250 likes | 473 Views
PET in the detection of breast cancer. The ClearPEM Project. Joao Varela IST/LIP, Lisbon, Portugal. Encontro Nacional de Ciência — Ciência 2009 Fundação Calouste Gulbenkian 29-30 de Julho de 2009. Breast Cancer. Breast Cancer Most common type of cancer among women
E N D
PET in the detection of breast cancer The ClearPEM Project Joao Varela IST/LIP, Lisbon, Portugal Encontro Nacional de Ciência — Ciência 2009 FundaçãoCalousteGulbenkian 29-30 de Julho de 2009
Breast Cancer • Breast Cancer • Most common type of cancer among women • Second deadliest cancer • One out of 9 women develop a form of breast cancer throughout her life • Mammography • Advantages • Low cost • Good sensivity/specificity • Disadvantages • Based in structural tissue changes • Less reliable in dense breasts • High false-positive rates
P.E.T. (Positron Emission Tomography) • PET • Basedonthedecayof a positronemittingradionuclide (tracer) • 18F-FDG mostcommonlyusedradiotracer • Basedonhistologicalandmetabolicalchangesofthetissue • PET vs. Mammography • Advantages • Notdependentontissuedensity • Verygoodsensivity • Disadvantages • More expensive • Lowsensivity for smalllesions (WholeBody-PET)
PET Mammography • PET and breast cancer • PET demonstrated good sensitivity to breast cancer • Radiotracer • FDG (fluoro-deoxyglucose): glucose labeled with the positron emitter F-18 • FDG has strong affinity to cancer cells • Other tracers under investigation • Dedicated PEM scanners • One commercial scanner (Naviscan, USA) • Two research prototypes in USA X-ray mammogram Dedicated Breast-PET Imaging Randolph Cancer Center West Virginia University
The ClearPEM Project • Goal • High performance scanner, toapproachthelimitsallowedbytracerphysiology • Framework • ProjectdevelopedintheframeworkoftheCrystalClearCollaboration, CERN • FundedbythePortugueseInnovationAgency (AdI) • 6 yearsdevelopment • 4.5 M€ investiment • Status • IP licensend to PETsys, SA • Scanner inclinicaltrials
The ClearPEM Project • Requirements • High counting sensitivity • Detector capable of image resolution of 1 mm • Detector capable to sustain a large flux of single photons (up to 10 MHz) • On-line coincidence trigger with few ns resolution • No data acquisition dead-time (up to 1 M coincidence events/s) • Measurement of individual hits of Compton events in the detector • Movable and compact dual-head detector plates with large active area • No parallax effect
Detector Technology • Crystals • Material: LYSO:Ce • Density: 7.4g.cm-3 • EmissionPeak: 420nm • LightYield: 27000 photons/MeV • TimeConstant: 40ns • Geometry: 2x2x20 mm3 • Avalanche Photodiodes • OperatingVoltage: 350-450V • DarkCurrent: ≤10nA • Gainuniformity (sub-array): ±15% • Detector Plates • 6144 crystals • 12288 readoutchannels • 160x180 mm2surfacearea • Front-backreadout for DoImeasurement
Frontend ASIC Characteristics Technology: CMOS 0.35μm Area: 70mm2 Input: 192 channels Output: 2 highest channels Max Input Charge: 90 fC Noise: ENC ~ 1300 e- Shaping: 40ns Analog Memories: 10 samples Clock Frequency: 50-100MHz Power: 3.6 mW/channel
Frontend (FE) Electronics • Frontend Board • Processes 384 APD channels • Contains 2 ASICs for signal selection • 2 High-speed ADCs (10bit, 100MHz) • 1 LVDS transmitter (600Mbps) • Supermodule • Comprises 2 FE Boards • Processes 768 APD channels • Detector Plate • Comprises 8 Supermodules • Processes 6144 APD channels • Contains one Service Board to control 192 high-voltage lines as well as power supply and clock distribution
FE Electronics Performance • Pulse Shape • Amplifier response rise time: 20ns • Variation in baseline <0.5% • Noise • ENC = 1300 e- r.m.s. • Inter-channel dispersion ~ 8% • (2.2 ADC Counts = 5keV) • (Noise measurements obtained with full electronics chain)
Data AcquisitionElectronics • DAE CrateSystem • DAE housedin a single 19” rackcrate • Uses twocPCIbackplanes • 1 TGR/DCC Board • 4 DAQ Boards • FE - DAE bandwidthup to 19.2Gbps • Sophisticatedcoincidencetrigger (36k calibrationconstants) • DAE-Acq Server bandwidthup to 6.4Gbps
Data Acquisition Electronics Performance • Trigger Performance • Events in coincidenceup to 2.5MHz • (This involves computation of energy and time and Compton grouping and transmission to the trigger processor) • Acquisition rateup to 0.8MHz • (This involves readout of the event dataframe after the issueing of a trigger) • Disk storage > 300MBps
Detector Performance • Performance results • Dispersionofchannelgain 15.3% • Energy resolution at 511 keV 15.9% • Dispersion of energy resolution 8.8% • Single photon time resolution 1.5 ns (RMS) • Coincidence time resolution 5.2 ns (FWHM) • Resolution in DoI 2 mm 137Cs Resolution ~12.5%
OSEM-3D OSEM-2D ClearPEM ImageResolution • Image Setup • 1mm Na-22 source • Gridwith 5mm pitch • Two acquisitions with orthogonal plate orientations for each source location (400-600 keV) • Simultaneous reconstruction of 16 source positions • Results • Horizontal FWHM: 1.3mm • Vertical FWHM: 1.2mm
ClearPEM ImageResolution • Parallax effect • Measurement of 3D photon interaction coordinates eliminates parallax effect • ClearPEM is unique in this respect (DoI resolution of 2mm) • Images without parallax correction show considerable blurring 5 mm 1 mm
ClearPEM ImageUniformity • Images of uniform Ge-68 source • Image artifacts due to detector effects are corrected • Absorption and scatter corrections are not applied (less intensity at the center) • Reconstruction with 4 orientations of the detector plates • Very good uniformity Cylinder filled with positron emitter Ge-68 2 mm slices
Simulations • SimulationModel • NURBS CArdiac Torso (NCAT) Phantom • Detector detaileddescription • Standard injectionof 10mCi (370MBq)
Clinical Trials • Scanner Installation • Hospital Garcia Orta at Almada not available • Obliged fall back solution at IPO, Porto • Phase 1 • Tuning the image reconstruction with real cases • Patients indicated for PET/CT (other disease) • Normally negative breast exams • Started in June 2009 • Phase 2 • Assessment of PEM sensitivity / specificity • Comparison to mammography and MRI • Patients with positive indication from x-rays mammography
ClinicalImages • Tuning image reconstruction • Phase 1 • Normalization correction • Correction of the effect of background radiation • Effect of scattered radiation • Measure detector sensitivity • Evaluate FDG uptake in the breast • Validate simulation results
ClearPEM andUltrasound • Multimodal PET – US • CERIMED and University Hospital Marseille • Ultra-sound probe with elastography capabilities coupled to ClearPEM • Cross-reference system and PET-US image fusion • Construction of second ClearPEM machine well advanced
Conclusions • ClearPEM technological developments were successfully completed • The detector performance is excellent • ClearPEM is one of the most innovative APD-based PET systems in clinical • Scanner is presently installed at IPO, Porto • Clinical Phase 1 is on-going