M4D1  Clinical Emission and Hybrid Tomography Instrumentation

Thursday, Nov. 5  16:30-18:00  Golden Pacific Ballroom

Session Chair:  Ling-Jian Meng, , ; Paul Marsden, King's College London, England, United Kingdom

Show/Hide All Abstracts

(16:30) M4D1-1, Final Design of the C-SPECT-I Lab-Prototype

W. Chang1, M. Rozler1, P. Sankar1, D. Stentz1, J. Strologas1, R. Arseneau1, S. Metzler2

1Dept. of Diagnostic Radiology, Rush University Medical Center, Chicago, IL, USA
2Department of Radiology, University of Pennsylvania, Philadelphia, PA, USA
Abstract? Improving system sensitivity without sacrificing imaging resolution is the key to improving the performance of cardiac SPECT imaging. This sensitivity increase is needed for reducing imaging time, or radiation dose, and/or motion artifacts for clinical imaging studies, as well as for exploration of new clinical applications. In addition, attenuation correction is necessary to yield quantitative information - the long-standing goal of SPECT imaging. These goals can be met by C-SPECT - our proposed dedicated cardiac platform for SPECT imaging. High sensitivity is accomplished by C-SPECT?s optimized detection and system geometry, which wraps around patients? left-front thorax and provides the highest practical geometric efficiency for any spatial resolution of collimation. The first generation of C-SPECT platform - C-SPECT-I ? is the simplified version that uses parallel axial collimation instead of the ultimate 3D converging collimation. In the presented final design, C-SPECT-I?s variable collimation system provides 5 slit-arcs, for transaxial collimation to match with 2 slat-stacks, for axial collimation, to provide high system sensitivity and a large number of simultaneous projections. This lab-prototype will provide a system sensitivity that is 2.5 times that of a dual-head SPECT system for heart imaging with the same hardware resolution at the center of the imaging volume as well as unprecedented functionality and operational versatility.

(16:45) M4D1-2, Implementation and First Results of the Fully Suspended Cone Beam CT and SPECT System for Dedicated Breast Imaging

J. Shah1,2, S. D. Mann2,3, R. L. McKinley4, M. P. Tornai1,2,3

1Biomedical Engineering, Duke University, Durham, NC, USA
2Radiology, Duke University Medical Center, Durham, NC, USA
3Medical Physics, Duke University Medical Center, Durham, NC, USA
4ZumaTek, Inc., Durham, NC, USA
Stand-alone cone beam CT and SPECT systems capable of complex sinusoidal acquisition trajectories have previously been developed for dedicated breast imaging and used in early clinical studies. The fully-3D motions of the SPECT system can view into the chest wall and throughout the breast volume. The polar tilting capability of the CT system has shown a marked improvement in sampling into the chest wall, with comparable <6mGy total dose delivered to the volume as with a circular orbit, while eliminating cone beam artifacts because of the fully-3D acquisitions. A hybrid SPECT-CT system, with each individual modality capable of independently traversing complex trajectories around a pendant breast, was recently designed and the practical implementation of this design is presented here. The CT system consists of a 40x30 cm^2 flat panel imager and an x-ray tube with a 16° anode angle, placed on opposing ends of the completely suspended gantry. A unique mechanism employing a linear stage was implemented to tilt the suspended gantry within 0.02° positioning error about the 3D center of rotation; the fully-3D SPECT system is nested inside the suspended CT gantry, perpendicular to the source-detector pair. Both sub-systems are positioned on an azimuthal rotation stage enabling spherical trajectories. Initial imaging results demonstrate the additional off-axis projection views of various phantoms, allowed by the ±15° polar tilting of the CT system. To date, this is the first implementation of a fully-3D positioning hybrid SPECT-CT system that we are aware of and could have various applications in diagnostic imaging.

(17:00) M4D1-3, Delay Grid Multiplexing: Time-Based Positioning Method for Silicon Photomultiplier

J. Y. Won, G. B. Ko, J. S. Lee

Dept. of Nuclear Medicine, Seoul National University, Seoul, Korea
In this paper, we propose a fully time-based two-dimensional multiplexing method using the principle of the global positioning system. Every SiPM channel is connected to the delay grid (PCB trace) and the identical signals with the differences of transit times are detected at the four corner nodes of the delay grid. The transit times from each SiPM channel to the four corner nodes are uniquely encoded, so that the position where the SiPM is firing is identified using the differences of transit times. This method uses the innate PCB trace differences between the SiPM channels for multiplexing, so that it does not demand analog components for multiplexing.
In order to verify the proof of concept, we used a 4x4 SiPM (S11064-050P, Hamamatsu Photonics K.K.), which is one-to-one coupled with an array of 4x4 LGSO crystals, of which size is 3x3x20mm3. We employed a fast digitizer (DT5742B, CAEN) based on the switched capacitor array to obtain the fine time resolution. We evaluated position decoding performance (distance-to-width ratio; DWR), energy resolution, and coincidence resolving time (CRT). The DWR was calculated as the ratio of distance between two adjacent spots in the flood map to the average FWHM of the two spots. The reference detector was an R9800 photomultiplier tube (Hamamatsu Photonics K.K.) coupled with a single 4x4x10 mm3 LYSO crystal. The single timing resolution was 217 ps FWHM.
The 16 crystals were clearly identified using only time information and the DWR was 3.94±1.06. The energy resolution was 12.1±0.8% FWHM and the CRT was 442±26 ps FWHM.
This time-based multiplexing with time-of-flight (TOF) capability would be useful in TOF PET/MR because it demands far fewer non-magnetic analog components than other multiplexing methods. Furthermore, there is no signal distortion induced by multiplexing, so that time-over threshold technique can be used and pulse pile-up correction is easier than other charge multiplexing methods.

(17:15) M4D1-4, A Light Sharing, Charge Multiplexed Time-of-Flight Depth-of-Interaction PET Detector

M. F. Bieniosek1,2, J. W. Cates2, C. S. Levin1,2,3,4

1Department of Electrical Engineering, Stanford University, Stanford, CA, USA
2Department of Radiology, Stanford University, Stanford, CA, USA
3Department of Bioengineering, Stanford University, Stanford, CA, USA
4Department of Physics, Stanford University, Stanford, CA, USA
Time-of-flight (TOF) and depth of interaction (DOI) measurements are both important capabilities to improve clinical PET imaging. The combination of these two measurements has been shown to improve image quality and uniformity. Current high performance TOF DOI detectors have a high level of complexity, requiring cooling, pulse shape information or increased numbers of photosensors. This works describes and characterizes an approach to TOF DOI using a two layer, 20mm thick, lutetium-yttrium oxyorthosilicate (LYSO) light sharing crystal array. The crystal array was coupled to a single ended SiPM readout with a novel binary encoded multiplexing scheme that has a 4:1 timing channel multiplexing ratio, and two position channels. Flood maps show show excellent crystal separation with a minimum ratio of distance between crystal peaks to standard deviation of crystal peaks of 7.8. The detector achieves 10mm DOI resolution with 205 +/-5ps FWHM coincident time resolution. Time resolution was very uniform between layers (205 +/- 5ps for the top layer and 204 +/- 5ps for the bottom layer). The simplicity of the readout scheme makes it a good candidate for scaling to a practical TOF DOI PET system. The detector presented in this work has single ended readout, no active cooling, a 4:1 timing channel multiplexing ratio, and requires no pulse shape information.

(17:30) M4D1-5, Effects of Operating Parameters on a Digital SiPM Module Directly Coupled with a Pixelated Scintillator Array for Positron Emission Tomography

S. I. Kwon, S. R. Cherry

Biomedical Engineering, University of California, Davis, Davis, CA, USA
A digital SiPM (PDPC) is a candidate photosensor for our project to develop a total-body PET scanner with a long axial field-of-view (~2 m). Because such a scanner will have very high sensitivity, and can therefore support high spatial resolution, it is desirable to use scintillator crystals that are smaller than the SiPM pixel, rather than standard 1:1 coupling and readout. A digital SiPM has a relatively wide selection of configurable parameters compared to conventional analog SiPMs, and some interesting behaviors were observed during experiments when using light sharing across multiple SiPM elements. The digital-SiPM/LYSO-array detector consisted of a digital SiPM module and an 8 × 8 LYSO array that comprised 3 × 3 × 20 mm3 polished LYSO crystals. Various coincidence measurements were performed with different digital SiPM parameter settings. All 64 crystals were distinguishable in the flood histograms. However, the pattern of the histograms were different for each parameter setting. Average coincidence resolving times rapidly increased with trigger thresholds from 220 ps to 1.2 ns. Measurement with "no" and "full" neighbor logic had the highest and the lowest coincidence event rate for each trigger threshold, respectively. The validation interval strongly affected the flood histogram and complete single event ratio (defined as the fraction of events in which all expected dies contained valid data). The complete single event ratio increased along with trigger threshold and validation interval. Moreover, it was observed that timestamps of die events were grouped with particular time durations and this time duration depended on the validation interval. Careful selection of single event time window along with the validation interval is required for optimal performance. These results are helpful in informing future PET development using digital SiPMs to decode scintillator arrays when 1:1 coupling is not used.

(17:45) M4D1-6, Time-of-Flight and Depth-of-interaction PET Detectors Using Phosphor-coated Crystals and FBK HD-RGB SiPMs

E. Roncali1, A. Ferri2, A. Gola2, F. Acerbi2, E. Berg1, S. I. Kwon1, C. Piemonte2, S. R. Cherry1

1University of California-Davis, Davis, CA, USA
2Fondazione Bruno Kessler, Trento, Italy
Recent research in the field of positron emission tomography (PET) focuses on scanners with thicker detectors, extended axial field of view, and time-of-flight (TOF) capability to improve sensitivity and signal-to-noise ratio. Thick detectors create a depth-dependency in the timing resolution that needs to be corrected to achieve the best possible timing performance for TOF PET scanners; in long AFOV scanners spatial resolution is degraded by parallax errors and will require depth-of-interaction (DOI) correction. We have proposed a method to encode DOI using phosphor-coated crystals and pulse shape discrimination techniques, and demonstrated how this approach could be combined to high timing resolution detectors. Our initial work with these TOF-DOI detectors was performed with fast PMTs. Although we showed that DOI could be encoded with minimal degradation of the timing resolution (~400 ps), a major limitation of PMTs for this specific method is their low quantum efficiency for yellow light, corresponding to the wavelengths of the scintillation light converted by the phosphor coating. This degrades both the DOI and timing resolution. Here, we assessed the potential of silicon photomultipliers (SiPM) developed by the Fondazione Bruno Kessler (FBK, Italy) to read out phosphor-coated crystals. Their high photodetection efficiency across the visible spectrum makes these SiPMs highly attractive for phosphor-coated TOF-DOI detectors. Phosphor-coated crystals with various coating schemes were coupled to the FBK SiPMs, and energy and timing resolution were measured. An outstanding timing resolution of 190 ps was obtained with single 3x3x20 mm3 LYSO crystals coated on a third of one of their lateral sides. The degradation due to the coating was only 25 ps. Energy resolution varied from 10.25% with uncoated crystals to 11.7% with the third-coated crystal. These results clearly demonstrated the potential of the HD-RGB SiPMs to read out phosphor-coated crystals.