1 1 X - Ray Computed Tomography Measures Tissue Properties from Macro to Micro Michael Andre, Ph.D. Department of Radiology University of California, San Diego San Diego VA Healthcare System Outline • Perspective on Computed Tomography (image reconstruction from projections) • Scientific basis and timeline • Evolution of medical design • Principles of image reconstruction • Current medical CT scanner design and applications • System performance and image display • Patient Dose Reduction and Dose Reports • Artifacts 2 Conventional Radiography: 2D map of 3D object 3 • 2D map of x-ray attenuation (e -μx ) • Superposition • Distortion due to non-uniform magnification • Non-uniform exposure • Widely varying image contrast Goals of CT • 2D image of 2D object without superposition • Uniform object contrast • Uniform contrast sensitivity • “Calibrated” image • Structure and function Cross-sectional anatomy was a new challenge to the medical community 5 Transmission Emission Reconstruction from Projections X - Ray, microwave, ultrasound, optical SPECT, PET, MRI , ) , ( , r ds y x f r p Interaction map 6 Transmission Computed Tomography (“Image Reconstruction”) I 0 I Assumed conditions for x-ray CT: • Straight-line propagation • Monoenergetic x-ray beam
22
Embed
Interaction map - UCSD RadResradres.ucsd.edu/secured/CH13 Andre.2015.CTlecture.pdf · Tissue Properties from Macro to Micro Michael Andre, Ph.D. Department of Radiology University
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
1
X-Ray Computed Tomography Measures
Tissue Properties from Macro to Micro
Michael Andre, Ph.D.Department of Radiology
University of California, San Diego
San Diego VA Healthcare System
Outline
• Perspective on Computed Tomography
(image reconstruction from projections)
• Scientific basis and timeline
• Evolution of medical design
• Principles of image reconstruction
• Current medical CT scanner design and applications
• System performance and image display
• Patient Dose Reduction and Dose Reports
• Artifacts
2
Conventional Radiography: 2D map of 3D object
3
• 2D map of x-ray attenuation (e-µx)
• Superposition
• Distortion due to non-uniform magnification
• Non-uniform exposure
• Widely varying image contrast
Goals of CT• 2D image of 2D
object without
superposition
• Uniform object
contrast
• Uniform contrast
sensitivity
• “Calibrated” image
• Structure and
function
Cross-sectional anatomy was a new challenge to the medical community
5
Transmission Emission
Reconstruction from Projections
X-Ray, microwave,
ultrasound, opticalSPECT, PET, MRI
,
),(,r
dsyxfrp
Interaction map
6
Transmission Computed
Tomography
(“Image Reconstruction”)
I0I
Assumed conditions for x-ray CT:
• Straight-line propagation
• Monoenergetic x-ray beam
2
7μ/ρ = Mass attenuation coefficient (cm2/gm)
Incident
Intensity
I0
Transmitted
Intensity
I
Interaction Map: I = I0 e-µx
X
Select “x”
for desired
resolution
and FOV
8
Equivalent transmitted intensity
Sum attenuation coefficients Solution: We need more views
9
Goals of CT• 2D image of 2D
object without
superposition
• Uniform object
contrast
• Uniform contrast
sensitivity
• “Calibrated” image
• Structure and
function
Cross-sectional anatomy was a new challenge to the medical community
11
First Commercial CT ScannerEMI Limited, 1973
I0
I
Translate
Rotate
• “First Generation” Scanner
• Translate-Rotate design
• 13 min to acquire singe slice
• Head scan only
• Rigid head holder and water path 12
1973 1979
First Application: Head Trauma
48x64 512x512
13 min scan, 13 min
reconstruction per slice!
2nd Generation: 30 sec scan,
1 min reconstruction
3
13
Second Generation Scanner
• Translate-Rotate design
• More detectors so fewer angles needed
• 30 sec to acquire singe slice
• Capable of body imaging with breath hold 14
Third Generation Scanner Geometry
• Tube and detector array rotate together
• Complete detector profile from single x-ray tube pulse
• 0.3 – 5 sec rotation speed to acquire single slice
• All current CT scanners use this design geometry
• Number and size of detector elements limits resolution
• Width of fan beam determines Field of View (FOV)
15
Fourth Generation Scanner
• Tube rotates, detectors stationary
• Allows oversampling in rotation to increase resolution significantly
• 1 sec to acquire singe slice
16
Electron Beam CT
• No mechanical motion
• Single slice in 30-50 msec
17
• Required two adjacent rooms to house system
• Remarkable cardiac images, poor for everything else
• Doomed by the advent of multi-slice helical scanners
EBCT Image Reconstruction Methods in Medicine
I. Analytic Methods
a. Filtered back-projection (Convolution, Radon, Fourier
filtering)
1) Dominant method for decades
2) Very fast reconstruction
b. Two- or three-dimensional Fourier reconstruction
1) MRI
II. Iterative Numerical Methods
a. Slower but more accurate (100-1000X longer than FBP)
b. Significant dose reduction
c. Simultaneous Iterative Reconstruction Technique (SIRT)
Average Annual Effective Dose to the US Population
Early 1980s 2006
MedicalMedical
BackgroundBackground
Average Annual Effective Dose from All Sources = 6 mSv
68
Medical = 50%
A major reason for
this increase is
greater utilization
of CT imaging.
High priority effort
in Radiology to
reduce all patient
doses has been
successful
especially in
pediatrics.
CT
Medical Sources of Radiation
Adapted from ICRP Publication 102: Managing Patient Dose in Multi-Detector Computed Tomography (MDCT), 102 Annals of the ICRP Volume 37/1, Chapter 4 (2007)
The effective dose of a single abdomen and pelvis CT scan is greater than three times that of a year of background radiation (what is that number again?)
DEXA scan 0 .004
70
Patient Dose Reports in CT
Dose Indices in CT:
• CT Dose Index (volumetric)
• Dose-Length Product
• CTDIVOL and DLP indicate
measured doses to 100 mm ion
chamber in a cylindrical plastic
phantom
• CT head phantom, 16 cm
• CT body phantom, 32 cm
• Not a measured dose to the
patient
71 72
13
CT Dose Report Definitions
73
Scanner Display
• CTDIvol (mGy)
Weighted avg measurement in
phantom w/pencil chamber
• DLP (mGy•cm)
CTDIvol x scan z-axis length
• Dose Efficiency (%)
Measure of z-axis beam usage
• Projected Series DLP
Based on tabulated measures
• Accumulated Exam DLP
Sum of series DLPs
Images CTDIvol
mGy
DLP
mGy•
cm
Dose Eff
%
Phantom
cm
1-8 10.44 20.88 92.70 Head 16
9-977 4.01 249.59 92.70 Body 32
978-
1025
4.65 27.32 92.70 Body 32
1026-
1217
18.98 113.94 92.70 Body 32
Projected series DLP:
Accumulated exam DLP:
532.03
1561.23
mGy•cm
mGy•cm
Methods to Reduce CT Dose
74
CT Exam Reference
CTDIvol
Notification
CTDIvol
Typical
Dose Equiv
Head 75 mGy 80 mGy 2 mSv
Adult Abdomen 25 mGy 50 mGy 8 mSv
Pediatric Abdomen (5 yo) 20 mGy 25 mGy 2 mSv
Brain Perfusion 500 mGy 600 mGy 12 mSv
Background Radiation Dose 3.1 mSv/yr
• Reduce mA
• Dose is proportional to mA
• Use adaptive mA (less for thinner pats or body part)
• Reduce kVp for thinner patients/pediatrics
•Dose increases exponentially with kVp
• Increase pitch, gantry speed or detector aperture
• Use iterative reconstruction
• Reduce number of series -- Do you really need pre-contrast?
• Angle gantry to avoid direct exposure of eyes, breast, gonads
Results
in more
noise
Automatic Tube Current Modulation
Fixed mA technique uses constant dose: too low or too high.
Auto-mA modulation may be vulnerable to user error.• Four adjustments per rotation: Sectors 1-4
• Noise index and slice thickness determine mA and thus dose.
• Reducing slice thickness without adjusting the noise index can cause