SmartPET: A Small Animal PET Demonstrator using HyperPure Germanium Planar Detectors

1
SmartPET A Small Animal PET Demonstrator using
HyperPure Germanium Planar Detectors R.J.
Cooper(1), A.J. Boston(1), H.C Boston(1), J.R.
Cresswell(1), A.N, Grint(1), A.R. Mather(1), P.J.
Nolan(1),D.P. Scraggs(1), G. Turk(1), C.J.
Hall(2), I. Lazarus(2), A.Berry(3) T.
Beveridge(3), J. Gillam(3), R.A. Lewis(3) (1)
Department of Physics, University of Liverpool,
UK (2) CCLRC Daresbury, Warrington, Cheshire,
UK (3) School of Physics and Materials
Engineering, Monash University, Melbourne,
Australia
Abstract The SmartPET project aims to exploit
advances in the sensitivity, timing, position and
energy resolution of HPGe detectors to construct
a small animal Positron Emission Tomography (PET)
system. The development of sophisticated
digital acquisition techniques 4 and the use of
Pulse Shape Analysis (PSA) 1 and Gamma Ray
Tracking (GRT) 1,2 will allow accurate position
and energy information to be extracted, allowing
scattered interactions to be identified and used
for image reconstruction.
Motivation In conventional PET systems 85 of
photons incident on the detectors will undergo
Compton scattering and will be rejected on the
basis of energy. The rejection of these events
leads to highly inefficient data collection.
This project aims to tackle the deficiencies in
current PET systems by utilising the excellent
energy resolution and position sensitivity
offered by highly segmented germanium detectors
to use a greater proportion of events (70).
This will provide increased patient throughput
and/or reduced patient dose while achieving
improved spatial resolution. The SmartPET system
will provide highly efficient dual modality
PET/SPECT imaging with fine spatial resolution
and the potential for PET/MRI fusion.
The SmartPET Detectors The SmartPET system is
based on two 60mmx60mmx20mm HyperPure germanium
(HPGe) crystals electrically segmented with
orthogonal 5mm strips providing a raw position
resolution of 5mmx5mmx20mm. The use of Pulse
Shape Analysis (PSA) 1 will improve this
allowing interaction positions to be defined to
an accuracy of 1mm3 1. This precise position
and energy information will then form the input
of Gamma Ray Tracking (GRT) 2 algorithms to
accurately track the path of gamma rays through
the detector thus facilitating the inclusion of
scattered events in the reconstruction data set.
Pulse Shape Analysis (PSA) Gamma Ray Tracking
(GRT) Analysing digitised pre-amp charge pulses
resulting from gamma ray interactions allows us
to calibrate the 3-D position sensitivity of the
entire crystal volume 3. Depth of interaction
information is obtained from rise-time analysis
while the x-y interaction position is ascertained
using image charge analysis 2. The kinematics
of Compton Scattering are then used to
reconstruct interaction sites on an event by
event basis.
Gamma Ray Tracking (GRT) The well defined
kinematics of Compton scattering allow gamma ray
paths to be reconstructed back to the source
according to the equations below. Equation (1)
derives the energy of the incoming photon from
the angle between scatters and the first two
energy deposits. Equation (2) allows the
azimuthal angle of the photon to be calculated
from the incoming photon energy and the first
energy deposit.
Incoming Gamma Ray, Energy Ei
(1)
Rise-time analysis Following a gamma ray
interaction electron-hole pairs are produced
which drift to the electrodes under the influence
of the applied electric field. As the applied
bias ensures the drift velocity of the carriers
is saturated, the charge collection time is a
function of interaction depth.
(2)
Image Reconstruction Analytic and statistical
algorithms are being developed to allow fast,
accurate 3-D image reconstruction. Two
dimensional reconstruction algorithms have been
tested using experimental data acquired from a
scintillation detector based coincidence system.
Image Charge Analysis The movement of charge
carriers induces signals (Qright and Qleft) on
adjacent strips. The relative magnitudes of
these image charges varies as a function of x-y
interaction position. We can therefore calibrate
this lateral position sensitivity according to
the image charge asymmetry parameter, (A). The
image on the right shows variation in image
charge magnitude.
References 1 Performance of the GRETA prototype
detector, K. Vetter et al, NIM A Vol. 452
(2000) 2 Three-dimensional position sensitivity
in two-dimensionally segmented HP-Ge detectors,
K. Vetter et al, NIM A Vol. 452 (2000) 3 The
position response of a large-volume segmented
germanium detector, M. Descovich et al,
NIM A 4 The GRT4 VME Pulse Processing Card for
Segmented Germanium Detectors, I. Lazarus et al
Future Work Two SmartPET crystals are being
characterised at the University of Liverpool to
provide a detailed detector response matrix
encompassing the entire active volume. Following
the completion of this work, electric field
simulation and the implementation of PSA and GRT
algorithms, the detectors will be mounted in a
rotating gantry to facilitate both PET and SPECT
imaging. Colleagues at Monash University are
investigating the use of cone beam reconstruction
for LoR definition.