Fujifilm AMULET Innovality

34 Pages·2017·0.73 MB·English

by  Public Health England

e
r
t
n
e
C

g
)
n M
i
t P
a
NHS Breast Screening Programme
C
n
i C
equipment report d
r N
o
Technical evaluation of Fujifilm (AMULET
-

o y
Innovality digital breast tComoshynthesis
p

l
system a a
n r
g
o
o
i
t m
a
N m

February 2018 a
e
M
h
t

f
m
o
Public Health England leads the NHS Screening Programmes

o
s
r c
f
i
e s
y
l
b
h
a
P
l
i
a e
v h
A
t

r
o
f
Technical evaluation of Fujifilm AMULET Innovality tomosynthesis system
About Public Health England

e
Public Health England exists to protect and improve the nation’s health and wellbeing, r
t
and reduce health inequalities. We do this through world-leading science, knowledge andn
intelligence, advocacy, partnerships and the delivery of specialist public health servicees.
C
We are an executive agency of the Department of Health and Social Care, and a distinct

delivery organisation with operational autonomy. We provide government, local government,
g
)
the NHS, Parliament, industry and the public with evidence-based professionnal, scientifiMc
and delivery expertise and support. i
t P
a
Public Health England, Wellington House, 133-155 Waterloo Road, London SE1 8UG
C
n
Tel: [phone] www.gov.uk/phe
i C
Twitter: @PHE_uk Facebook: www.facebook.com/PublicHealthdEngland
r N
o
(
-
About PHE screening
o y
C h
p
Screening identifies apparently healthy people who may be at increased risk of a disease
l
a a
or condition, enabling earlier treatment or better informed decisions. National population
n r
screening programmes are implemented in the NHS on gthe advice of the UK National
o
Screening Committee (UK NSC), which makes indepoendent, evidence-based
i
t m
recommendations to ministers in the foaur UK countries. The Screening Quality Assurance
Service ensures programmes are saNfe and effemctive by checking that national standards
are met. PHE leads the NHS Sc reening Programmes and hosts the UK NSC secretariat.
a
e
M
h
www.gov.uk/topic/population-screening-programmes
t

Twitter: @PHE_Screenin g Blog: phescreening.blog.gov.uk
f
m
o
Prepared by: CJ Strudley, A Hadjipanteli, JM Oduko, KC Young

o
For queries relating to this docsument, please contact: [email protected]
The image on parge 9 is coucrtesy of Fujifilm.
f
i
e s
y
l
b
© Crown copyright h2018
a
P
You may re-use this information (excluding logos) free of charge in any format or
l
i
maedium, undeer the terms of the Open Government Licence v3.0. To view this licence,
vvisit OGL.h Where we have identified any third party copyright information you will need
A
to obtatin permission from the copyright holders concerned.

r
o
Published March 2018
f
PHE publications PHE supports the UN
gateway number: 2017807 Sustainable Development Goals
2
Technical evaluation of Fujifilm AMULET Innovality tomosynthesis system

Acknowledgements

e
The authors are grateful to the staff at Burnley Hospital for facilitating the evaluation of r
t
the unit at their site. n
e
C

g
)
n M
i
t P
a
C
n
i C
d
r N
o
(
-

o y
C h
p

l
a a
n r
g
o
o
i
t m
a
N m

a
e
M
h
t

f
m
o

o
s
r c
f
i
e s
y
l
b
h
a
P
l
i
a e
v h
A
t

r
o
f
3
Technical evaluation of Fujifilm AMULET Innovality tomosynthesis system

Contents

e
About Public Health England 2 r
t
n
About PHE screening 2
e
Acknowledgements C 3

Contents g 4
)
n M
Executive summary 5
i
t P
1. Introduction a 6
C
n
1.1 Testing procedures and performance standards for digital mammography 6
i C
1.2 Objectives d 6
r N
2. Methods o 7
(
2.1 System tested - 7
o y
2.2 Dose and contrast-to-noise ratio under AEC 9
C h
2.3 Image quality measurements 11
p

2.4 Geometric distortion and reconstruction lartefacts 12
a a
2.5 Alignment 13
n r
2.6 Image uniformity and repeatability g 13
o
2.7 Detector response o 14
i
2.8 Timings t m 14
a
2.9 Modulation transfer function (MTF) 14
N m
2.10 Local dense area 14

3. Results e a 15
M
h
3.1 Dose and contrast-to-noise ratio under AEC 15
t
3.2 Image quality m easuremen ts 20
f
m
3.3 Geometric distortion ando resolution between focal planes 22
3.4 Alignmento 25
s
3.5 Image urniformity acnd repeatability 25
3.6 Dete cftor responise 25
3.7 Timeings s 26
y
3.8 bMlTF 26
h
3.9 a Local dense area 28
P
4. lDiscussion 29
i
a e
5. Conclusions 32
v h
A
Referetnces 33

r
o
f
4
Technical evaluation of Fujifilm AMULET Innovality tomosynthesis system

Executive summary

e
The technical performance of the Fujifilm AMULET Innovality digital breast r
t
tomosynthesis system was tested in the 2 tomosynthesis modes available, Standard n
(ST) and High Resolution (HR). The mean glandular dose (MGD) to the standard e
C
breast was found to be within the remedial dose levels, except at the high (H) dose

setting in HR mode. The threshold gold thicknesses measured with the CDMAM test
g
)
object are better than the achievable level for 2D, for details of 0.2mm and anbove. M
i
t P
Technical performance of this equipment was found to be satisfactory, sao that the
C
system could proceed to practical evaluation in a screening centre. Tnhis report provides
i C
baseline measurements of the equipment performance including:d
r N
o
 dose
(
-

 contrast detail detection o y
 contrast-to-noise ratio (CNR) C h
 reconstruction artefacts, z-resolution p
l
 detector response a a
n r
 projection modulation transfer function g
o
o
i
The MGD and CNR measurements in 2tD mode wemre close to those measured and
a
reported previously.7 N m

a
e
M
h
t

f
m
o

o
s
r c
f
i
e s
y
l
b
h
a
P
l
i
a e
v h
A
t

r
o
f
5
Technical evaluation of Fujifilm AMULET Innovality tomosynthesis system

1. Introduction

e
1.1 Testing procedures and performance standards for digital mammographyr
t
n
e
This report is one of a series evaluating commercially available mammography systems
C
on behalf of the NHS Breast Screening Programme (NHSBSP). The testing methods

and standards applied are those of the relevant NHSBSP protocols, which are g
)
published as NHSBSP Equipment Reports. Report 06041 describes the testinng of full M
field digital mammography systems used for 2D imaging and Report 14072i describes
t P
a
the testing of digital breast tomosynthesis.
C
n
i C
NHSBSP protocols1,2 are similar to European protocols,3,4,5 but thde European protocols
r N
also provide some additional or more detailed tests and standards, some of which are
o
included in this evaluation. (
-

o y
Additional tests were carried out according to the UK reCcommendahtions for testing
mammography X-ray equipment as described in IP EM Report 8p9.6
l
a a
1.2 Objectives n r
g
o
o
i
The aims of the evaluation were to: t m
a
N m
 measure the technical performance of the Fujifilm AMULET Innovality system in

a
tomosynthesis mode e
M
h
 verify that the doste and noise were as previously reported when the system is

f
operating in 2Dm mode (Reoport 1601).7
o
s
r c
f
i
e s
y
l
b
h
a
P
l
i
a e
v h
A
t

r
o
f
6
Technical evaluation of Fujifilm AMULET Innovality tomosynthesis system

2. Methods

e
2.1 System tested r
t
n
e
Details of the system tested are given in Table 1.
C

Table 1. System description
g
)
Manufacturer Fujifilm n M
Model AMULET Innovality i
t P
Target material Tungsten (W) a
Added filtration 50μm Rhodium (Rh) for 2D n C
700μm Aluminium (Al) for tomiosynthesisC
d
Detector type Amorphous selenium
r N
Detector serial number J125579 o
(
Image pixel size 50µm in 2D images, 1-00µm in ST and HR

reconstructed focal polanes, 150yµm and
C h
100µm in ST and HR projections
p
Detector size 240mm x 300m m
l
Source to detector distance 650mm a a
Source to table distance 633mm n r
g
Automatic exposure control (AEC) AEC, ioAEC
o
modes avaiilable in 2D and tomosynthesis modes
t m
a
AEC dose levels High (H), Normal (N), Low (L)
Tomosynthesis projections NFifteen promjections without anti-scatter grid
equally spaced covering range ±7.5˚ (ST)
a
e
and ±20˚(HR)
M
h
Reconstructed focal planes Focal planes at 1mm intervals, number
t
equals compressed breast thickness in mm
f
m
o plus 5
Software version o FDR-3000AWS Mainsoft V7.0
s
r c
f
i
e s
The system has 2 tomyosynthesis modes:
l
b
h
 Staandard (ST)P mode which uses a narrow angular range of projections (15°)
l
i
a e
 High Resolution (HR) mode which uses a wide angular range of projections (40°)
v h
A
t
There is a facility available to carry out a combination exposure, in which a 2D and a
r
tomoosynthesis exposure are performed within a single compression.
f
Fujifilm set up the system for testing in service mode, which has reconstructed
tomosynthesis quality control (QC) images available as sets of 2D images
corresponding to the individual focal planes, in Digital Imaging and Communications in
Medicine (DICOM) CT format. In normal clinical use the reconstructed images would be
7
Technical evaluation of Fujifilm AMULET Innovality tomosynthesis system

available in the standard BTO DICOM format, and this would be a more convenient
format for routine QC testing, but should make no difference to the results of these
tests. In the CT format images tested the pixel spacing in reconstructed focal planes

e
differs from the image pixel sizes given in Table 1, which are the nominal pixel spacing
r
at the detector. The pixel spacing reduces with increasing height above the detector. t
n

e
There is a logarithmic relationship between pixel value and detected radiation dose in
C
Fujifilm 2D images and projections. For 2D QC analysis it is necessary to linearize pixel

g
values with respect to dose. To standardise the linearization process the “S” and “L” )
n M
values used to set the pixel values in the image were set to 121 and 4.0 respectively. In
i
reconstructed tomosynthesis images pixel values have a complex relationtship to dosPe.
a
They are by definition heavily processed and can therefore not be linenarized in a C
manner analogous to the linearization of 2D images. Clinical reconistructed C
d
tomosynthesis images from this system are created using the logarithmic projections.
r N
o
For this evaluation Fujifilm also made available reconstructions created using linearized
(
-
projections, and analysis was carried out using both typeso of reconsytr uction to compare
their merits for QC purposes. C h
p
l
The reconstructed tomosynthesis images availabale in QC moade as used for this
n r
evaluation excluded some of the image processing appliged to clinical images. There are
o
2 types of post-reconstruction processing available to oclinical images: Pattern 1 and
i
Pattern 2. Pattern 2 is less commonly usted and themrefore the QC reconstructions for
a
this evaluation used Pattern 1. N m

a
e
The system generated a synthetic 2D view (‘S-view’) for each ST and HR
M
h
tomosynthesis reconstruction, but these were not evaluated
t

f
m
o
The AMULET Innovality is shown in Figure 1 – image courtesy of Fujifilm.

o
s
r c
f
i
e s
y
l
b
h
a
P
l
i
a e
v h
A
t

r
o
f
8
Technical evaluation of Fujifilm AMULET Innovality tomosynthesis system

e
r
t
n
e
C

g
)
n M
i
t P
a
C
n
i C
d
r N
o
(
-

o y
C h
p

l
a a
Figure 1. The Fujifilm AMULET Innovality dnigital breasrt tomosynthesis system
g
o
o
i
t m
a
2.2 Dose and contrast-to-noise ratio under AEC
N m

a
2.2.1 Dose measurement e
M
h
t

Measurements were mad e of half value layer (HVL) and tube output across the clinically
f
m
o
relevant range of kV and filter combinations. Output measurements were made on the

o
midline at the standard positiosn, 40mm from the chest wall edge of the breast support
r c
platform. The cofmpression paddle was in the beam, raised well above the ion chamber.
i
As the systeme uses diffesrent target filter combinations for 2D and tomosynthesis, output
y
l
measurembents were made in both modes. In tomosynthesis mode the stationary
h
exposuare option was used.
P
l
i
a e
In both 2D and tomosynthesis modes, exposures of a range of thicknesses of
v h
polymethyl methacrylate (PMMA) were made under AEC. For each measurement the
A
t
height of the paddle was set to match the indicated thickness to the equivalent breast
r
thicokness for that thickness of PMMA. In 2D mode exposures were made both with and
f
without the intelligent AEC setting (iAEC) which adjusts exposures according to
localised densities in the breast.

MGDs for the standard breast model for 2D and tomosynthesis exposures were
calculated using the methods described in the UK and European protocols.1-5 The
9
Technical evaluation of Fujifilm AMULET Innovality tomosynthesis system

method of measuring MGD in tomosynthesis mode described in the UK protocol differs
slightly from the method described by Dance et al 8 in that the incident air kerma is
measured with the compression paddle well above, instead of in contact with, the ion

chamber. Measurements on other systems 9,10 show that this difference reduces the air e
r
kerma and thus the MGD measurement by 3% to 5%. t
n

e
C
2.2.2 Contrast-to-noise ratio

g
)
n M
For CNR measurements a 10mm x 10mm square of 0.2mm thick aluminium foil was
i
included in the phantom described above, positioned 10mm above the tatble on the P
a
midline, 40mm from the chest wall edge. (The standard position is 60nmm from theC chest
wall edge.) i C
d

r N
o
CNR in 2D mode was assessed using 5mm x 5mm ROIs positioned in the centre of the
(
-
aluminium square and 2 background positions, to the cheost wall andy n ipple sides of the
square, as shown in Figure 2. The CNR in tomosyntheCsis mode whas measured in the
focal plane in which the aluminium square was brou ght into focpus. Because the
l
aluminium square was positioned closer than usuaal to the chaest wall edge and there
n r
was a gradient in pixel value perpendicular to the chest wgall, alternative ROI positions
o
were selected. The ROIs were subdivided into 1mm xo 1mm elements and the
i
background ROIs were positioned at thet same distamnce from the chest wall as the
a
aluminium square, as shown in FiguNre 3. The CmNR in tomosynthesis mode was
calculated using the average of the mean and standard deviation in pixel values for

a
e
each 1mm x 1mm element.
M
h
t
CNR was also assessed in the unpro cessed tomosynthesis projections acquired for the
f
m
o
above images, using a 5mm x 5mm ROI.

o
s
Variation of CNRr with dose cwas assessed in the reconstructed focal planes for a
f
simulated brea st thicknesis of 53mm (using a 45mm thickness of PMMA). The variation
e s
in central projection CyNR with breast thickness and the variation in projection CNR with
l
b
h
projection angle for a 53mm thick breast were also assessed.
a
P
l
i
a e
v h
A
t

r
o
f
10