8/2/2016

Advanced X-ray Breast Imaging

Andy Smith, Ph.D.
Vice President, Image Research
Hologic 2016. 1

Learning Objectives

• Tomosynthesis
• Theory, QC
• C-View™ Synthetic 2D Imaging
• Theory, QC
• I-View™ Dual Energy Mammography
• Theory, QC
• Tomo-Guided Biopsy
• Theory, QC

Hologic 2016. 2

Hologic Breast Tomosynthesis

Tube moves in a 15o arc

• 15 low dose images are acquired. Total dose ~ 1 FFDM
• 3.7 second continuous motion sweep
• X-rays are pulsed on and off

Images are reconstructed into 1 mm slices

• ~ 100 micron in-plane resolution

Hologic 2016. 3

1
 8/2/2016

X-Ray Generation

• Tungsten (W) Anode

X-Ray Tube • LFS: 0.3 mm; SFS: 0.1 mm

• FFDM: 50 μm Rh; 50 μm Ag
X-Ray Filtration • Tomo: 0.7 mm Al
• I-View: 0.3 mm Cu

• 200 mA max for LFS

• 50 mA max for SFS
X-Ray Generator • Max mA varies with kVp
• mA adjusts to target time range

Hologic 2016.

Selenia Dimensions: Technicals

Conventional 2D Imaging Tomosynthesis Imaging
• a-Se detector, 24×29 cm area • a-Se detector, 24×29 cm area

• 70 μm pixel size • 140 μm pixel size

• Rh and Ag filters • Al filter

• 20-39 kVp • 20-49 kVp

• HTC grid in contact mode • No anti-scatter grid

• No grid in magnification mode • Moving tube, 15° sweep

• 15 projections

• Moving detector

• 3.7 seconds acquisition

• Reconstruction
• ~100 μm pixel size
• 1 mm slice spacing Hologic 2016.

System Configurations

Selenia Dimensions Screening

(Avia)
2D Screening Configuration
Selenia Dimensions Mobile Upgradable to Stereo/Tomo
2D or 2D/3D Configurations

Selenia Dimensions 2D/3D

AWS 5000/6000/8000/9000
Configuration

Selenia Dimensions 2D
AWS 5000/6000/8000/9000
Configuration

Hologic 2016.

2
 8/2/2016

Clinical Utility of Combo (Tomo+2D) imaging

(FDA approved mode for tomo in screening)

Tomo (or DBT)

Visualization of masses and architectural distortions

2D
Rapid review of calcification clusters
Comparison to 2D priors and left/right asymmetry
CAD operates on the 2D images

Clinical Results
2D+3D superior to 2D in cancer detection and false positive rate (Friedewald, …)

Hologic 2016.

C-View™ Image

Provides a 2D-like image

• Helps the radiologist facilitate review
– Quick overview of breast
– Compare to 2D Priors
– Maintain familiar workflow

Does not require additional x-ray exposures

• Original Combo imaging required both 2D and Tomo acquisitions

FDA approved to replace the 2D images in tomo screening exam

Clinical Studies (Skaane, Bernardi, Conant): C-View+3D comparable to 2D+3D

Hologic 2016. 8

C-View ™ (Synthesized 2D) ~60 Tomosynthesis Slices

How does it work?

• Perform a standard tomosynthesis scan
• Reconstruct tomosynthesis slices
• Generate the 2D image
• ‘Intelligent’ MIP Tomo  2D
Algorithm
• No need for 2D exposure
• Shorter compression time

C-View 2D Image

Hologic 2016.

3
 8/2/2016

C-View™ Image and FFDM

• Not intended to have identical appearances

• Small but distinctive differences
- Important details from the tomosynthesis slices are
preserved and enhanced
- Architectural distortions and calcifications may be
more conspicuous

2D C-View™ image

2D C-View™ image

4
 8/2/2016

2D C-View™ image

Conspicuity of the calcifications in dense tissue can be

higher with C-View™ images.

2D C-View™ image

C-View™ images can maintain details even in

dense tissue

RCC LCC

5
 8/2/2016

RMLO LMLO

RMLO LMLO

2D Tomosynthesis slices

6
 8/2/2016

2D Tomosynthesis Slice
Invasive Ductal Carcinoma

2D C-View™ image

Contrast Enhanced 2D Mammography – I-View™

Hologic Proprietary and Confidential

7
 8/2/2016

Dual Energy 2D Imaging

Designed to image an iodine contrast agent

Two exposures are made in rapid sequence:
1. Low kV (normal mammogram, ~28-30 kV, Rh/Ag filters). Below Iodine’s k-edge of 33 keV
2. High kV (~45-49 kV, Cu filter). Above Iodine’s k-edge.

Subtraction gives a 2D iodine contrast image

Repeat as desired
Imaging window ends after ~6 minutes due to contrast redistribution

Hologic Proprietary and Confidential Hologic 2016.

Dual-Energy Subtraction
• You cannot see the iodine contrast drug in high or low kV images
• But you can when you subtract the images:
Subtraction = high kV - low kV

Low kV High kV Subtraction

• Only the low kV and subtraction images are viewed

Hologic Proprietary and Confidential Hologic 2016.

CE2D – Dual Energy 2D

Low kV, High kV images acquired for each view
Views can be any order
Ipsilateral Breast Contralateral Breast

MLO CC CC MLO

H H H H
E ne rgy

L L L L

inject

0s ~120 s ~ 180 s ~ 240 s ~ 300 s ~ 480 s

Hologic Proprietary and ConfidentialHologic 2016.

8
 8/2/2016

Dual Energy 2D Combo Imaging

Three exposures are made in rapid sequence:

1. Low kV tomosynthesis scan (~30 kV, Al filter)

2. Low kV (normal mammogram, 28-30 kV, Rh/Ag filters)

3. High kV (~45-49 kV, Cu filter)

Subtraction gives a 2D contrast image

Tomo image is co-registered to the contrast 2D image
Repeat as desired
Imaging window ends after ~6 minutes due to contrast redistribution

Hologic Proprietary and Confidential Hologic 2016.

Combo CE2D – Dual Energy 2D with Tomosynthesis

Tomo, Low kV, High kV images acquired for each view

Views can be any order

Ipsilateral Breast Contralateral Breast

MLO CC CC MLO

H H H H
Energy

L L L L

inject

0s ~120 s ~180 s ~240 s ~300 s ~480 s

Hologic Proprietary and ConfidentialHologic 2016.

Example Image
Courtesy of Andrea Woodroof, Kentucky Breast Care

Hologic Proprietary and Confidential

9
 8/2/2016

CE2D for Discordant Findings

• 52 y.o. female, presented for screening
• 0.8 cm spiculated mass in the left axillary tail noted on both 2D
and 3D
• Focal area of possible distortion noted on 3D

Hologic Proprietary and Confidential

CE2D for Discordant Findings

Hologic Proprietary and Confidential

CE2D for Discordant Findings

• Area of distortion couldn’t definitively be correlated on
ultrasound for US biopsy

Hologic Proprietary and Confidential

10
 8/2/2016

CE2D for Discordant Findings

Hologic Proprietary and Confidential

CE2D for Discordant Findings

• Area of distortion couldn’t definitively be correlated on
ultrasound for US biopsy
• The enhancing mass on CE2D was correlated with the tomo
distortion
• Tomo-guided biopsy ensued

Hologic Proprietary and Confidential

Quality Control Procedures: Digital Breast Tomosynthesis

Hologic 2016.

11
 8/2/2016

Image Acquisition Modes

• Conventional
• Acquires 2D images only

• Tomo
• Acquires tomosynthesis images only

• TomoHD
• Acquires tomosynthesis images only
• Produces C-View images

• Combo
• Acquires 2D images
• Acquires tomosynthesis images

• ComboHD
• Acquires 2D images
• Acquires tomosynthesis images
• Produces C-View images Hologic 2016.

Image Acquisition Modes (I-View™)

• 2D Contrast Acquisition
• 2D low-energy image
• 2D high-energy image

• Combo 2D Contrast Acquisition

• Tomosynthesis projections
• 2D low-energy images
• 2D high-energy images

Hologic 2016.

QC Modes

• Quality control procedures test

• Conventional
• Tomo
• Combo
• I-View™

• The following modes do not

require separate QC testing
• C-View™
• TomoHD
• ComboHD

Hologic 2016.

12
 8/2/2016

DBT QUALITY CONTROL

DBT
Quality Control

Hologic 2016.

Medical Physics
Medical Equipment Evaluation (MEE) Testing

12 tests to be performed by
the medical physicist

5 extra tests that are usually

performed by technologist

Hologic 2016.

Medical Physics
QC Tests

12 tests to be performed by
the medical physicist

7 of them have DBT

components/requirements

Hologic 2016.

13
 8/2/2016

Medical Physics
QC Tests

3 of them have I-View

components/requirements

Hologic 2016.

Technologist QC
Tests

12 tests to be performed by
the technologist

3 of them have DBT

components/requirements

Hologic 2016.

Technologist QC
Tests

2 of them have I-View

components/requirements

Hologic 2016.

14
 8/2/2016

Alternative
Standard 9
• Category A
• Immediate corrective action is required
• Issue needs to be resolved before resuming imaging

• Category B
• Affects interpretation devices
• Interpretation device stays out of use until issue is resolved
• Alternative interpretation device may be used
• Imaging may continue

• Category C
• Issue needs to be resolved within a predefined time
window
• Imaging may continue up to the end of that time window

Hologic 2016.

Alternative
Standard 6
• Additional evaluations must be conducted after major
component change or repair
• Software changes are considered as major repairs
• Alternative Standard 6 allows manufacturers to:
• Evaluate software update for QC requirements
• Define a series of tests to be performed by the
technologist

Hologic 2016.

Mobile Quality Control Checks

Additional checks must be repeated

by technologist every time the
system moves and before patient
imaging
• Compression thickness
• Artifact evaluation
• Phantom image quality
• Signal-to-noise and contrast-to-noise

Hologic 2016.

15
 8/2/2016

Hologic QC Manuals

MAN-01965 MAN-01476
R008 R001

MAN-03706
R002

• http://www.hologic.com/support/dimensions-3d-breast-tomosynthesis-
dimensions-2d-full-field-digital-mammography
• http://www.hologic.com/support/selenia-digital-mammography
Hologic 2016.

Software Version and QC Manuals

MAN-01965
• Applies to software version prior to v1.8.x

MAN-03706
• Applies to software version starting at v1.8.x

Important differences
• Appendix D: CNR Correction Tables

Hologic 2016.

Tomosynthesis Option

Tomosynthesis specific
tests are marked with an
icon
Icon indicates that a
special action is required
under tomosynthesis
NOTE: When testing FFDM only,
these instructions are ignored

Hologic 2016.

16
 8/2/2016

Diagnostic Option

Diagnostic specific tests

are marked with an icon
Icon indicates that listed
action is only applicable
to systems licensed for
diagnostic use
NOTE: When testing screening-only
FFDM systems, these instructions
are ignored

Hologic 2016.

Contrast Option

Iodine contrast specific

tests are marked with an
icon
Icon indicates that listed
action is only applicable
to systems licensed for
iodine contrast imaging
NOTE: When testing systems that
are not licensed for iodine contrast,
these instructions are ignored

Hologic 2016.

Tomosynthesis-Specific Quality Control

• Tomo Geometry
Calibration
• Usually performed by
technologist

Hologic 2016. 51

17
 8/2/2016

SYSTEM DESCRIPTION AND USER INTERFACE

Quality Control
Tests

Hologic 2016.

CAUTION

Direct x-ray exposure of the image receptor may damage the receptor
The image receptor should be covered with lead or copper during testing for exposures
other than those required to qualify image quality
Following the test procedures in the Hologic QC Manual will ensure the safety of the
image receptor

Hologic 2016.

CAUTION

System must go from Detector Warming status to

All Ready status to perform image quality tests

Hologic 2016.

18
 8/2/2016

Follow the 1999 ACR Mammography Quality Control Manual

1. MAMMOGRAPHIC UNIT ASSEMBLY

EVALUATION

Hologic 2016.

Follow the Hologic Selenia Dimensions Quality Control Manual

2. COLLIMATOR ASSESSMENT

Hologic 2016.

2a. X-Ray Field to Light Field

ONLY use the 24x29 cm compression paddle

Cover the image receptor if repeated, high exposures are required (i.e. self-
developing film)

Hologic 2016.

19
 8/2/2016

2b. X-Ray Field to Image Receptor

Test with the 24x29 cm compression paddle

Test left, center and right x-ray fields with the 18x24 cm compression paddle
Use the Zero-Degree Tomo view to test under tomosynthesis
Follow the directions in the QC manual
Hologic 2016.

2c. Compression Paddle to Image Receptor

Compression paddles
Manufactured as single pieces

Do not have adjusting parts

Designed to comply with the regulations

Design assumes mild compression (~10lb) to remove play

Hologic 2016.

Follow the Hologic Selenia Dimensions Quality Control Manual

3. ARTIFACT EVALUATION

Hologic 2016.

20
 8/2/2016

Procedure Highlights
DICOM printer
Send an artificial flat field image to the printer

FFDM testing
Test all focus/filter combinations
(LFS/Rh; LFS/Ag; SFS/Rh; SFS/Ag; LFS Cu)
Preview image in full resolution

DBT testing
Test using middle projection
Preview image in full resolution

Hologic 2016.

Follow the 1999 ACR Mammography Quality Control Manual

4. KVP ACCURACY AND

REPRODUCIBILITY

Hologic 2016.

Procedure Highlights

Conventional Tomo Stationary Tomo

Cover the image receptor to protect it from radiation exposure

FFDM extends to 39 kVp; DBT extends to 49 kVp
Use the Zero-Degree Tomo mode to test beyond 39 kVp, if needed
Non-invasive meters must be calibrated to the specific filters and energy range used
Hologic Service can assist with equipment
Hologic 2016.

21
 8/2/2016

Follow the 1999 ACR Mammography Quality Control Manual

5. BEAM QUALITY ASSESSMENT—

HVL MEASUREMENT

Hologic 2016.

Procedure Highlights

Conventional Tomo Stationary Tomo

Cover the image receptor to protect it from radiation exposure

Use the Zero-Degree Tomo mode to measure HVL under DBT (Al filter)
NOTE: compression thickness should be <24cm for the system to allow exposure
Non-invasive meters must be calibrated to the specific filters and energy range used
[HVL > (kVp/100) + 0.03] in mm Al
Hologic 2016.

Follow the Hologic Selenia Dimensions Quality Control Manual

6. EVALUATION OF SYSTEM RESOLUTION

Hologic 2016.

22
 8/2/2016

Procedure Highlights
Place the line pair phantom on
top of the 4 cm acrylic block
Rotate the line pair phantom 45°
Apply 15-20 lb of compression to
avoid vibration during DBT
Use the Flat Field view (no image
processing)
Resolution guidelines:
FFDM: > 7 lp/mm @ 45°
DBT: > 3 lp/mm @ 45°
Hologic 2016.

Procedure Highlights

Hologic 2016.

Follow the Hologic Selenia Dimensions Quality Control Manual

7. AEC FUNCTION PERFORMANCE

Hologic 2016.

23
 8/2/2016

AEC Function Description

AEC modes
• Auto-Filter
• Auto-kV
• Auto-Time

AEC positions
• Auto AEC: two ~1cm2 floating sensors in 5 x 14cm 2 area
• One of seven manual positions (marked on compression paddle)

AEC function
• kVp and filter parameters are determined by compression thickness and AEC technique tables
• kVp can be adjusted upwards if the exposure time will be too long
• Starting mAs is determined from short pre-exposure targeting a specific exposure index (EI)
• Final mAs is adjusted by CNR correction factor
Hologic 2016.

Procedure Highlights
Compression
Thickness
Compression thickness must be
set using the compression
display
FFDM testing
• Range of phantom thickness
• Different operating modes (i.e. mag)
• Exposure compensation steps

DBT testing
• Range of phantom thickness
Hologic 2016.

Exposure Index (EI)

EI is defined as the digital value

of a detector element
“Raw” EI values need to be
corrected by
• Subtracting the DC offset (value of 50) 𝑅𝑂𝐼 𝑚𝑒𝑎𝑛 − 𝐷𝐶 𝑂𝑓𝑓𝑠𝑒𝑡
𝑃𝑖𝑥𝑒𝑙 𝑉𝑎𝑙𝑢𝑒 =
• Normalizing by the CNR correction 𝐶𝑁𝑅 𝐶𝑜𝑟𝑟𝑒𝑐𝑡𝑖𝑜𝑛 𝐹𝑎𝑐𝑡𝑜𝑟
factor (given in Appendix D of the
Hologic QC Manual)

Hologic 2016.

24
 8/2/2016

CNR Correction Factors, FFDM

LFS

Hologic 2016.

CNR Correction Factors, FFDM

SFS

Hologic 2016.

CNR Correction Factors, FFDM

LFS

Hologic 2016.

25
 8/2/2016

Calculation Example

Hologic 2016.

Follow the Hologic Selenia Dimensions Quality Control Manual

8. BREAST ENTRANCE EXPOSURE,

AEC REPRODUCIBILITY AND AGD

Hologic 2016.

Procedure Highlights

Wait until the image receptor

goes from Warming to
Ready status
Use ACR Phantom view to
overwrite compression
thickness to 4.2 cm

Hologic 2016.

26
 8/2/2016

Procedure Highlights
Test AGD in all three modes
• FFDM
• Tomo
• Combo
Hologic AGD recommended
dose for ACR phantom
• FFDM: 1.2 mGy
• Tomo: 1.45 mGy
Performance criteria
• AGD < 3 mGy
Hologic 2016.

Follow the Hologic Selenia Dimensions Quality Control Manual

9. RADIATION OUTPUT RATE

Hologic 2016.

Procedure Highlights

Maximum exposure time under LFS

• 2.5 sec

X-ray tube current at 28 kVp

• 160 mA

Output rate requirement for W/Rh at

4.5 cm above breast platform
support
• 230 mR/s (2.0 mGy/s air kerma)

Hologic 2016.

27
 8/2/2016

Follow the Hologic Selenia Dimensions Quality Control Manual

10. PHANTOM IMAGE QUALITY

EVALUATION

Hologic 2016.

Procedure Highlights
Wait until the image receptor
goes from Warming to
Ready status
Use ACR Phantom view to
overwrite compression
thickness to 4.2 cm

Hologic 2016.

Phantom Scoring
Score phantom on AWS display
Review image in full resolution
FFDM scoring
• 5 fibers, 4 specs, 4 masses
• Due to phantom variations a score of 4.5/4.0/3.5 is
acceptable providing SNR and high contrast
resolution tests pass

DBT scoring
• Scroll to the slice that puts the different elements
in focus
• 4 fibers, 3 specs, 3 masses
Hologic 2016.

28
 8/2/2016

Follow the Hologic Selenia Dimensions Quality Control Manual

11. SNR AND CNR MEASUREMENTS

Hologic 2016.

Procedure Highlights

Wait until the image receptor goes

from Warming to Ready status
Use ACR Phantom view to overwrite
compression thickness to 4.2 cm
ACR Phantom view allows
automatic SNR/CNR calculations
Test is performed under FFDM mode
only

Hologic 2016.

Automatic Computation

Hologic 2016.

29
 8/2/2016

Follow the Hologic Selenia Dimensions Quality Control Manual

12. DIAGNOSTIC REVIEW

WORKSTATION QUALITY CONTROL

Hologic 2016.

Procedure Highlights

The Hologic QC Manual offers an

alternative QC procedure for the
review workstation
Most review workstation offer their
own QC software and QC
procedures
Follow their QC procedures and
performance requirements

Hologic 2016.

Follow the Hologic Selenia Dimensions Quality Control Manual

13. DETECTOR GHOSTING

(TROUBLESHOOTING USE ONLY)

Hologic 2016.

30
 8/2/2016

Procedure Highlights

Test to be performed if ghosting is

noticed on clinical images
Is not required under acceptance or
annual evaluation
Wait until the image receptor goes from
Warming to Ready status
Typical reasons for failure
• Erase LED array failure

Hologic 2016.

Procedure Highlights

3 2 1

Hologic 2016.

Computation of Ghosting

𝑚𝑒𝑎𝑛𝑅3 − 𝑚𝑒𝑎𝑛𝑅2
𝐺ℎ𝑜𝑠𝑡 =
𝑚𝑒𝑎𝑛𝑅1 − 𝑚𝑒𝑎𝑛𝑅2

Hologic 2016.

31
 8/2/2016

Tomo-guided Biopsy

Hologic 2016.

How do you biopsy this lesion, occult in 2D and U/S and

MRI?

2D
Tomo

Hologic 2016. 95

Affirm™
Interventional Add-On Device Utilizing Tomosynthesis Localization

Hologic 2016.

32
 8/2/2016

Biopsy: Stereo vs. Tomo-Guided

Stereotactic Biopsy
• Take 2D scout exposure
• Acquire ±30° stereo pairs
• Use triangulation to determine x,y,z
lesion location

Tomo-guided Biopsy
• Take Tomo exposure
• Use tomo slices to determine x,y,z
lesion location

Hologic 2016. 97

Tomosynthesis procedure –
Target

Hologic 2016. 98

Tomosynthesis procedure -
Target

Targeting Lesion coordinates

Hologic 2016.

33
 8/2/2016

Prone Tomosynthesis-Guided Breast Biopsy

Hologic 2016. 100

Prone Biopsy System Technicals

• a-Se detector, 12.5 x 14.3 cm area

• 70 μm pixel size

• Tungsten anode, 200 mA max

• 20-49 kVp

• Al 0.70 mm, Ag 0.050 mm filters

• No anti-scatter grid
• 15° sweep tomo, 30° stereo

Hologic 2016. 101

Tomo-Guided Biopsy Quality Control

Refers back to 1999 ACR Stereotactic Breast Biopsy QC Manual

Hologic 2016. 102

34
 8/2/2016

Tomo-Guided Biopsy Quality Control

Refers back to 1999 ACR Stereotactic Breast Biopsy QC Manual

Hologic 2016. 103

QUALITY CONTROL PROGRAMS, HOLOGIC, INC.

Digital Mammography
QC Program:
TRAINING

Hologic 2016.

Required Training

Hologic 2016.

35
 8/2/2016

FDA Training Requirements: FFDM

FFDM training is specific to the type of user

Everybody needs FFDM training
• 8 hours for Medical Physicists
• 8 hours for Technologists
• 8 hours for Radiologists

Hologic 2016.
PRE-00023
PRE-00023

FDA Training Requirements: DBT

DBT is considered a new imaging modality

DBT requires its own training
DBT training is specific to the type of user
Everybody needs additional training
• 8 hours for Medical Physicists
• 8 hours for Technologists
• 8 hours for Radiologists

Hologic 2016.
PRE-00023
PRE-00023

Medical Physicist DBT Training

8 hours of FFDM training is required

8 hours of DBT training is required
Available training sources
• This AAPM meeting
• MTMI hands-on workshops
• Hologic
– On-line training
– Field training during system installation

Hologic 2016.
PRE-00023
PRE-00023

36
 8/2/2016

Facility Certification

2D MQSA certification through ACR or other approved accreditation body

No approved accreditation bodies DBT today
– DBT systems will be accredited under facility’s existing FFDM certification through FDA’s
Certification Extension Program
– Facility must be FFDM MQSA certified before applying for DBT extension certification

Certification Extension Program

Division of Mammography Quality and Radiation Programs
FDA/CDRH/OCER
10903 New Hampshire Avenue, WO66-4621
Silver Spring, MD 20903-0002
Phone: 301-796-5710 Fax: 301-847-8502
Hologic 2016.
PRE-00023
PRE-00023

Quality Control with new ACR Protocol/Phantom

Hologic 2016. 110

Quality Control with new ACR Protocol/Phantom

Hologic 2016. 111

37
 8/2/2016

Quality Control with new ACR Protocol/Phantom

• Designed to replace the manufacturer’s FFDM QC protocols

• Does not apply to systems with tomosynthesis options
• Does not apply to systems with dual-energy contrast options
• What about clinics with some systems with tomo and other
systems FFDM only?

Hologic 2016. 112

Thank you!

Email:
[email protected]

Hologic 2016. 113

Hologic 2016. 114

38