Mammography Imaging Optimization for Early Detection of Cancer

1
Mammography Imaging Optimization for Early
Detection of Cancer

• Dr. Nabil Maalej
• Physics Dept., KFUPM
• Dec 2007

2
Outline

• Introduction
• Image Formation in Mammography
• Image Quality
• Contrast and resolution
• Scatter and Scatter Reduction
• Quality Assurance and Dosimetry
• Mammography grid optimization
• Conclusion

3
Breast Cancer (Highest)

• Cancer Statistics (2002 Saudi Cancer Registry)

4
Early Diagnosis

• Early diagnosis and treatment improves the chance
  of survival

Detection Status Survival
Before Starts to spread 100
Smaller than 2 cm and hasnt spread 98
2-5 cm and spread to auxiliary lymph nodes 88
Over 5 cm and hasnt spread to lymph nodes 76
5
Breast Screening

• In 2002, the US Preventative Services Task Force
  concluded there was fair evidence that
  mammography screening every 1-2 years could
  reduce breast cancer mortality by approximately
  20 percent to 25 percent over 10 years for women
  age 40 and over.
• Problem Estimates of the numbers of cancers
  missed by mammography are usually around 10-20.

6
Image Formation
7
Mammography Machine
8
X-ray Tube
9
X-ray Tube
10
Mammography Spectrum
11
Photoelectric Interaction
T
h?
h? T BEelectron
12
Compton Scattering
T
h?
h?
h? T BEelectron h?
13
Thomson (Rayleigh) scattering
DEMO
14
IMAGE QUALITY
15
Signs of Cancer in a Mammogram
Microcalcifications
Mass
16
Resolution

• Spatial resolution of an imaging system can be
  defined in terms of the smallest spacing between
  two objects that can be imaged clearly
• Example for Radiography spatial resolution is
  about 0.01 mm, for CT it is about 1 mm
• Good resolution is essential to see
  micro-calcifications in a mammogram

17
Resolution
18
Measurement of Resolution
19
Contrast

• Image Contrast is a measure of difference between
  regions in an image

20
Measuring Contrast
21
Measuring Contrast Detail Curve
22
Contrast

• Image contrast in mammography is due to the
  difference in linear attenuation coefficient and
  thickness of different types of tissue

Io
x
23
Attenuation Coefficient
Contrast decreases toward higher energies ? the
recommended optimum for mammography is in the
region 18 - 33 keV depending on tissue thickness
and composition.
24
Contrast Degradation Due to Scatter
25
Contrast Degradation with Scatter
26
How Can We Reduce Scatter?
X-ray Source
Compression Paddle
Compressed Breast
Image Detector
27
How Can We Reduce Scatter?

• Use an Anti-Scatter Grid
• Use and Air Gap between the breast and the
  detector
• Line Scanning System

28
Anti-scatter grid
breast
leadsepta
detector
29
Air Gap (Magnification)
30
Line Scanning Mammography
31
Image Quality Assurance Measurements
32
Importance of Image Quality
33
Quality Assurance Tests
Mammographic unit Assembly inspection Uniformity of screen speed
Collimation Assessment Breast entrance exposure
Evaluation of the focal spot size Average Glandular dose
kVp accuracy and reproducibility Image quality evaluation
Beam quality assessment (HVL) Artifact evaluation
Automatic Exposure Control performance
Automatic Exposure Control reproducibility
34
RMI 156 Phantom
35
RMI 156 Phantom
ACR minimum score is 4 visible fibers, 3 speck
groups and 3 masses
36
RMI 156 Phantom
37
Radiation Dose to The Breast
38
Risk of Radiation Induced Breast Cancer
39
Mean Glandular Dose

• The mean, or average, dose to the glandular
  tissue within the breast
• 3 mGy per view is the federal and ACR dose limit
  for an average-size patient

40
Entrance Skin Exposure

• The entrance skin exposure is the amount of
  radiation delivered to the patient skin at the
  entrance point of the X-ray

41
SETUP For Entrance Skin Exposure Measurement
X-Ray Tube
Ionization Chamber
Breast Equivalent Phantom
X-Ray Film
Table
Film
42
Entrance Skin Exposure Measurement
Dosimeter Mammography Phantom
43
Grid Geometry Optimization
44
Linear Grid Geometry
45
Grid Optimization
46
Monte Carlo MCNP Code

• Monte Carlo simulation is a stochastic technique
  that uses random numbers and probability
  statistics
• First Developed for the Manhattan Project at the
  Los Alamos National laboratory during World War
  II.
• MCNP A general purpose Monte Carlo n-particle
  code
• Used for particle (neutron, photon and electron)
  transport and interaction with matter as well as
  radiation protection and dosimetry
• Neutron energy (l0-11 MeV to 20 MeV)
• Photon and electron energies (1 keV to 1 GeV).
• We use Monte Carlo code MCNP 5

47
Simulation Geometry
48
Detector Positions

49
Input Parameters

• Source Mo-Mo Anode filter Spectrum
• Compression Paddle Lexan (C16H14O6)n
• Supporting Table Carbon Fiber
• Breast Phantom 50 Glandular and 50 Adipose

50
Validation
51
SPR for Different Phantom Thicknesses
52
SPR for Different kVp
53
SPR Vs. Interspace Thickness for Different Grid
Heights
54
SPR Versus Grid Ratio
55
Contrast Improvement Factor Versus Grid Ratio
56
Bucky Factor Vs. Septa Thickness for Different
Grid Heights
57
Bucky Factor Vs. Grid Ratio
58
Conclusions

• For 3, 5 and 8 cm thick phantom the optimal grid
  design gt h 0.9 mm
• d 12 µm
• D 100 µm

59
Optimal Grid Performance
Phantom Thickness SPR Without Grid SPR with Optimal Linear Grid SPR with Optimal Focused Grid Bucky Factor with Optimal Linear Grid Bucky Factor with Optimal Focused Grid
3 cm 0.379 0.153 - 2.24 -
5 cm 0.581 0.203 0.084 2.43 1.79
8 cm 0.789 0.330 - 2.44 -
60
Results

• Highest SPR occurs near the center of mass of the
  phantom
• SPR increases with phantom thickness
• Optimal linear grid (for h 0.9 mm, t12 mm and
  D100 mm) gives 58-68 scatter reduction and a
  Bucky factor 2.24-2.44
• Focusing the grid septa to the x-ray source gives
  further improvement with scatter reduction of
  86 and Bucky factor of 1.79

61
Conclusion

• Breast cancer is the most prevalent type of
  cancer
• Early detection through screening mammography can
  save lives
• For accurate diagnosis regular testing and
  quality assurance is essential
• Medical physicists are trained to establish
  quality assurance programs to ensure best image
  quality, reduce the dose to the patient and
  reduce the chances of misdiagnosis
• Saudi hospitals and clinics should implement
  quality assurance and radiation protection
  programs to provide quality care for the patients

62
Some Chemistry Cancer Research Ideas for
Investigation

• Design of better biocompatible and biodegradable
  contrast agents that will enhance cancer cell
  imaging
• Design of biocompatible chemicals that will
  generate free radicals and kill cancer cells when
  exposed to an external stimulus
• Find better cancer targeting chemicals for
  targeted cancer therapy to reduce effect on
  normal cells

63
Acknowledgement

• KFUPM support
• Prof. Akhtar Naqvi
• Abdullah Al-Kafi

64
THANK YOU
65
Imaging Quality Parameters
CONTRAST
RESOLUTION
NOISE
66
The Nationwide Evaluation of X-ray Trends (NEXT)

• The Nationwide Evaluation of X-ray Trends (NEXT)
  is a U.S. national program conducted annually to
  measure the x-ray exposure that a standard
  patient receives for selected x-ray examinations

67
QA Frequency
Task Minimum Freq. Approx. time to do a test (minutes)
Darkroom cleanliness Daily lt5
Screen cleanliness Weekly 5
Phantom Image Monthly 10
Repeat Analysis Quarterly 15 to 30
Analysis for Fixer retention in film Quarterly lt5
Darkroom fog Semi-annually 15
Compression Semi-annually Compression
68
Contrast Detail Analysis

• Contrast detail analysis an observer reports the
  size of the smallest object he or she can
  perceive at a certain contrast level and with a
  given noise level in the image.

69
Measuring Contrast Detail Curve
70
Measuring Contrast Detail Curve
71
Measuring Contrast Detail Curve
72
Measurement of Resolution
73
Mammography Spectrum
74
Anti-scatter grid
Scattered Rays
Primary Rays
Grid Septa
Image Receptor
75
Air Gap Effect
76
SPR as a function of Air gap for different
Phantom Thicknesses
77
Object resolution as a function of Air Gap size
78
SPR Vs. Septa Thickness for Different Grid Heights
79
SPR Vs. Septa Thickness for Different Interspace
Thicknesses
80
Bucky Factor Vs. Septa Thickness for Different
Grid Heights
81
SPR Vs. Grid Height for Different Interspace
Thicknesses
82
Pair Production
83
Phtodisintegration
84
Breast Cancer (Highest)

• Cancer Statistics (2001 Saudi Cancer Registry)