Sergei Ivanovich Dolinsky

Multi-Pixel Photon Counters for TOF PET detector and its challenges

Multi-Pixel Photon Counter (MPPC) is a Geigermode APD developed by Hamamatsu Corp. We proposed that it could be a suitable photo-sensor for next-generation time-of-flight PET detectors due to mainly its high photon detection efficiency. Therefore, we concentrated on the measurement of coincidence timing performance of various MPPCs in conjunction with LYSO crystal scintillators. With 3mm × 3mm devices of 50µm sub-pixels coupled to 3mm × 3mm × 10mm crystals, we have demonstrated a strong dependence of timing performance on over-voltage and temperature, and analyzed how changes in photon detection efficiency and dark counts would explain the measurements. The best coincidence timing resolution we have achieved between two identical LYSO/MPPC detectors was 240ps in FWHM. This was worse than the timing resolution of 220ps obtained with Hamamatsu H6533 fast PMT, and contradicted the expected improvement from higher photon detection efficiency. The contradiction could be explained by slow rise-time of MPPC pulse shape, transit time spread, dark counts and electronics noise from large capacitance of MPPC. In particular, the slow rise-time of MPPC pulse suggested that the need for a very low trigger threshold on the timing circuit. Since it in turn makes the detector system more sensitive to noise, this poses additional challenges for ganging multiple devices together into a commercially viable time-of-flight PET block detector. We will discuss it in detail including other challenge imposed by MPPC characteristics.

Multi-Pixel Photon Counters for TOF PET Detector and Its Challenges

The Multi-Pixel Photon Counter (MPPC) is a Geiger-mode avalanche photo-diode (APD) developed by Hamamatsu Corp. We propose that it could be a suitable photo-sensor for next-generation time-of-flight PET detectors due to its high photon detection efficiency. We concentrate on the measurement of coincidence timing performance of various MPPCs in conjunction with LYSO crystal scintillators. With 3 mm times 3 mm devices of 50 mum sub-pixels coupled to 3 mm times 3 mm times 10 mm LYSO crystals, we have demonstrated a strong dependence of timing performance on over-voltage and temperature, and analyzed how changes in photon detection efficiency and dark counts would explain the measurements. The best coincidence timing resolution we have achieved between two identical LYSO/MPPC detectors was 240 ps in FWHM. This was worse than the timing resolution of 220 ps obtained with a Hamamatsu H6533 fast PMT, and contradicted the expected improvement from higher photon detection efficiency. The contradiction could be explained by the slow rise-time of MPPC pulse shape, transit time spread, dark counts and electronic noise from the large capacitance of the MPPC. In particular, the slow rise-time of the MPPC pulse suggests the need for a very low trigger threshold on the timing circuit. Since this makes the detector system more sensitive to noise, this poses additional challenges for ganging multiple devices together into a commercially viable time-of-flight PET block detector. We will discuss this work in detail including other challenge imposed by MPPC characteristics.

Time-of-flight PET-MR detector development with silicon photomultiplier

Silicon photomultiplier technology based on Geigermode avalanche in p-n junction has made unforeseen progress for the last 5-6 years. With its high gain and high photon detection efficiency, it has shown that it could replace PMTs in many applications including time-of-flight PET. Also, its magnetic immunity and compactness made it very suitable for PET-MR detector. In this paper, we present time-of-flight PET-MR detector based on silicon photomultiplier from its selection, prototype test with discrete electronics, ASIC and a module design. Also, the system performance data is presented.

Sensitivity improvement of time-of-flight (ToF)-PET detector through recovery of Compton scattered annihilation photons

In PET detector designs, the scintillator material can be partitioned such that a block of crystals share timing/energy readout electronics. A fraction of the incident 511 keY photons produce simultaneous events in two adjacent block readouts due to Compton scattering followed by escape from the primary block. These inter-block Compton scatter events are typically not processed in current PET scanners. We have used radiation transport simulations to determine the fraction of inter-block Compton events for different block sizes using LYSO scintillator. The simulations showed the statistical distribution of the energy signals and time stamps in the two blocks and guided the selection of energy and time criteria for an event recovery algorithm. With a particular block size, we experimentally demonstrated that inter-block Compton events may be recovered as valid events with a corrected time stamp and estimated position of initial interaction. The results showed a significant improvement in detector sensitivity at the expense of a small degradation in the timing resolution of the detector block.

Timing resolution performance comparison for fast and standard outputs of SensL SiPM

A new silicon photomultiplier (SiPM) with a unique fast output signal feature was recently introduced by SensL. For a SiPM device with 3×3 mm2 sensitive area, the single photo electron response from the fast output is less than 2 ns wide compared with 50 ns width observed at the standard output. Using the fast readout signal, a coincidence resolving time (CRT) of 190 ps (FWHM - full width at half maximum) was measured with a 3×3×10 mm3 LYSO crystal optically coupled to the SiPM mounted on the SensL MicroFB-SMA evaluation board. Applying additional C-R high pass filters to the standard SiPM output, we shortened the SPE response from 50 ns to 3 ns width and measured CRT values similar to those obtained with fast output signals. We have studied the effect of external filtering on CRT with two SensL 3×3 mm2 devices with different microcell sizes (35 μm and 50 μm). For both devices studied, we demonstrated similar CRT performance between the fast output and standard output with optimized pulse shaping filter. However, for applications requiring simultaneous measurement of intensity and timing of light pulses, the availability of a separate fast output directly from the sensor is convenient because it reduces the need for additional circuits and provides CRT performance equal to external filters optimized for the standard outputs.

Dependence of timing resolution on crystal size for TOF PET

Recently, with the prospect of great improvement in image quality, the development of time of flight technology has become an exciting topic for positron emission tomography. The excitement was further accelerated by the introduction of various fast and high light output scintillators as well as photosensors. However, the development of improved time of flight detectors is not only about the selection of crystals and photosensors, but also about how detectors are assembled to optimize their performance. For example, depending on crystal block structure, photo-sensor layout, and coupling methods, a detectors timing resolution can be drastically different. Since the effect of block structure for timing resolution is complex and less understood it is essential to first dissect the block structure and understand the impact of its basic components on timing resolution. In this paper, we will present the dependence of timing resolution on varying the dimensions of the scintillator crystals that are the main component of a block detector.

Multiplexing requirements for solid state photomultipliers in time-of-flight PET

The solid state photomultiplier (SSPM), an array of Geiger-mode APDs, developed by Hamamatsu Corp. has been evaluated for time-of-flight PET detectors. As it was demonstrated in previous work the SSPM is a promising photo sensor to achieve very good timing resolution for PET applications. Due to relatively small size of individual sensor (3×3 mm2) a large number of readout channels will be required in applications such as a whole body PET scanner. The obvious solution for this problem is multiplexing several devices into a single readout channel. To better understand the limits and trade offs involved for timing resolution a detailed analysis of the effects of dark current, amplifier noise, bandwidth and amplifier input impedance was done.

Effect of microcell size on timing performance of silicon photomultipliers for ToF-PET imaging

The timing resolution of a Silicon Photomultiplier (SiPM)-based scintillator detector depends on its ability to detect optical photons in the first ns after a scintillation event and is thus directly proportional to its photon detection efficiency (PDE). As a result, an SiPM with a larger microcell size, and thus larger fill factor and PDE, is expected to demonstrate better timing resolution. However, coincidence timing resolution measurements we performed with Hamamatsu 3×3 mm2 SiPMs with 25, 50 and 100 um microcells showed that the timing resolution of larger microcell devices are adversely impacted by larger dark count rates. A simple high pass filter can be used to reduce the impact of the dark counts on the baseline of an SiPM output and thus improve the timing resolution to reflect what is expected from a device with larger fill factor and PDE.

Design of a Modular and Efficient CAMAC/Lab VIEW-Based Data Acquisition System for a Time of Flight PET Test-Bed

A high-speed data acquisition (DAQ) system has been designed for a time of flight PET test-bed. The requirements of the system were flexibility, data throughput and data integrity. The software is modular so that modifications and additions can be integrated easily into the existing software architecture. The program operation is driven by commands read from a script file, simplifying implementation of complex acquisition sequences. The heart of the program is the DAQ module, which efficiently transfers data from CAMAC to file. Another software module offers online or offline analysis capabilities. The software, written in LabVIEW, communicates with a novel high-speed USB2 CAMAC controller (CCUSB). The CCUSB offers significant improvements over its GPIB predecessor, supporting FIFO buffered DAQ and a variety of data readout modes. Four readout modes have been evaluated in order to maximize the DAQ rate for this particular system. A highest sustained data rate of 15.7 k events/s was achieved for approximately 60 input channels using a 22Na flood phantom. Flexibility in the software design accommodates both current and future hardware configurations without the need to edit the LabVIEW code.

Simulation study for designing a compact brain PET scanner

Summary form only given. The desire to understand normal and disordered human brain of upright, moving persons in natural environments motivates the development of an ambulatory micro-dose brain PET imager (AMPET) [1]. An ideal system would be light weight and have high sensitivity and spatial resolution. These requirements are often in conflict with each other. Therefore, we performed simulation studies to search for the optimal system configuration and to evaluate the improvement in performance over existing scanners. An intuitive design to achieve high sensitivity is to use a tight geometry that covers the brain. However, a tight geometry also increases parallax error in peripheral lines of response, which may increase the variance in ROI quantification. In this study, we first simulated cylindrical PET with different ring diameters. All PET configurations are subjected to the same maximum weight constraint by restricting the amount of detector materials. We computed the Cramér-Rao variance bound, which is the lower bound of the variance for an unbiased estimator, to compare the performance for region of interest (ROI) quantification using different scanner geometries. The results show that while a smaller ring diameter can increase photon detection sensitivity and hence reduce the variance in the center of the field of view, it can result in higher pixel variance in peripheral regions when the length of detector crystal is 15 mm or more. The variance can be substantially reduced by adding depth of interaction (DOI) measurements to the detectors. Our simulation study also shows that the relative performance highly depends on the size of the ROI, and a large ROI favors a tighter geometry even without DOI information. Based on the 2D simulation results, we proposed a helmet scanner design with DOI detectors as shown in Fig. 1, which is similar to the design in [2]. This helmet scanner consists of three parts: a top panel, side rings with varying diameters, and a bottom panel. We used the Siemens brain MR-PET scanner geometry as a reference for comparison. The detector block parameters and the diameter of the bottom ring for the helmet scanner are the same as the reference cylinder scanner. Parameters of the side rings of the helmet scanner are listed in Table I. The bottom panel contains 4×4 detector blocks and the top panel contains 52 detector blocks. Distance between the bottom flat panel and the bottom ring is about 160 mm and the axial gap between the top panel and the top ring is 2.5 mm. GATE V6.2 [3] was used to perform Monte Carlo simulations. GATE simulation results of the cylindrical scanner and the helmet with side rings only were cross-validated by SimSET simulation results [4]. The results showed that the sensitivity of the helmet scanner is about 4 times that of the reference cylindrical scanner. The sensitivity improvement is also position dependent. The bottom panel mainly improves the sensitivity in the lower portion of the scanner FOV, while the top panel mainly improves the sensitivity in the upper portion of the FOV. The maximum improvement is near the top with a gain factor up to 35. Reconstructions of the simulated Hoffman phantom [5] data showed that the helmet scanner can substantially improve the image quality over the reference cylindrical scanner.