DIGITAL vs. FILM-SCREEN MAMMOGRAPHY
by EDWARD L. NICKOLOFF, D.Sc. DEPARTMENT OF RADIOLOGY COLUMBIA UNIVERSITY NEW YORK, N.Y. 10032
GOALS
• COMPARE THE VARIOUS DIGITAL SYSTEMS TO FILM-SCREEN SYSTEMS • DETERMINE SOME KEY DIFFERENCES BETWEEN THE SYSTEMS: IMAGE QUALITY, RADIATION DOSES, OPERATIONAL FEATURES & PHYSICS TESTING ISSUES • IDENTIFY ADVANTAGES & DIS-ADVANTAGES OF EACH MAMMOGRAPHY SYSTEM
THE EQUIPMENT
GE 2000-D DIGITAL
GE DMR FILMSCREEN
THE EQUIPMENT
LORAD CCD DIGITAL -SELENIA SIMILAR
LORAD M IV PLATINUM FILM-SCREEN
Fischer Imaging SenoScan FFDM System
DIGITAL MAMMOGRAPHY DETECTORS
GE 2000-D IGITAL MAMMOGRAPHY UNIT
Contact Leads For Read-Out Electronics Contact Fingers Amorphous Silicon Array
Scintillator (CsI)
Glass Substrate
FROM GE
LORAD CCD DIGITAL MAMMOGRAPHY UNIT
CARBON FIBER RECEPTOR SURFACE CsI
SCINTILLATOR
40 m LIGHT PIPE LIGHT PIPE LIGHT PIPE
10 m CCD CCD
CIRCUIT BOARD
AMORPHOUS SELENIUM DETECTORS
AMORPHOUS SELENIUM DETECTOR
READOUT
VOLTAGE
-- --
IONIZATION
DETECTOR ASSEMBLY MOVES IN AN ARC
Cesium Iodide
CsI SCINTILLATOR
Fiber Optic Plate 4 CCD’s
Fischer Imaging Corporation
WORK LOAD EFFICIENCY COMPARISON
Digital Mammography - Productivity
Conventional Exam With Film
1:30 Patient in Room First Exposure 4:45 Last Exposure 2:15 First Film Out of Processor 8% Repeat Rate :05 Last Film Out of Processor :10 QC :30 Release Patient
12:10 min
Exam Using Digital Detector
1:30 Patient in Room First Exposure & QC 1:45 Last Exposure & QC :45 Release Patient / Send Study To WS
4:00 min
Minimal Repeats
Estimated > 60% Reduction In Exam Times From GE
MY EFFICIENCY ESTIMATE
• FILM-SCREEN SYSTEMS
– TALK WITH PATIENT & POSITION FOR 4 FILMS = 6 MINUTES – MARK CASSETTES WITH I.D. & BAR CODE STUDY = 1 MINUTES – WALK TO PROCESSOR WITH CASSETTES = 2 MINUTES – INSERT CASSETTES IN DAYLIGHT LOADER COUPLED TO PROCESSOR & WAIT FOR PROCESSED FILMS = 7 MINUTES – QC FILM & PLACE IN JACKET = 1 MINUTES – TALK TO PATIENT, RADIOLOGIST AND WAIT TO DRESS & LEAVE = 2 – 4 MINUTES
• TOTAL = 20 MINUTES
MY EFFICIENY ESTIMATE
• DIGITAL MAMMOGRAPHY
– TALK WITH PATIENT & POSITION FOR 4 IMAGES = 6 MINUTES – KEY IN COMPUTER INFORMATION & I.D. & BAR CODE STUDY = 2 MINUTES – REVIEW IMAGES AND TRANSMIT TO RADIOLOGIST REVIEW STATION = 2 MINUTES – TALK TO PATIENT, RADIOLOGIST AND WAIT TO DRESS & LEAVE = 2 - 4 MINUTES
• TOTAL = 13 MINUTES
COMMENTS ON RELATIVE EFFICIENCY
• ONE DIGITAL MAMMO CAN EASILY REPLACE TWO FILM-SCREEN MAMMO UNITS • ROUTINE RT QC IS BETTER … IN COMPUTER • FEW DRIFTS WITH DIGITAL UNIT • FEWER REPEATS WITH DIGITAL UNIT
RADIATION DOSES FOR MAMMOGRAPHY UNITS
COMPARISON OF RADIATION DOSES (GE DMR FILM-SCREEN vs. GE 2000-D DIGITAL)
400
AVERAGE GLANDUALAR DOSE (mRADS)
350 300 250 200 150 100 50 0 2 3 4 5 6 THICKNESS OF BR-12 (cm)
DIGITAL AOP STANDARD FILM AoP STANDARD
7
DIGITAL AoP CONTRAST FILM AoP CONTRAST
DIGITAL AoP DOSE FILM AoP DOSE
COMPARISON OF RADIATION DOSES (LORAD FILM-SCREEN WITH DIGITAL)
600
AVERAGE GLANDUALAR DOSE (mRADS)
500 400 300 200 100 0 2 3 4 5 6 THICKNESS OF BR-12 (cm) 7 8
CCD DIGITAL AT 28 kVp FILM AUTO-FILTER 200 mAs LIMIT FILM AUTO-TIME
CCD DIGITAL WITH kVp ADJUST FILM AUTO-kVp AUTO-mAs LORAD SELENIA DIGITAL
SUMMARY OF MAMMO DOSES
• IN GENERAL, THE RADIATION DOSE FOR FILM-SCREEN & DIGITAL MODES WERE SIMILAR • HOWEVER, DIGITAL DETECTORS ARE VERY EFFICIENT EVEN WITH HIGH ENERGY XRAYS... SO HIGHER KVP’s AND FILTRATION WERE USED.... RESULTING IN SLIGHTLY LOWER RAD DOSES • FILM CONTRAST MODES DELIVERED MORE RAD DOSE BECAUSE IT USED LOWER KVP’S
SUMMARY OF RADIATION DOSES
• BEWARE, UNLIKE FILM-SCREEN, DIGITAL DETECTORS CAN OPERATE OVER A WIDE RANGE OF RAD DOSES & NOISE IMPROVES AT HIGHER DOSES • DIGITAL IMAGES WITH HIGHER RADIATION DOSES LOOK BETTER • PHYSICISTS MUST GUARD AGAINST ABUSE !!
GOALS OF AUTO AEC SYSTEMS
• UTILIZE EXPOSURE TIMES <2.5 SEC • UTILIZE Dg < 300 mrads FOR < 4.2 cm COMPRESSED BREAST THICKNESS • INCREASE Dg < 150 mrads @ +1 cm • USE HIGHER kVp’S FOR THICKER / DENSE BREASTS • MAINTAIN “ESE” < 10,000 mR • IF AVAILABLE, USE Rh ANODE AND/OR FILTER FOR THICKER / DENSE BREASTS
AUTOMATED AEC SYSTEMS
• USE DYNAMIC MEASUREMENTS DURING AN EXPOSURE TO SET PARAMETERS
– ANODE – FILTER – kVp – mAs
• INVOLVES LOOK-UP TABLES (LUT) BASED UPON PATIENT BREAST
TECHNIQUES BASED UPON AVERAGE GLANDULAR DOSE MEASUREMENTS IN BR-12
CO M PR ES SED BR E AS T THIC K NES S (C M ) X -R AY TU B E P O TEN TIAL (kV p)
< 3 3 - 4 4 - 5 5 - 6 6 - 7 7 - 8
25 26 27 28 29 31
-
26 27 28 29 30 32
COMPARISON OF kVp & ANODE / FILTER FOR FILM-SCREEN vs. DIGITAL GE MAMMO IN AoP CONTRAST MODE
35
kVp UTILIZED
30
25
20 0 1 2 3 4 5 6 7 8 THICKNESS OF BR-12 (cm)
FILM Mo/Mo DIGITAL Mo/Mo FILM Mo/Rh DIGITAL Mo/Rh FILM Rh/Rh DIGITAL Rh/Rh
COMPARISON OF RADIATION DOSES (GE DMR FILM-SCREEN vs. GE 2000-D DIGITAL )
500 450
AVERAGE GRANDULAR DOSE (mRADS)
400 350 300 250 200 150 100 50 0
2 3 4 5 6 THICKNESS OF BR-12 (cm)
DIGITAL AoP CONTRAST FILM AoP CONTRAST
7
8
COMPARISON OF kVp & ANODE / FILTER FOR FILM vs. DIGITAL GE IN AoP STANDARD
34 32
kVp UTILIZED
30 28 26 24 22 20 0 1 2 3 4 5 THICKNESS BR-12 (cm)
FILM Mo/Rh DIGITAL Mo/Rh
6
7
8
FILM Mo/Mo DIGITAL Mo/Mo
FILM Rh/Rh DIGITAL Rh/Rh
COMPARISON OF RADIATION DOSE (GE DMR FILM-SCREEN vs. GE 2000-D DIGITAL)
450
AVERAGE GLANDULAR DOSE (mrads)
400 350 300 250 200 150 100 50 0 2 3 4 5 6 PHANTOM THICKNESS (cm) 7 8
DIGITAL AOP STANDARD
FILM AoP STANDARD
AutoCell:
Breast w/Photo Cells
DDD
Digital vs FilmScreen
SUMMARY OF AEC SYSTEMS
• DIGITAL SYSTEMS USE HIGHER kVp’S THAN FILM-SCREEN SYSTEMS • DIGITAL SYSTEMS USE ANODE / FILTERS WHICH PROVIDE HIGHER ENERGY XRAYS • DIGITAL SYSTEM ARE MORE FLEXIBLE WITH AEC SENSOR AREA • DIGITAL SYSTEM ARE EXPECTED TO BE SLIGHTLY LESS DOSE • DIGTIAL SYSTEMS CAN BE ADJUSTED FOR A RANGE OF DOSES
HIGH CONTRAST SPATIAL RESOLUTION
WHAT AFFECTS HIGH CONTRAST SPATIAL RESOLUTION?
•
X-RAY TUBE FOCAL SPOT SIZE – DIMENSION – GEOMETRY … MAGNIFICATION, SID – L = M / [ (M -1) x f ] – INTENSITY DISTRIBUTION • IMAGE RECEPTOR INHERENT LIMITS – FILM-SCREEN – DIGITAL SYSTEM
CONTACT FILM-SCREEN MAMMOGRAPHY SPATIAL RESOLUTION AS A FUNCTION OF BREAST THICKNESS & FOCAL SPOT SIZE
25
OVERALL SPATIAL RESOLUTION (LP/mm)
20 15 10 5 0 0 1 2 3 4 5 6 7 8 9 10 BREAST THICKNESS (CM)
F=0.3 mmf=0.6mm f=0.6 mm f=1.2 mm
1.5x MAG FILM-SCREEN MAMMOGRAPHY SPATIAL RESOLUTION AS A FUNCTION OF BREAST THICKNESS & FOCAL SPOT SIZE
18 16 14 12 10 8 6 4 2 0 0 1 2 3 4 5 6 7 8 9 10 BREAST THICKNESS (CM)
F=0.1 mm f=0.3 mm f=0.6 mm
OVERALL SPATIAL RESOLUTION (LP/mm)
6
CONTACT DIGITAL MAMMOGRAPHY SPATIAL RESOLUTION AS A FUNCTION OF BREAST THICKNESS & FOCAL SPOT SIZE
OVERALL SPATIAL RESOLUTION (LP/mm)
5 4 3 2 1 0 0 1 2
F=0.3 mm
3 4 5 6 7 BREAST THICKNESS (cm)
f=0.6 mm
8
9
10
f=1.2 mm
1.5x MAG DIGITAL MAMMOGRAPHY SPATIAL RESOLUTION AS A FUNCTION OF BREAST THICKNESS & FOCAL SPOT SIZE
9 8 7 6 5 4 3 2 1 0
OVERALL SPATIAL RESOLUTION (LP/mm)
0
1
2
F=0.1
3 4 5 6 7 8 BREAST THICKNESS (cm)
F=0.3
9
F=0.6
10
PIN-HOLE CAMERA OF X-RAY FOCAL SPOT
HIGH CONTRAST SPATIAL RESOLUTION (GE DMR FILM-SCREEN vs GE 2000-D DIGITAL)
18
SPATIAL RESOLUTION (LP/mm)
16 14 12 10 8 6 4 2 0 0 1 2 3 4 5 6 7 8 9 PHANTOM THICKNESS (cm)
DIGITAL MEASURED THEORY FILM-SCREEN FILM MEASURED THEORY DIGITAL
HIGH CONTRAST SPATIAL RESOLUTION (LORAD MIV PLATINUM FILM-SCREEN vs LORAD CCD & SELENIA DIGITAL)
18
SPATIAL RESOLUTION (LP/mm)
16 14 12 10 8 6 4 2 0 0 1 2 3 4 5 6 7 8 PHANTOM THICKNESS (cm)
CCD DIGITAL MEASURED THEORY FILM-SCREEN SELENIA DIGITAL MEASURED FILM MEASURED THEORY CCD DIGITAL THEORY SELENIA DIGITAL
SUMMARY OF HIGH CONTRAST SPATIAL RESOLUTION
• DIGITAL MAMMOGRAPHY UNITS HAVE LOWER SPATIAL RESOLUTION THAN THE FILM SCREEN UNITS ( 5 – 8 LP/mm) • MEASURED DIGITAL RESOLUTION WAS 30% TO 60% OF FILM-SCREEN RESOLUTION (15 – 20 LP/mm) • SPATIAL RESOLUTION DOES NOT TELL THE WHOLE STORY!
Flat Panel-Light Sensor
scan line FET
Very High Fill Factor
Fill Factor= Sensitive Area
Pitch x Pitch
data line
Pitch
FROM GE
Pitch
IMAGE RECEPTOR DEPENDENCE
• FILM-SCREEN SYSTEMS – DESIGN OF SCREEN (THICKNESS, DYES, REFLECTIVE LAYER, PHOSPHOR CRYSTALS) – TYPICALLY….. 15 TO 25 LP/mm • DIGITAL SYSTEMS – PITCH ( 25m To 100m ) – L = 1 / (2 x PITCH) – TYPICALLY …… 5 TO 10 LP/mm
LOW CONTRAST DETECTABILITY
LOW CONTRAST DETECTABILITY
RMI 180 C-D PHANTOM
• NINE HOLE DEPTHS RANGE FROM 0.062 mm TO 1.0 mm • SUBJECT CONTRAST = EXP (-m T) • TEN HOLE DIAMETERS RANGE FROM 0.312 mm TO 7.07 mm
COMPARISON OF LOW CONTRAST DETECTION OF FILM-SCREEN vs. DIGITAL MAMMO SYSTEMS ---- GE 100
% 0F LOW CONTRAST TARGET VISUALIZED (%)
90 80 70 60 50 40 30 20 10 0 0 2 4 6 8 ACRYLIC THICKNESS OF PHANTOM (cm)
GE 2000-D DIGITAL Linear (GE 2000-D DIGITAL) GE DMR+FILM Linear (GE DMR+FILM)
10
COMPARISON OF LOW CONTRAST DETECTION OF FILM-SCREEN vs. DIGITAL MAMMO SYSTEMS --- LORAD 100
% OF LOW CONTRAST TARGET VISUALIZED (%)
90 80 70 60 50 40 30 20 10 0 0 2 4 6 8 ACRYLIC THICKNESSOF PHANTOM (cm)
LORAD CCD DIGITAL Linear (LORAD CCD DIGITAL) Linear (LORAD SELENIA DIGITAL)
10
LORAD MIV + FILM LORAD SELENIA DIGITAL Linear (LORAD MIV + FILM)
DIGITAL SYSTEM SIGNAL-TO-NOISE RATIO (SNR)
DIGITAL MAMMOGRAPHY SYSTEMS SNR vs. RADIATION EXPOSURE 1000
SIGNAL-TO-NOISE RATIO (SNR)
y = 12.216x 100
0.5374
y = 9.759x 10
0.5552
1 0.1 1
LORAD DIGITAL Power (LORAD DIGITAL)
10
100
GE 2000-D DIGITAL Power (GE 2000-D DIGITAL)
1000
ENTRANCE RADIATION EXPOSURE (mR)
OTHER CONSIDERATIONS
DYNAMIC RANGE CONSIDERATIONS
FILM-SCREEN CHARACTERISTIC CURVE -GE DMR WITH KODAK MIN-R 2000
4 3.5 3 2.5 2 1.5 1 0.5 0 0 0.5 1 1.5 2 2.5 LOG [EXPOSURE (mR)]
USEFUL RANGE = 78.9 mR / 8.9 mR = 8.7 28 kVp, 3 cm ACRYLIC, INTO BUCKY KODAK 270 RA PROCESSOR WITH KODAK CHEMISTRY FILM DYNAMIC RANGE
FILM DENSITY [O.D.]
FILM-SCREEN CONTRAST GRADIENT -GE DMR WITH KODAK MIN-R 2000 FILM
4.5 4
FILM CONTRAST GRADIENT
3.5 3 2.5 2 1.5 1 0.5 0 0 0.5 1 1.5 2 2.5 3 3.5 4 FILM OPTICAL DENSITY (O.D.) FILM-SCREEN DYNAMIC
DYNAMIC RANGE OF DIGITAL MAMMOGRAPHY -- GE 2000-D DIGITAL 10000
RELATIVE DIGITAL SIGNAL
DIGITAL DYNAMIC RANGE
1000
100
USEFUL RANGE = > 260 mR / 0.1 mR = > 2600
10
1
0.1
1
10
100
1000
ENTRANCE EXPOSURE ( mR)]
LORAD 2000 DIGITAL MAMMOGRAPHY: SIGNAL vs RADIATION EXPOSURE
100000
DIGITAL DYNAMIC RANGE
SIGNAL (relative units)
10000
1000
100
USEFUL RANGE = 130 mR / 0.1 mR = 1300
10
1 0.1 1 10 100 1000 10000 ENTRANCE EXPOSURE (mR)
TISSUE EQUALIZATIONS
Thickness Compensation
GOAL: EQUALIZATION OF DENSITIES TO SKIN LINE. Algorithm: 1. Estimate the thickness of the breast at all pixels 2. Add a nearly-equivalent thickness [of water] to raise gray level.
Line-Profile through Breast
Original
TC-Processed
Compensation Thickness Breast Thickness
{
{
FROM GE
Breast 3 RMLO no TE
FROM GE
Breast 3 - RMLO TISSUE EQUILIZATION
COMPUTERS IN MAMMOGRAPHY
• • • • • • Computer Aided Diagnosis (CAD) Data Compression PACS Image Enhancement Dual Energy Subtraction Tomosynthesis
COMPUTER AIDED DIAGNOSIS
FROM : L CLARKE et al., Acad. Radiol. 1997
GE C ompany Proprietary - Not for External Distribution
October 6, 1998
COMPUTER-AIDED DETECTION
Applications • Automated detection • Automated classification
TM Current ImageCheckerTM Performance • Flagged 85% of overlooked cancers • Flagged 83% of diagnosed cancers
Increase the sensitivity for cancer detection when combined with the radiologist
From GE
DISPLAY MONITORS
VDT verses VIEWBOX
• BRIGHTNESS
– VDT = 200 TO 400 LUX – VIEWBOX = 3000 - 4000 LUX
• VDT HAS WINDOWS & LEVELS FOR VARIABLE CONTRAST / BRIGHTNESS • COLOR TINT DIFFERENCES • CALIBRATION REQUIREMENTS & DRIFT
MONITOR CHARACTERISTICS
THE SMPTE TEST PATTERN WITH 13 SHADES OF GRAYSCALE PATCHES WAS UTILIZED TO MEASURE THE GRAYSCALE “LUT “.
SMPTE TEST PATTERN
VISIBILITY OF BARS EACH SHADE OF GRAY BLOCK DIFFERENT
LINES STRAIGHT
0% / 5% CONTRAST BLOCK VISIBLE
100% / 95% CONTRAST BLOCK VISIBLE
IMAGE MANIPULATION
LMLO - Full Zoom, no Invert
FROM GE
LMLO - Full Zoom, Invert
FROM GE
UNSHARP MASK ENHANCEMENT
DENSITY DENSITY
ORIGINAL
SMOOTH
DISTANCE
DISTANCE
DENSITY
ORIGINAL - SMOOTH
DENSITY
ORIGINAL + 4 DIFFERENCE
DISTANCE
DISTANCE
Tomosynthesis
Series of low-dose images used to reconstruct tomography images at any level
FROM GE
ADVANTAGES OF DIGITAL TECHNOLOGY
• LARGE DYNAMIC RANGE -- NOT TOO DARK OR TOO LIGHT • WINDOW & LEVEL CAPABILITY • IMAGE PROCESSING • FAST RETREIVAL & LAYOUT • LOWER REPEAT RATES • LESS PATIENT RADIATION DOSE • COMPUTER AIDED DIAGNOSIS (CAD) • TELERADIOLOGY CAPABILITY
ADVANTAGES OF DIGITAL TECHNOLOGY
• BETTER SENSITIVITY & EARLY DETECTION • COMBINE WITH RIS/HIS DISPLAY • COMPACT IMAGE STORAGE • BETTER PATIENT TRACKING & BILLING • SERVICES TO REMOTE LOCATIONS • LATITUDE & CONTRAST
DIS-ADVANTAGES OF DIGITAL TECHNOLOGY
• • • • • HIGH INITIAL COSTS HIGHER MAINTENANCE COSTS TECHNOLOGY OBSOLENCE ISSUES HIGH QUALITY SUPPORT PERSONNEL NEED FOR NETWORK & NETWORK SUPPORT • LOWER SPATIAL RESOLUTION • NETWORK TRANSFER SPEED LIMITS • MORE COMPLEX VIEWING PROCESS
DIS-ADVANTAGES OF DIGITAL TECHNOLOGY
• ENTERING INCORRECT ID# CAN MISPLACE STUDIES & BILLING • PACS/NEMA COMPATABLE HARDWARE MAY NOT BE INSTALLED • INTRA-VENDOR COMPATABILITY • POWER / COMPONENT FAILURES • LONG-TERM DATA RETRIEVAL DELAYS • LACK OF INSTANT FLUORO REVIEW
CONCLUSIONS
• DIGITAL IS MORE TIME EFFICIENT • RAD DOSE CAN BE SIMILAR • FILM-SCREEN HAS BETTER HIGH CONTRAST RESOLUTION & DIGITAL BETTER LOW CONTRAST SENSITIVITY • DIGITAL MAMMO SYSTEMS ARE MORE ROBUST:
– WIDER DYNAMIC RANGE – WINDOW / LEVEL ADJUSTMENT – COMPUTER IMAGE ENHANCEMENT
ARRET