Innovative detection systems for digital mammography Edoardo Castelli University PowerPoint PPT Presentation

presentation player overlay
1 / 33
About This Presentation
Transcript and Presenter's Notes

Title: Innovative detection systems for digital mammography Edoardo Castelli University


1
Innovative detection systems for digital
mammographyEdoardo CastelliUniversity INFN -
Trieste (Italy)
  • Workshop on
  • Particle Accelerators and Detectors from Physics
    to Medicine

2
Motivations
  • Modern radiology aims at the improvement of
    diagnostic accuracy while delivering low dose to
    the patient
  • Mammography is a typical example for extreme
    requirements on the X-ray source as well as on
    the detection system (detection of
    microcalcifications and low contrast masses,
    possible indicators of early breast cancer)

3
Motivations
- Breast is one of the most radiosensitive organs
  • - Mammography is far from being perfect
    sensitivity ranges from 75 to 90
  • specificity from 90 to 95.

4
The problem of the diagnostic mammography an
example
  • In one year, at the University Radiological
    Department in Trieste, 5000 patients have a
    conventional imaging mammographic
    examination300 are positives, i.e. have
    suspected abnormalities at breast imaging200
    have a following histopathologic analysis75
    have malignant breast lesions
  • The ratio false positives/true positives is
    therefore 225/75 3assuming a 90 sensitivity,
    the prevalence results 1.6 and the specificity
    95.
  • The mammography (and/or ultrasonogrphy) is a
    first level exam. At this level there is a great
    demand for a substantial improvement of the
    sensitivity.At a second level, i.e. after an
    initial mammography (and/or ultrasonogrphy),
    being relevant the number of false positives,
    the main aim is a substantial reduction of this
    number, aiming at a consequent reduction of
    histopathologic analyses.

5
Possible solutions
  • No X-ray mammography
  • Diagnostic mammography with magnetic resonance
    imaging
  • Breast scintigraphy and SPECT
  • X-ray mammography
  • Dual-energy contrast-enhanced digital subtraction
    mammography (possible contrast agent).
  • Tomographic techniques linear tomography and
    digital tomosynthesis (conventional geometry).
  • Breast computer tomography (rotation of both
    X-ray tube and detector around the pendulant
    breast).
  • Development of dedicated digital detectorsINFN
    experiments MEDIPIX (Pisa et al.) and MATISSE
    (Trieste)
  • Synchrotron radiation mammography (Phase contrast
    imaging)SYRMEP Project - Synchrotron Radiation
    for Medical Physics (Trieste)

6
Standard mammographic equipment
GE DMR
7
MEDIPIX Collaboration
  • MEDIPIX1-IMI
  • Dipartimento di Fisica dellUniversità e Sezione
    INFN - Pisa
  • S.R.Amendolia, M.G.Bisogni, P.Delogu,
    M.E.Fantacci, M.Novelli, M.Quattrocchi, V.Rosso,
    A.Stefanini
  • AMS, CAEN, Gilardoni, Laben
  • MEDIPIX2
  • CERN, INFN (Pisa, Napoli, Cagliari), European
    Research Institutions
  • http//medipix.web.cern.ch/medipix

8
(No Transcript)
9
Detectors
  • Geometry
  • Pixel 150 mm, pitch 20 mm
  • 64 x 64 pixel for 1.2 cm2

Detector efficiency at 20 keV for different
thickness Si 300 mm 26 Si 525 mm 40 Si
800 mm 55 GaAs 200 mm 98
10
Si detectors comparison
Exposition parameters anode W, 40 kV, 40 mA,
630 ms
C4,75
100 micron Al
75 micron Al
C3,58
50 micron Al
C2,40
25 micron Al
C1,21
800 mm
11
Acquisition of a large area image
Our detection area is 1.2 cm2. In order to obtain
a complete image, the RMI phantom was exposed to
X-rays and moved using motors controlled by
PC. We acquired 64 different images. Each image
has been equalized by weighting with a high
statistical image obtained in the same beam
conditions. A Pb collimator, with a square
opening 1.5 cm in side, has been used so to
irradiate only an area comparable to the area of
our detector because this improves the image
quality.
The RMI 156 phantom
Tube settings 25 kV and 80 mAs
In the image we can see five nylon fibers, three
microcalcification groups and four tumor-like
masses. The experimental contrast of tumor-like
masses varies from 6.7 (detail 12) to 1.7
(detail 15). For the last detail the SNR is only
1.2.
8 cm
8 cm
12
The GaAs detectors
  • Goals obtained
  • 200 ?m thick ( 98 detection efficiency _at_22keV)
  • Breakdown voltage gt 500 V
  • At the 350 V operating voltage leakage current lt
    50nA/mm2, charge collection efficiency gt 75

SNR as a function of the mAs for the detail 12 (2
mm tumor mass) of the RMI 156 phantom, obtained
at 40 kV with GaAs and Si 525 mm thick detectors.
GaAs detectors will be used in a prototype of a
mammographic head
13
Acquisition Geometry
scan direction
focus
105 mm
3 mm
Pb collimator
440 mm
phantom
42 mm
  • Move and tile technique
  • 56 different acquisitions
  • scanning performed with stepper motors controlled
    by the PC

50 mm
detector
Detector-focus distance 640 mm
14
Full field image
4 of 5 tumor-like masses (2.00 mm, 1.00 mm, 0.75
mm, 0.50 mm thickness) 3 of 5 groups of simulated
micro-calcifications (0.54 mm, 0.40 mm, 0.32 mm
diameter) 4 of 6 different size nylon fibers
(1.56 mm, 1.12 mm, 0.89 mm, 0.75 mm thickness)
Tube settings -25 kVp -80 mAs -Texp 0.8 s MGD
0.68 mGy
Image processing Flat field equalization Normaliza
tion Noisy and dead pixels recovery
15
SYRMEP Project, MATISSE
Dipartimento di Fisica dellUniversità e Sezione
INFN - Trieste F. Arfelli, A. Bergamaschi, E.
Castelli, D. Dreossi, R. Longo, A. Olivo, S.
Pani, L. Rigon, T. Rokvic, C. Venanzi Dipartiment
o Universitario Clinico di Scienze
Cliniche, Morfologiche e Tecnologiche -
Università di Trieste F. Brizzi, L. Dalla Palma,
M. Tonutti, F. Zanconati Sincrotrone Trieste
SCpA A. Abrami, F. Billè, V. Chenda, L. Mancini,
R.H. Menk, G. Tromba, F. Zanini
16
The SYRMEP Beamline
SRlaminar beam
IonizationChamber
Exit SlitSystem
SampleMovement
Elettra Bending Magnet
EntranceSlitSystem
Si 111 Double Crystal Monochromator
Sample
  • Bending Magnet Beamline Source size 1.1 x 0.1
    mm2
  • Source-to-Sample distance 23 m
  • Monochromatic beam with tuneable energy (8.5 -
    35 keV)
  • Laminar beam cross section 4 x 150 mm2

17
The holder for the patient
Digital stereotactic biopsy table
Giottoproduced by IMS (Bologna, Italy)
18
Patient and detector holders
19
The movement stage of the holder
Rotation movement Tomographic mammography
Patient Table
Vertical (scanning)movement Planar mammography
Compressor
Floor
20
New beamline layout
Experimental room
Patient Room
Radiologist room
21
Previous results phase contrast imaging of
breast tissue
Phase Contrast
Thickness 3 cm Energy 17 keV MGD 0.5 mGy
22
The silicon micro-strip detector edge-on
geometry
  • Advantages of edge-on geometry
  • High absorption efficiency
  • Natural pixel linear array pixel aperture is
    given by the strip pitch (horizontal) and by the
    detector thickness (vertical)
  • Problems
  • Dead (undepleted) volume in front of the
    sensitive region that reduces the detection
    efficiency, in particular at low energies

F.Arfelli et al., An 'edge-on' silicon strip
detector for X-ray imaging, IEEE Trans. Nuclear
Science,Vol.44, n.3, p. 874 (1997)
23
The silicon micro-strip detector
micro strips
  • The guard ring is present on only 3 sides
  • The distance between the bias ring and the scribe
    line is kept 200 mm shorter than the standard
  • 256 to 1024 strips
  • Strip depth 2 cm
  • 100 mm strip pitch
  • Detector thickness 300 mm
  • Dead entrance window 200 - 400 mm
  • Detection efficiency 80 (20 keV)

guard ring
bias ring
Pixel aperture 100 x 300 mm2
24
MATISSEMAmmographic and Tomographic Imaging
with Silicon detectors and Synchrotron radiation
at Elettra
  • Monochromatic synchrotron radiation beam
  • Energy optimization
  • No beam hardening
  • Negligible divergence
  • Edge on silicon microstrip detector
  • High efficiency
  • High spatial resolution
  • Single photon counting
  • Scattering reduction
  • Several edge-on silicon microstrip detectors
    equipped with single photon counting read-out
    electronics have been previously tested for
    synchrotron X-ray imaging by the SYRMEP
    collaboration, with serious limitations for
    possible use in clinical, i.e. with patients,
    mammography
  • - excellent S/N but low counting rate
    capabilities
  • good speed characteristics but problems with S/N
    at low energy and with several ASICs operating in
    parallel

25
Previous results contrast-detail phantom
  • Single photon counting capability gt High
    contrast resolution

26
Previous results in vitro experimentation
  • Very promising results
  • Optimal for thick and dense breasts
  • Allows to recognize the shape of the lesions

27
Previous results in vitro breast tissue imaging
a) SR digital image Energy 17 keV Scan step 100
mm MGD 1 mGy
Conventional X-ray tube 26 kVp MGD 1 mGy
b) SR digital image Energy 20 keV Scan step 100
mm MGD 0.33 mGy
3 cm thick in vitro human breast tissue
28
The MATISSE detector
  • 20 cm wide
  • 2 layers
  • 100 ?m resolution
  • 80 efficiency
  • High speed single photon counting read out

29
MATISSE prototype
30
Bench tests Rate response
  • Pulses injected through the calibration capacitor
  • Optimal linearity up to 3.5 MHz pulse frequency

31
Beam tests First images
  • Contrast/Detail phantom
  • Resolution 100 x 100 ?m2
  • Contrast of least detectable detail is lower than
    1, with about 10000 photons/pixel

32
Beam tests Contrast measurement
28 keV
32 keV
33
Conclusions
  • MEDIPIX (Pisa et al.) An improvement can be
    expected both in sensitivity and specificity with
    low dose examinations, when the detector will be
    used in a conventional X-ray equipment for first
    level mammography.
  • MATISSE (Trieste)The synchrotron radiation
    mammography, in this phase, is a second level
    examination, with the aim of improving the
    specificity, therefore reducing, in perspective,
    the biopsy number.As a validation study, in vivo
    mammographic analyses are foreseen on a limited
    number (70 - 100) of volunteer patients with
    uncertain diagnosis the first patients are
    scheduled for the end of 2004.Phase contrast
    radiography with conventional films will be the
    first step. Digital planar and tomo mammography
    with silicon microstrip detector will be the next
    steps.
Write a Comment
User Comments (0)
About PowerShow.com