA and
1 second), GE 16-slice

1.25

¬

SingleSingle-Slice Detector
(Other factors constant at 140
kVp, 200 mA and 1 second),
GE single-slice

¬
¬
¬

¬

¬

CTDIw - Head CTDIw - Body
76.3 mGy

41.8 mGy

5

73.5 mGy

40.3 mGy

10

58.4 mGy

32.0 mGy

15

54.7 mGy

30.0 mGy

20

50.0 mGy

27.4 mGy

Kanal

93

¬

¬

American College of Radiology guidelines
¬ Exam
CTDIvol
X Adult Head CT
< 60 mGy
X Peds Abdomen CT
< 25 mGy
X Adult Abdomen CT
< 35 mGy

Kanal

94

Reducing Patient Radiation Dose – Impact on
Diagnostic Image Quality

Reducing Patient Radiation Dose
¬

Patient size – dose more for smaller patients
For same technical factors,
CTDIw – Head > CTDIw – Body because of the greater attenuation of
x-rays in body
Reduce technique factors when scanning smaller adults and
pediatric patients

FDA notice dated 1111-2-01: www.fda.gov/cdrh/safety.html
For Pediatric and small patients

X Reduce tube mA (current)
X Increase pitch
X Develop mA settings based on patient weight or diameter and
body region
X Reduce number of multiple scans without contrast
X Eliminate inappropriate referrals for CT

¬

Decrease mA or current
¬ Increase noise

¬

Increase pitch
¬ Increase volume averaging

¬

Increase axial increment
¬ Introduce gaps

NCI published a guideline and circulated to all ACR members
X www.cancer.gov/cancerinfo/causes/radiationwww.cancer.gov/cancerinfo/causes/radiation-risksrisks-pediatricpediatric-CT

Kanal

Kalpana M. Kanal, Ph.D.

95

Kanal

96

24


Computed Tomography – Chapter 13 Bushberg

Diagnostic Imaging Physics Course
10 March – 7 April 2005

Effective Dose Comparison with Chest PA Exam

Effective doses in CT and Radiographic Examinations
CT examination

Effective dose
[mSv]

Radiographic
examination

Procedures

Eff. Dose [mSv]

Equivalent no.
of chest xx-rays

Approx. period
of background
radiation

Chest PA

0.02

1

3 days

Pelvis

0.7

35

4 months

Abdomen

1.0

50

6 months

CT Chest

8

400

3.6 years

CT Abdomen or
Pelvis

1010-20

500

4.5 years

Effective dose
[mSv]

Head

2

Skull

0.07

Chest

8

Chest PA

0.02

Abdomen

1010-20

Abdomen

1.0

Pelvis

1010-20

Pelvis

0.7

Ba swallow

1.5

Ba enema

7

Typical Background Radiation - 3 mSv per year

Kanal

97

Kanal

98

Current Modulation in CT

Organ Doses in CT
¬

¬

¬

¬

Head CT
X Thyroid - 1.9 mGy
X Eye lens - 40 mGy
Chest CT
X Breast - 21 mGy
Abdomen CT
X Uterus – 8 mGy
X Gonads - 8 mGy
Pelvis CT
X Uterus – 26 mGy
X Gonads - 23 mGy

Kanal

Kalpana M. Kanal, Ph.D.

¬

¬

Patient skin doses are typically
between 20 (body) and 40
mGy (head) or 2 to 4 rad
Induction of erythema is
typically 2 Gy

99

Kanal

¬

Modern CT scanners are capable of modulating the mA (current)
during the scan

¬

The rationale behind this technique is that it takes fewer photons
photons
(lower mA) to penetrate thinner tissue, and more xx-ray photons
(higher mA) are needed to penetrate thicker projections through the

body

¬

Dose can be reduced

¬

Because mA is reduced per gantry rotation, xx-ray tube loading is
reduced and helical scans can be performed for longer periods and
and
with greater physical coverage

¬

Auto mA on GE Scanners

100

25