8/2/2017. Disclosure. Philips Healthcare (Cleveland, OH) provided the precommercial

Transcription

1 8//0 AAPM0 Scientific Symposium: Emerging and New Generation PET: Instrumentation, Technology, Characteristics and Clinical Practice Aug Wednesday 0:4am :pm Solid State Digital Photon Counting PET/CT Instrumentation and Technology Jun Zhang, PhD, DABR Assistant Professor Wright Center of Innovation in Biomedical Imaging Department of Radiology The Ohio State University Wexner Medical Center Disclosure Philips Healthcare (Cleveland, OH) provided the precommercial investigational PET/CT system (Vereos). DPC PET/CT (Vereos, Philips Healthcare) at OSU Pre-commercial release unit of digital photon counting PET/CT system (Vereos, Philips Healthcare) operational at The Ohio State University (PI: Michael V Knopp, MD, PhD) since October 04. PET detector technology moving from PMT to solid state DPC is a technology leap.

2 8//0 From PMT to Solid State DPC Light Spreading PMTs Positioning Logic Crystal High Energy Photon Scintillation photons Analog PMT Scintillator Crystals Analog Many-to-many coupling High Energy Photon Digital Photon Counter Scintillation photon DPC tile Digital One-to-One coupling Vereos PET Detector Geometry 8,040 x,00 } 0000 DPC Data Acquisition Sequence Biggest contribution to the overall sensor DCR is caused by only a small percentage of the cells such as ~0 % cells responsible for 0-80% DCR. Disabling high DCR cells on small active sensor area leads to significantly reduced overall sensor DCR. Digital Photon Counting Digital Time of Flight : Coupling. Philips Vereos PET/CT Whitepaper. Haemisch Y, et al. Physics Procedia. 0;:4 0

3 Energy Resolution Energy Resolution (% FWHM) [% FWHM] Timing Resolution [ps FWHM] Timing Resolution [ps FWHM] JNM 48(00) 4 8//0 Energy Resolution A FWHM.% Energy [kev] 0.0% 8.0%.0% 4.0%.0% 0.0% 8.0%.0% 4.0%.0% 0.0% B Energy resolution (%) TOF Timing Resolution Axial FOV 4ps FWHM 8 ps FFWHM 8 F Na point source Timing [ps] [ps] Experimental setup (ref ) Vereos PET/CT Gemini TF (ref ) Avg Singles Rate [Mcps] Timing Resolution (ps) NEMA NU 0 Spatial Resolution

4 Sensitivity cps/mbq Sensitivity cps/mbq Effective Sensitivity [kcps/mbq] Effective Sensitivity [kcps/mbq] //0 NEMA NU 0 Spatial Resolution FWHM FWTM at cm at 0cm at 0cm at cm at 0cm at 0cm Axial Transverse (Radial) Transverse (Tangential) R(FWHM) K r R i + R p + R a + R l K r R i + R p + R a where R is the overall spatial resolution, K r a scale factor, R i the intrinsic resolution determined by crystal width, R p the error due to positron range, R a the error from noncollinearity, and R l is the decoding error (anger logic positioning localization). For DPC PET, R l is zero due to that : coupling eliminates need for position decoding.. WW Moses. Nucl Instrum Methods Phys Res A. 0; 48 (Supplement ): S S40. Jun Zhang, Michael V Knopp. J Nucl Med 0; 8 (supplement ): Compared to conventional photo-multiplier PET systems, spatial resolution of the SiPM digital photon counting PET system is NOT affected by A. Crystal size B. Positron Range C. Noncollinearity / Acollinearity D. Localization decoding SAM Question Answer: (D) : coupling eliminates need for position decoding 0 NEMA NU 0 Sensitivity A 0cm 0cm B 00 NEMA PET Sensitivity Phantom (Data Spectrum Corporation, Hillsborough, NC) Axial Sensitivity A Profile 0 B cm 0cm Slice Number Vereos PET Analog PET Vereos PET Analog PET Courtesy of Philips Activity at Measurement Time [mci] G TOF D time-of-flight gain cδt Slice Number Average at 0 and 0cm: ~. kcps/mbq Activity at Measurement Time [mci] S eff (A) S NEMA (A) G TOF Budinger. JNM 4(): -8 4

5 NECR NECR(r) (kcps) [kcps] TOF Effective NECR (kcps) TOF Effective NECR(r) [kcps] Rate 0 0 (kcps) Rate (kcps) Contrast (%) Background Variability (%) 8//0 NEMA NU 0 NECR and Scatter Total Randoms TRUE Scatter A 00 OSU System x V 000 V, cm V, cm Activity Concentration [kbq/ml] Activity Concentration (kbq/ml) B NECR Scatter Fraction 0% Activity Concentration (kbq/ml) Peak NECR kbq/ml TOF peak NECR 4 kbq/ml OSU System x V Activity Concentration [kbq/ml] Activity (kbq/ml) NECR T / (T + S + R) Measured TOF Gain - OSU ps G 4. V ps G 4. Analog ~ 00 ps. For a 48cm diameter object, an estimated sensitivity gain and the SNR improvement of the digital photon counting TOF PET (0ps, 4.8cm uncertainty) compared to nontof PET are about A..0 and.0 B. 0.0 and 0.0 C. 0.0 and. D. 0.0 and.0 SAM Question Answer: (C) G TOF D D time-of-flight gain cδt Δx D G SNR Δx NEMA NU 0 Image Quality Sphere (mm) Sphere (mm)

6 SUVmean % //0 Phantoms IQ Sphere Contrast (%) Background Variability (%) 0 SPHERE SIZE (MM) LBR Alpha-Bravo Uniformity SUV Axial Profile - Brain FOV.0mCi 8:4am on 09/9/04, phantom on the table, acq started at 0:49am, SUVmax0.98, SUVmin0.89, SUVmean0.9, % variance8.% LBR Vereos Gemini Slice # System Specifications: DPC vs PMT Specifications Vereos Gemini TF 4 Detector Design Direct Photon Counting PMT Anger Logic Ring Diameter (mm) 4 4 Crystal LYSO LYSO Crystal Size (mm ) 4 x 4 x 9 4 x 4 x # of Detector Elements, PMT Crystal-Detector Coupling : multi : multi Transaxial FOV (mm) / / / / Axial FOV (mm) 4 80 Overlap in Axial FOV 9% % Energy resolution (%). Energy calibration : multi : multi Timing Resolution (ps) 0 0 Timing Calibration : multi : multi Coincidence window (ns).0 / 4.0 / 4..0 / 4.0 / 4. Reconstruction Algorithm D OSEM TOF D OSEM TOF, D Ramla Reconstruction Matrix (pixel) up to x up to 8 x 8 Reconstruction Voxel (mm ) 4 / 8 / 4 / 8 # Voxels per slice up to 4,9 up to 0,84 Resolution Recovery PSF No PSF PMT PET vs DPC PET Gemini Vereos

7 8//0 PMT PET vs DPC PET Non-TOF TOF Non-TOF TOF Non-TOF ps TOF 0ps 4mm PET ps 4mm PET PET Image Resolution (Recon) Spacing (mm) 4 x 4 x 4 Volume (mm ) 4 FOV mm 44 x 44 matrix FOV mm 9 x 9 matrix FOV mm N/A # Voxels of V N (e.g., 0,) PSF & Filter on/off Recon D TOF OSEM x x 8 88 x 88 matrix 8 x 8 matrix 8 x 8 matrix 8N (e.g.,,888) on/off D TOF OSEM x x x matrix x matrix x matrix 4N (e.g.,,,04) on/off D TOF OSEM Small Lesion Detectability with Less PVE 4x4x4 mm xx mm xx mm

8 8//0 Lesion Detectability with Less PVE Gemini (4mm ) Vereos (mm ) Brain PET A B C A B PET/CT Gemini TF Vendor 4 Gemini A TF 4 resolution mm mm XY Slice Thickness mm mm FDG Injected 4.mCi.0mCi Uptake time 4min 4min Acquisition Time 0min 0min XY resolution mm mm mm mm Slice Thickness mm mm mm mm FDG Injected 4.mCi.0mCi 0.0mCi.0mCi Uptake time 4min 4min 0min 49min Acquisition Time 0min 0min 0min 0min A B C D E F PET/CT Gemini TF 4 Gemini TF 4 mct mct Vereos Vereos XY resolution mm mm mm mm mm mm Slice Thickness mm mm mm mm mm mm FDG Injected 4.mCi.0mCi 0.0mCi.0mCi.9mCi.9mCi Uptake time 4min 4min 0min 49min 8min 8min Acquisition Time 0min 0min 0min 0min 90sec 90sec D E F A B C D PET/CT Gemini TF 4 Gemini TF 4 mctvendor B mct Brain PET Redefinition Speed, Convenience, Dose, IQ mm Brain of FOV 4mm Brain of WB mm Brain of WB mm Brain of WB Wholebody FOV 0min FDG 90s/bed FDG 8

9 8//0 WB PET Redefinition 0s/bed 0s/bed s/bed 90s/bed WB NaF PET 0s/bed 0s/bed 0s/bed s/bed s/bed s/bed Clinical Ultrafast PET 90s/bed min 0s/bed min 9s/bed.min 9

10 8//0 Low Dose FDG PET NIH R0 PIs: Knopp, Zhang Clinical Post 90 Y Microsphere Radioembolization Bremsstrahlung SPECT/CT DPC PET A B A 0min/bed; B - min/bed 0

11 8//0 9% (9bed) % (8bed) 0% (bed) 0% (bed).mci; NaF; 90s/bed; 0cm 9% % 0% 0% Highlights : coupling eliminates need for position decoding Timing and energy calibrations per crystal Improved timing resolution Reduced dead time Increased count rate performance Early digitization eliminates as much as possible analog noise and distortion.

12 8//0 How much SiPM DPC Detector Coverage Do We Really Need? A B C D E F A Full ring B Clinical Evaluation 0% detector C detectors reduction 0% detector reduction % detector % detector D E F reduction reduction 40% tile-ring reduction 8 Acknowledgement This research was supported by the Ohio Third Frontier OSDA TECH (PI: Michael V Knopp, MD PhD) and NCI R0CA9 (PI: Michael V Knopp, MD PhD; Jun Zhang, PhD) Philips Healthcare provided the pre-commercial investigational PET/CT system (Vereos) hardware and technical support Wright Center of Innovation in Biomedical Imaging (PI: Michael V Knopp, MD PhD)

13 8//0 Thank You