Slide 1
Slide 2
Introduction
Optimizing Dose and
Image Quality in Digital
Mammography
Eric A. Berns,
Berns, PhD
Certified Statistics – 15 Months
April 1, 2005
August 1, 2006
Difference
Total Certified
Mammo
Facilities
8,929
8,829
-100
Total Accredited
Mammo Units
13,640
13,556
-84
Certified
Facilities with
FFDM units
607
1,130
+523
Accredited
FFDM units
819
1,604
+785
eberns@radiology.northwestern.edu
Northwestern University Medical School
Lynn Sage Comprehensive Breast Center
Chicago, IL
Slide 3
Slide 4
Introduction
FFDM vs. Date
Fuji FCRm
App: Jul ‘06
Detector Types
12%
Siemens Novation
App: Aug ‘04
% Fac w Dig
10%
% Units FFDM
6%
Cesium Iodide with Silicon Diode Array (GE)
Lorad Selenia
App: Oct ‘02
Lorad CCD
App: Mar ‘02
8%
Selenium with Silicon Diode Array (Lorad
(Lorad,, Siemens,
Fischer Senoscan
App: Sept ‘01
GE Essential
App: Apr ‘06
4%
Kodak, Agfa)
GE 2000D
App: Jan 2000
2%
0%
Jul-98
Slot Scanning CCD Array (Fischer)
GE DS
App: Feb 2004
Dec-99
Apr-01
Sep-02
Jan-04
May-05
Oct-06
Computed Radiology (Fuji, Kodak, Konica, Agfa)
1
Slide 5
Slide 6
Equipment
GE Senographe 2000D
Slide 7
Equipment
GE Senographe DS
Slide 8
Equipment
Equipment
GE Senographe 2000D & DS & Essential
– FOV: 19.2 x 23.0 cm & 24.0 x 31.0 cm
– Spatial resolution: 100 microns (5.0 lp/mm)
lp/mm)
GE Senographe Essential
2
Slide 9
Slide 10
Equipment
Equipment
Siemens Mammomat NovationDR
Lorad Selenia
Slide 11
Slide 12
Equipment
Equipment
Lorad Selenia & Siemens Novation
– FOV: 24 x 29 cm
– Spatial resolution: 70 microns (7.14 lp/mm)
lp/mm)
Fischer Senoscan
3
Slide 13
Slide 14
Equipment
Equipment
Computed Radiology
Fischer Senoscan
– Fuji – FDA approved
– FOV: 21 x 29 cm (std), 11 x 15 cm (high res)
– Kodak
– 54 & 27 microns – 9.3 lp/mm
lp/mm in 54 µm mode
– Konica
– Scan time:
time: 5.2 seconds
– Agfa
Slide 15
Slide 16
Equipment
Equipment
Fuji FCRm
Fuji FCRm
– FOV: 18 x 24 and 24 x 30 cm
– Spatial resolution: 50 microns (10.0 lp/mm)
lp/mm)
– 60 to 80 imaging plates per hour
Flash Plus IIPm
FCSm
4
Slide 17
Slide 18
Equipment
Equipment
Kodak DirectView CR Mammography System
Sectra Microdose
– SlotSlot-scanning photon counter detector
– FOV: 24 x 26 cm
– Spatial resolution: 50 microns (10.0 lp/mm)
lp/mm)
Slide 19
Slide 20
Equipment
Equipment
• FDAFDA-Approved Laser Imagers ~ 40 µ m spot
•Agfa LR5200 Laser Imager (Wet Chemistry)
•Agfa DS4500M
•Kodak 8600 Laser Imager
•Kodak 8610 Laser Imager
•Kodak 8900
•Fuji Drypix 4000, 5000, 7000
•Fuji Drypix FMFM-DP L
•Codonics Horizon Ci,
Ci, GS, SF
Planmed Nuance – Amorphous Selenium – 85 µm
•Konica Minolta DryPro 793
5
Slide 21
Slide 22
Workstations
Workstations
What is a Multimodality Review Workstation?
Until recently, workstations came with acquisition
units
Those days are over
Spurred by FDA approval of third party RWS, high
resolution displays, and PACS
As a result, FFDM is becoming more a la carte to
allow best of each (acquisition, RWS, PACS, etc.)
Slide 23
Slide 24
Multimodality Workstations
Workstations
FDA Approved Multimodality Workstations
– AGFA IMPAX MA3000
Current challenges
– Sectra IDS5/mx
– Kodak DirectView PACS System
– Device to device connectivity
– McKesson’
McKesson’s PACS Mammo Station
– iCad Second Look 500M
– DICOM incompatibilities between acquisition and
– Cedara I-ReadMammo
– Fuji Synapse
displays
– GE Seno Advantage & Seno Adv. 2
– Fischer Senoview Plus (Cedara
(Cedara))
• Can result in image degradation
– Hologic SecureView DX
– Siemens MammoReportPlus
– PACS connectivity
6
Slide 25
Slide 26
Workstations
Detector Size (cm)
AEC System Function
vs. Image size (pixels)
5625
GE Essential
GE (2000D, DS, Essential)
Fischer
Mo & Rh Targets, Mo & Rh Filters
Lorad & Siemens
Lorad & Siemens
4100
3100
31
29
Fischer
23
Monitor
3 Modes
2560
2304
GE
2000D &
DS
– Mo/Mo, Mo/Rh
Mo/Rh,, Rh/Rh
GE Essential
– Con – Thin, less dense (2000D)
GE
2000D &
DS
– Std – Intermediate (DS & Essential)
– Dos – Thick, dense breasts
19 21 24
24003400 4096
1920
2048
Slide 27
Slide 28
AEC System Function
Lorad – Mo target, Mo & Rh Filters
– Mo/Mo, Mo/Rh
Mo/Rh
Modes
–
–
–
–
AutoAuto-Filter – AEC sensor, exposure adjustment
AutoAuto-kV – Filter, AEC sensor, exposure adjustment
AutoAuto-Time – kV, filter, AEC sensor, exposure adjustment
TEC – (Tissue Exposure Control) - Breast density
AEC System Function
Siemens – Mo & W Targets, Mo & Rh Filters
– Mo/Mo, Mo/Rh
Mo/Rh,, W/Rh
W/Rh
Modes
– OPDOSE Mode (recommended using W/Rh
W/Rh))
• Compress breast to given compression
• (enhanced manual mode)
– Manual
Recommended Mode – AutoAuto-Filter
• T/F & kVp prepre-selected by vendor lookup tables
– Manual
7
Slide 29
Slide 30
2000 FFDM
Variability Data
AEC System Function
AEC Summary
How Does Digital Compare to ScreenScreen-film
Mammography in Terms Of:
– Manufacturer only recommends, user decides
• Exposure
– Ultimately up to the Radiologist and/or Physicist
Times?
• Breast Dose?
to decide which mode to use
• Detection of LowLow-contrast Lesions?
– Make mode choice based on knowing effect on
3 AOP Modes:
dose and image quality
Slide 31
Contrast
Standard O
Dose
+
-
Slide 32
Methods
ContrastContrast-Detail image analysis
Methods
Contrast Detail Phantoms
– Acquire images at recommended techniques
• 2, 4, 6, 8 cm
• Calculate Dose
• Score contrastcontrast-detail image for image quality
ACR Phantom
– Calculate Dose
– Scores
Compare to screenscreen-film data
8
Slide 33
Slide 34
Methods
Methods
Decreasing Diameter
0.25 mm
4.0 mm
4%
Decreasing
Contrast
0.25%
Slide 35
Slide 36
Methods
ACR Phantom
ACR Phantom
9
Slide 37
Slide 38
Average Glandular Dose vs. Thickness
38 ScreenScreen-film Units, 18 FFDM Units
Exposure Time vs. Thickness
38 ScreenScreen-film Units, 18 FFDM Units
700
6
CMAP ACR Phantom
CMAP 6 cm
500
CMAP 8 cm
FFDM - Con/Auto
400
74% of
FSM > 2 s
CMAP 2 cm
CMAP 4 cm
Exposure Time (seconds)
Average G landular Dose (m rad)
CMAP 2 cm
600
3 mGy Limit for
ACR Phantom
FFDM - Std/Auto
FFDM - Dos/Auto
300
200
100
0
5
CMAP 4 cm
CMAP ACR Phantom
CMAP 6 cm
4
16% of
FSM > 2 s
CMAP 8 cm
FFDM - Con/Auto
3
FFDM - Std/Auto
2 seconds
FFDM - Dos/Auto
2
1
0
0
1
2
3
4
5
6
7
8
9
10
0
1
2
3
Thickness (cm)
5
6
7
8
9
10
Thickness (cm)
Slide 39
Slide 40
Contrast Detail Scores vs. Thickness
38 ScreenScreen-film Units, 18 FFDM Units
18
Methods
2002
GE 2000D Opto Data
16
14
C D S co res
4
12
CMAP 2 cm
CMAP 4 cm
CMAP 6 cm
CMAP 8 cm
FFDM - Con/Auto
FFDM - Std/Auto
FFDM - Dos/Auto
10
8
6
0
2
4
6
Thickness (cm)
8
10
Medical Physics. March 2003. 30 (3) pages 334-340
10
Slide 41
Slide 42
Methods
Results
CD Score vs. kVp
2 cm Breasts
15
Objective: To determine optimized technique
14
Mean CD Score
factors for fullfull-field digital mammography
system (GE 2000D) for lowlow-contrast lesion
detection
13
12
11
2 cm Mo/Mo
10
Optimization done under condition of matched
9
patient dose to screenscreen-film mammography
8
2 cm Mo/Rh
Mo/Rh
2 cm Rh/Rh
ScreenScreen-Film
20
25
30
Compare fullfull-field digital results to screenscreen-film
results
35
45
50
Mo/Mo Trend pp-value = 0.0752
Mo/Rh
p-value = 0.1369
Mo/Rh Trend pRh/Rh Trend pp-value = 0.0985
Optimized Technique Comparison
p-value = 0.47
Slide 43
40
kVp
Slide 44
Results
Results
CD Score vs. kVp
6 cm Breasts
15
15
14
14
Mean CD Score
Mean CD Score
CD Score vs. kVp
4 cm Breasts
13
12
11
4 cm Mo/Mo
10
4 cm Mo/Rh
Mo/Rh
13
12
11
6 cm Mo/Mo
10
6 cm Mo/Rh
Mo/Rh
6 cm Rh/Rh
4 cm Rh/Rh
9
9
ScreenScreen-Film
ScreenScreen-Film
8
8
20
25
30
35
kVp
Optimized Technique Comparison
p-value = 0.013
40
45
50
Mo/Mo Trend pp-value = 0.2221
Mo/Rh
Mo/Rh Trend pp-value = 0.5691
Rh/Rh Trend pp-value = 0.0123
20
25
30
35
kVp
Optimized Technique
Comparison
p-value < 0.0001
40
45
50
Mo/Mo Trend pp-value = 0.5710
Mo/Rh
Mo/Rh Trend pp-value = 0.6496
Rh/Rh Trend pp-value = 0.2735
11
Slide 45
Slide 46
Results
Conclusions
CD Score vs. kVp
8 cm Breasts
15
LowLow-contrast lesion optimization for FFDM (GE 2000D)
8 cm Mo/Mo
Mean CD Score
14
8 cm Mo/Rh
Mo/Rh
8 cm Rh/Rh
13
– Thin breasts (< 2 cm): Mo/Mo with low kVp
ScreenScreen-Film
12
– Intermediate breasts (~4 cm): Insensitive to targettarget-filter and kVp
11
10
selection
9
– Thick breasts (>5 cm): Rh/Rh with higher kVp
8
20
25
30
35
40
kVp
45
50
Mo/Mo Trend pp-value < 0.0428
Mo/Rh
p-value < 0.9453
Mo/Rh Trend pRh/Rh Trend pp-value < 0.0121
Optimized Technique Comparison
p-value < 0.0001
Slide 47
Slide 48
ACRIN Data
Compressed Breast Thickness vs. Dose - All GE Sites (5)
6
Compressed Breast Thickness vs. Dose - All Fischer Sites
Mean dose is 32% lower
with FFDM than SFM
6
Film-screen
5
Digital
5
4
Dose (mGy)
Dose (mGy)
4
3
2
3
2
1
1
Film Screen
Digital
0
0
1
2
3
4
5
6
7
8
9
Compressed Breast Thickness (cm)
10
0
0
Fischer FFDM doses were 32% lower than SFM doses
2
4
6
8
10
Compressed Breast Thickness (cm)
12
Slide 49
Slide 50
Compressed Breast Thickness vs. Dose - All Lorad Sites
Compressed Breast Thickness vs. Dose - All Fuji Sites
6
6
Film-screen
Filmscreen
Digital
5
4
4
Dose (m Gy)
Dose (mGy)
Digital
5
3
2
2
1
1
0
0
0
2
4
6
8
0
10
2
4
6
8
10
Compressed Breast Thickness (cm)
Compressed Breast Thickness (cm)
Lorad FFDM doses were 2% higher than SFM doses
Fuji FFDM doses were 5% lower than SFM doses
Slide 51
Slide 52
Mean Dose Comparison by Digital Manufacturer
3.0
Screen-Film
Overall: Digital Doses
22% Lower Than SFM
Digital
2.6
mGy))
Mean Glandular Dose ((mGy
3
2.4
2.5
ARRS 2006
2.4
2.3 2.3
2.3
2.2
2.0
1.9
1.8
1.7
1.5
1.0
0.5
Comparison of Image Quality
and Average Glandular Dose
on Four FDA-approved FullField Digital Mammography
Systems
0.0
Fischer
Fuji
GE
Lorad
All Sites
13
Slide 53
Slide 54
Purpose
Methods
To Measure and Compare
4 FDAFDA-approved FFDM systems
– Image Quality
– GE 2000D
– Average Glandular Dose
• On 4 FDAFDA-approved FFDM systems
– GE Senographe DS
• Across full range of breast thicknesses
– Lorad Selenia
• Using each manufacturers’
manufacturers’ recommended techniques
Slide 55
– Siemens Novation
Slide 56
Methods
Results
Average Glandular Dose Comparison
To acquire images
– Position the phantom like a clinical acquisition
4
DS
3.5
– Apply 10 dN compression force
technique
– Record technique factors
– Measure HVL’
HVL’s and entrance exposures
– Calculate AGD
– Measure and calculate ContrastContrast-Detail scores
Dose (mGy)
– Acquire image using manufacturer’
manufacturer’s recommended
Lorad
2000D
3
Lorad
GE 2000D
Siem
2.5
Range up
to 2.9
Siemens
2
GE DS
1.5
1
0.5
0
0
2
4
6
Breast Thickness (cm)
8
10
14
Slide 57
Slide 58
Results
Results
Contrast-Detail Scores
17
16
Contrast-Detail Scores
Average Glandular Dose Comparison
4
Lorad
GE 2000D
GE DS
Siemens
14
18
DS
3.5
3
13
Lorad
12
Siemens
GE DS
GE 2000D
11
17
2000D
Contrast-Detail Score
15
Dose (mGy)
Contrast-Detail Score
18
Lorad
Siem
2.5
2
1.5
1
0.5
15
14
13
12
11
10
9
8
0
10
16
0
2
4
6
Breast Thickness (cm)
8
10
0
2
4
6
Breast Thickness (cm)
8
10
9
8
0
2
4
6
Breast Thickness (cm)
8
10
Slide 59
Slide 60
Results
Conclusions
ACR Phantom Average Glandular Dose
3
– Average glandular doses varies by up to a factor
of 2.9
2.5
AGD (mGy)
Results indicate that
2
– There are significant differences in image quality
1.5
1
– Technique factors and automatic exposure mode
selection can play an important role in clinical
image quality and patient dose
0.5
0
Siemens
GE DS
GE 2000D
Lorad
15
Slide 61
Slide 62
Digital
Proposed
Phantom
Slide 63
Slide 64
Phantom Comparison
ACR
Phantom
Object #
ACR
Proposed
1
1.56
2
1.12
3
0.89
4
0.75
5
0.54
0.54
6
0.40
0.40
Fiber Diameter
(mm)
0.89
0.82
0.75
0.65
0.30
16
Slide 65
Slide 66
Phantom Comparison
ACR
Phantom
Object #
ACR
Proposed
Fiber Diameter (mm)
ACR
Proposed
Speck Diameter
(mm)
Phantom Comparison
ACR
Phantom
Object #
ACR
Proposed
Fiber Diameter (mm)
ACR
0.54
1
1.56
0.54
1.12
0.40
2
1.12
0.40
3
0.89
3
0.89
0.82
4
0.75
0.75
0.24
0.54
0.54
6
0.40
0.24
4
0.75
0.20
0.16
0.89
0.32
0.82
0.28
0.65
5
0.32
0.75
Proposed
2.00
1.56
2
0.32
ACR
Mass Thickness
(mm)
1
0.89
Proposed
Speck Diameter
(mm)
1.00
0.32
0.75
0.28
0.24
0.65
0.24
0.50
0.20
0.16
0.16
0.75
0.68
0.50
0.38
0.16
5
0.54
0.54
0.25
0.40
0.13
6
0.40
0.40
0.13
0.25
0.20
0.30
0.10
0.30
0.10
0.15
Slide 67
17
Slide 72
Optimization 2006
Objective: To determine optimized technique factors for
clinically available fullfull-field digital mammography systems
On 6 FFDM Units (Siemens counts as 2)
Using
– ContrastContrast-Detail Phantoms
4, 3, 3
3, 1, 1
– ACR Phantom
– Proposed ACR Digital Phantom
18
Slide 73
Slide 74
Optimization 2006
Optimization 2006
Measure dose at all recommended techniques
Evaluate performance of the ACR phantom and
and compare to ACRIN doses
compare to our proposed ACR digital phantom
Compare dose and image quality within each
Evaluate image quality as a function of dose for
system using each mode
different modes
Compare dose and image quality for each
Ultimately find the optimum techniques to provide
system using their recommended mode
Slide 75
the highest image quality with the lowest dose
Slide 76
HVL vs. kVp
HVL vs. kVp
0.65
GE 2000D Mo/Mo
GE 2000D Mo/Rh
GE 2000D Rh/Rh
GE DS Mo/Mo
GE DS Mo/Rh
GE DS Rh/Rh
GE Essential Mo/Mo
GE Essential Mo/Rh
GE Essential Rh/Rh
Lorad Mo/Mo
Lorad Mo/Rh
Siemens Mo/Mo
Siemens Mo/Rh
Siemens W/Rh
0.55
0.45
0.35
0.25
0.55
HVL (mm Al)
HVL (mm Al)
0.65
0.45
GE 2000D Mo/Mo
GE DS Mo/Mo
GE Essential Mo/Mo
Lorad Mo/Mo
Siemens Mo/Mo
0.35
0.25
20
30
40
kVp
50
20
25
30
35
40
45
50
kVp
19
Slide 77
Slide 78
HVL vs. kVp
0.65
0.65
0.55
0.55
HVL (mm Al)
HVL (mm Al)
HVL vs. kVp
0.45
GE 2000D Mo/Rh
GE DS Mo/Rh
GE Essential Mo/Rh
Lorad Mo/Rh
Siemens Mo/Rh
0.35
0.45
GE 2000D Rh/Rh
GE DS Rh/Rh
GE Essential Rh/Rh
Siemens W/Rh
0.35
0.25
0.25
20
25
30
35
40
45
50
20
25
30
kVp
Slide 79
40
45
50
Slide 80
Entrance Skin Exposure vs. kVp
Entrance Skin Exposure vs. kVp
45
45
GE 2000D Mo/Mo
GE 2000D Mo/Rh
GE 2000D Rh/Rh
GE DS Mo/Mo
GE DS Mo/Rh
GE DS Rh/Rh
GE Essential Mo/Mo
GE Essential Mo/Rh
GE Essential Rh/Rh
Lorad Mo/Mo
Lorad Mo/Rh
Siemens Mo/Mo
Siemens Mo/Rh
Siemens W/Rh
35
30
25
20
15
10
5
40
ESE (mR/mAs)
40
ESE (mR/mAs)
35
kVp
35
30
25
20
GE 2000D Mo/Mo
GE DS Mo/Mo
GE Essential Mo/Mo
Lorad Mo/Mo
Siemens Mo/Mo
15
10
5
0
0
20
30
40
kVp
50
20
25
30
35
40
45
50
kVp
20
Slide 81
Slide 82
Entrance Skin Exposure vs. kVp
Entrance Skin Exposure vs. kVp
45
45
40
GE 2000D Mo/Rh
GE DS Mo/Rh
GE Essential Mo/Rh
Lorad Mo/Rh
Siemens Mo/Rh
35
30
25
20
15
GE 2000D Rh/Rh
GE DS Rh/Rh
GE Essential Rh/Rh
Siemens W/Rh
35
ESE (mR/mAs)
ESE (mR/mAs)
40
30
25
20
15
10
10
5
5
0
0
20
25
30
35
40
45
20
50
25
30
35
Slide 83
40
45
50
kVp
kVp
Slide 84
Average Glandular Dose vs. Breast Thickness
4.0
GE 2000D - Std/Auto
GE 2000D - Dos/Auto
3.0
AGD (mGy)
Dose
GE 2000D - Con/Auto
3.5
GE DS - Con/Auto
GE DS - Std/Auto
2.5
GE DS - Dos/Auto
GE Essential - Con/Auto
2.0
GE Essential - Std/Auto
GE Essential - Dos/Auto
1.5
Lorad - Auto-Filter
1.0
Siemens - OpComp - Mo/Mo
Siemens - OpComp - W/Rh
0.5
ACRIN - Film Dose
ACRIN - Digital Dose
0.0
0
2
4
6
8
10
Breast Thickness (cm)
21
Slide 85
Slide 86
Average Glandular Dose vs. Breast Thickness
Average Glandular Dose vs. Breast Thickness
4.0
4.0
GE 2000D - Con/Auto
3.5
GE 2000D - Dos/Auto
GE DS - Std/Auto
3.0
AGD (mGy)
AGD (mGy)
3.0
GE DS - Con/Auto
3.5
GE 2000D - Std/Auto
2.5
2.0
1.5
GE DS - Dos/Auto
2.5
2.0
1.5
1.0
1.0
0.5
0.5
0.0
0.0
0
2
4
6
Breast Thickness (cm)
8
0
10
4
6
8
10
Breast Thickness (cm)
Slide 87
Slide 88
Average Glandular Dose vs. Breast Thickness
Average Glandular Dose vs. Breast Thickness
4.0
4.0
GE Essential - Con/Auto
3.5
GE 2000D - Con/Auto
3.5
GE Essential - Std/Auto
3.0
GE DS - Std/Auto
3.0
GE Essential - Dos/Auto
AGD (mGy)
AGD (mGy)
2
2.5
2.0
1.5
2.0
1.5
1.0
1.0
0.5
0.5
0.0
GE Essential - Std/Auto
2.5
0.0
0
2
4
6
Breast Thickness (cm)
8
10
0
2
4
6
8
10
Breast Thickness (cm)
22
Slide 89
Slide 90
Average Glandular Dose vs. Breast Thickness
Average Glandular Dose vs. Breast Thickness
4.0
4.0
3.5
Siemens - OpComp - Mo/Mo
3.0
GE DS - Std/Auto
GE Essential - Std/Auto
3.0
Siemens - OpComp - W/Rh
AGD (mGy)
AGD (mGy)
GE 2000D - Con/Auto
Lorad - Auto-Filter
3.5
2.5
2.0
1.5
Siemens - OpComp - W/Rh
ACRIN - Film
Acrin-Digital
2.0
1.5
1.0
1.0
0.5
0.5
0.0
Lorad - Auto-Filter
2.5
0.0
0
2
4
6
8
10
0
2
4
Breast Thickness (cm)
Slide 91
6
8
10
Breast Thickness (cm)
Slide 92
Contrast-Detail Score vs. Breast Thickness
17
GE 2000D - Std/Auto
Contrast-Detail Score
CD Scores
GE 2000D - Con/Auto
16
15
GE 2000D - Dos/Auto
GE DS - Con/Auto
14
GE DS - Std/Auto
13
GE DS - Dos/Auto
GE Essential - Con/Auto
12
GE Essential - Std/Auto
11
GE Essential - Dos/Auto
Lorad - Auto-Filter
10
Siemens - OpComp - Mo/Mo
9
Siemens - OpComp - W/Rh
8
0
2
4
6
8
10
Breast Thickness (cm)
23
Slide 93
Slide 94
Contrast-Detail Score vs. Breast Thickness
17
17
16
16
Contrast-Detail Score
Contrast-Detail Score
Contrast-Detail Score vs. Breast Thickness
15
14
13
12
11
10
GE 2000D - Con/Auto
GE 2000D - Std/Auto
GE 2000D - Dos/Auto
9
15
14
13
12
11
10
GE DS - Con/Auto
GE DS - Std/Auto
GE DS - Dos/Auto
9
8
8
0
2
4
6
8
10
0
2
Breast Thickness (cm)
Slide 95
6
8
10
8
10
Slide 96
Contrast-Detail Score vs. Breast Thickness
Contrast-Detail Score vs. Breast Thickness
17
17
16
16
Contrast-Detail Score
Contrast-Detail Score
4
Breast Thickness (cm)
15
14
13
12
11
GE Essential - Con/Auto
10
GE Essential - Std/Auto
9
15
14
13
12
11
Lorad - Auto-Filter
10
Siemens - OpComp - Mo/Mo
9
GE Essential - Dos/Auto
8
Siemens - OpComp - W/Rh
8
0
2
4
6
Breast Thickness (cm)
8
10
0
2
4
6
Breast Thickness (cm)
24
Slide 97
Slide 98
Contrast-Detail Score vs. Breast Thickness
17
ACR
Phantom
Contrast-Detail Score
16
15
14
13
12
GE 2000D - Con/Auto
11
GE DS - Std/Auto
10
GE Essential - Std/Auto
Lorad - Auto-Filter
9
Siemens - OpComp - W/Rh
8
0
2
4
6
8
10
Breast Thickness (cm)
Slide 99
Slide 100
Fibers for the ACR Phantom
7
2.50
6
5
2.00
Fibers
AGD (mGy)
AGD for the ACR Phantom
3.00
1.50
1.00
3
2
0.50
0.00
4
1
Con
Std
2000D
Dos
Con
Std
DS
Dos
Con
Std
Essential
Auto Mo/Mo W/Rh
Filter
Siemens
Lorad
Dos
0
Con
Std
2000D
Dos
Con
Std
DS
Dos
Con
Std
Essential
Dos
Auto Mo/Mo W/Rh
Filter
Siemens
Lorad
25
Slide 101
Slide 102
Masses for the ACR Phantom
6
5
5
4
4
Masses
6
3
3
2
2
1
1
0
Con
Std
2000D
Dos
Con
Std
DS
Dos
Con
Std
Essential
Dos
0
Auto Mo/Mo W/Rh
Filter
Siemens
Lorad
Slide 103
Con
Std
Dos
Con
2000D
Std
Dos
DS
Con
Std
Essential
Dos
Auto Mo/Mo W/Rh
Filter
Siemens
Lorad
Slide 104
AGD for "Proposed" ACR Digital Phantom
Proposed ACR
Digital Phantom
3.00
2.50
AGD (mGy)
Speck Groups
Speck Groups for the ACR Phantom
2.00
1.50
1.00
0.50
0.00
Con
Std
2000D
Dos
Con
Std
DS
Dos
Con
Std
Essential
Dos
Auto Mo/Mo W/Rh
Filter
Siemens
Lorad
26
Slide 105
Slide 106
Fibers for "Proposed" ACR Digital Phantom
8
7
7
6
6
Speck Groups
8
5
Fibers
Speck Groups for "Proposed" ACR Digital Phantom
4
3
5
4
3
2
2
1
1
0
Con
Std
Dos
Con
2000D
Std
Dos
DS
Con
Std
Dos
Essential
0
Auto Mo/Mo W/Rh
Filter
Siemens
Lorad
Slide 107
Con
Std
2000D
Dos
Con
Std
DS
Dos
Con
Std
Essential
Dos
Auto Mo/Mo W/Rh
Filter
Siemens
Lorad
Slide 108
Masses for "Proposed" ACR Digital Phantom
Vary Dose
from Rec.
Techs.
8
7
Masses
6
5
4
3
2
1
0
Con
Std
2000D
Dos
Con
Std
DS
Dos
Con
Std
Essential
Dos
Auto Mo/Mo W/Rh
Filter
Siemens
Lorad
27
Slide 109
Slide 110
Contrast-Detail Scores vs. % Vendor AGD
For 6 cm Breast Thickness
Contrast-Detail Scores vs. % Vendor AGD
For 6 cm Breast Thickness
GE 2000D - Con/Auto
16.0
GE 2000D - Std/Auto
15.0
GE 2000D - Con/Auto
15.0
GE DS - Con/Auto
14.0
GE DS - Std/Auto
13.0
GE Essential Con/Auto
GE Essential Std/Auto
Lorad
12.0
11.0
Siemens - Mo/Mo
10.0
Contrast-Detail Scores
Contrast-Detail Scores
16.0
GE 2000D - Std/Auto
14.0
Rh/Rh – 29 kVp
13.0
Rh/Rh – 31 kVp
12.0
11.0
10.0
Siemens - W/Rh
9.0
9.0
0%
50%
100%
150%
200%
0%
250%
50%
100%
Slide 111
200%
250%
Slide 112
Contrast-Detail Scores vs. % Vendor AGD
For 6 cm Breast Thickness
Contrast-Detail Scores vs. % Vendor AGD
For 6 cm Breast Thickness
16.0
GE DS - Con/Auto
GE DS - Std/Auto
15
14
Rh/Rh – 29 kVp
13
Rh/Rh – 29 kVp
12
11
10
9
Contrast-Detail Scores
16
Contrast-Detail Scores
150%
% Vendor AGD
% Vendor AGD
GE Essential - Con/Auto
15.0
Rh/Rh – 29 kVp
GE Essential - Std/Auto
Rh/Rh – 29 kVp
14.0
13.0
12.0
11.0
10.0
9.0
0%
50%
100%
150%
% Vendor AGD
200%
250%
0%
50%
100%
150%
200%
250%
% Vendor AGD
28
Slide 113
Slide 114
Contrast-Detail Scores vs. % Vendor AGD
For 6 cm Breast Thickness
16.0
Contrast-Detail Scores
Lorad
15.0
Vary kVp
Mo/Mo – 29 kVp
Siemens - Mo/Mo 28 kVp
Siemens - W/Rh 28 kVp
14.0
13.0
12.0
11.0
10.0
9.0
0%
50%
100%
150%
200%
250%
% Vendor AGD
Slide 115
Slide 116
Summary
Contrast-Detail Scores vs. kVp
At Matched Average Glandular Doses to Film and Digital
GE 2000D - Mo/Mo - Film
GE 2000D - Rh/Rh - Film
16.0
GE 2000D - Mo/Rh - Digital
15.0
GE DS - Mo/Mo - Film
4.0
GE DS - Mo/Rh - Film
3.5
GE 2000D - Mo/Mo - Digital
Average Glandular Dose vs. Breast Thickness
Contrast-Detail Score vs. Breast Thickness
GE 2000D - Rh/Rh - Digital
17
GE 2000D - Con/Auto
GE DS - Mo/Rh - Digital
GE DS - Rh/Rh - Digital
GE Essential - Mo/Mo - Film
12.0
GE Essential - Mo/Rh - Film
11.0
GE Essential - Mo/Mo - Digital
10.0
GE Essential - Rh/Rh - Digital
GE Essential - Rh/Rh - Film
GE Essential - Mo/Rh - Digital
Lorad - Mo/Mo - Film
Lorad - Mo/Rh - Film
9.0
3.0
GE Essential - Std/Auto
2.5
Lorad - Auto-Filter
Contrast-Detail Score
GE DS - Mo/Mo - Digital
13.0
16
GE DS - Std/Auto
GE DS - Rh/Rh - Film
14.0
AGD (mGy)
Contrast-Detail Scores
GE 2000D - Mo/Rh - Film
17.0
Siemens - OpComp - W/Rh
2.0
1.5
1.0
15
14
13
12
GE 2000D - Con/Auto
11
GE DS - Std/Auto
10
0.5
9
0.0
8
GE Essential - Std/Auto
Lorad - Auto-Filter
Siemens - OpComp - W/Rh
0
2
4
6
Breast Thickness (cm)
8
10
0
2
4
6
8
10
Breast Thickness (cm)
Lorad - Mo/Mo - Digital
Lorad - Mo/Rh - Digital
8.0
Siemens - Mo/Mo - Film
23 25 27 29 31 33 35 37 39 41 43 45 47 49
Siemens - Mo/Mo - Digital
Siemens - W/Rh - Film
Siemens - W/Rh - Digital
kVp
29
Slide 117
Conclusions
Conclusions
What’
What’s responsible for different image
Better calibration files on some digital
manufacturer’
manufacturer’s systems
quality scores?
– Better calibration files on some digital
manufacturer’
manufacturer’s systems
fewer artifacts
vs.
– Different breast doses
– Different postpost-processing algorithms
Slide 119
Slide 120
Conclusions
Dose
Conclusions
Image Quality
– Digital in general has higher image quality scores
– Digital in general has lower doses than film
– For the same phantom, dose varies widely by mode
– For the same phantom, image quality varies widely by
mode
– Breast thickness has big effect on image quality
– Breast thickness makes big contribution to dose
– Dose by thickness tracks pretty well across all vendors
– Dose is affected by selection of mode being used
– Image quality by thickness tracks pretty well across all
vendors
– Image quality is affected by selection of mode being used
30
Slide 121
Slide 122
Conclusions
Conclusions
Still to be done
Take home messages
– Phantom images scored by several readers
– Dose makes more of a difference on image quality than
– Analyzed for statistical differences and trends
kVp
– Digital has lower doses than Film
What I didn’
didn’t mention
– Some systems may set dose to low for their
– Viewing conditions
– Vendor QC
recommended mode
– Monitor calibration
– Pay attention to what mode is used clinically
– SNR Data
Slide 123
Slide 124
Conclusions
Take home messages
Thank You
– Digital mammography needs a more sensitive phantom
– There is a wide range of image quality scores and doses
across FFDM systems
– Evaluate systems the way they are to be used clinically
– Manufacturer only recommends AEC mode, user
ultimately decides how the system is to be used
31