Tianpeng Xu

A novel sub-millimeter resolution PET detector with TOF capability

We have developed a PET detector module based on sub-millimeter lutetium oxyorthosilicate (LSO) crystal block and prototype Silicon Photomultiplier (SiPM) arrays with additional fast timing output. The LSO array consists of 20×20 crystals of which each pixel is 0.78 × 0.78×3 mm3 and the pitch is 0.8 mm due to the reflector between pixels. The FM SiPM array and FB SiPM array are arranged 4×4 with 4×4mm2 size for each pixel and 0.2mm gap between adjacent pixels. Through dual readout of both fast and standard output signals from this new SiPM, good energy and timing performance can be achieved at the same time. Through multiplexing and summing scheme and charge division resistor network, sixteen channels of fast output and sixteen channels of standard output are reduced to one fast timing signal and four energy signals, making the detector block compact and scalable in building PET scanners. The performance of FM and FB detector module was evaluated. 16×16 LSO crystals were resolved in the flood image and the edge crystals could be identified if a light guide was used. The average energy resolution of FM detector and FB detector were 15.3% ±1.4% and 11.3% ± 0.7%,respectively, at 511 KeV. Through time pick-off method of leading edge discriminator (LED), the average coincidence resolving timing (CRT) of the block against a 3x3x20mm3 coupled with a FB30035 pixel were 401± 10 ps for FM detector and 303 ± 16 ps for FB detector. The average block-to-block CRT were estimated to be 495 ps for FM detector and 324 ps for FB detector. The FB array has demonstrated a better performance due to its much higher PDE at 420nm wavelength than FM array. We conclude that the proposed design is feasible for developing sub-millimeter resolution TOF PET detectors.

Development of multi-channel fast SiPM readout electronics for clinical TOF PET detector

Silicon photomultipliers (SiPM) are rapidly adapted in simultaneous time of flight (TOF) PET and MRI imaging system in recent years due to its fast timing response and immunity to magnetic field. The aim of this work is to develop multi-channel fast SiPM readout electronics for clinical TOF PET detector. The detector module is assembled by directly coupling a 12×12 LYSO crystal block to an 8×8 Sensl FB30035 SiPM array. We are developing two approaches of compact readout electronics to handle the large number of SiPM output channels. One is a low noise and high bandwidth, 128-channel ASIC to readout the 64 fast outputs and 64 standard outputs of the 8×8 SiPM array. The other approach is a conventional electronics based multiplexing readout scheme using discrete components. A charge division resistor network reduced 64 channels of standard output to four energy signals, and multiplexing and summing circuits reduced 64 channels of fast output to one timing signal, making the detector block highly compact and scalable. The ASIC design has been finished and its ready for tape-out. Preliminary performance of the detector module readout by conventional electronics was evaluated. 12×12 LYSO crystals were well resolved and identified in the flood image and the average energy resolution was 15.6% ±1.3%. For single crystal pair, 240ps coincidence resolving timing (CRT) was achieved. It was dropped down to 500ps for detector block to single crystal configuration, due to the combination effects of: 1) degradation of fast signal rising edge by output capacitance addition of SiPMs and 2)large white electronic noise and ultrahigh dark count rate. The readout design will be further optimized by increasing the signal bandwidth and reducing noise to get better timing performance at a cost of increasing circuit complexity to a reasonable level. The performance of ASIC readout will be evaluated and good timing performance is expected due to the individual pixel timing pick-off.

Optical simulation of a 9×9 LYSO block detector with PQS technology using GATE

Design of the light sharing mechanisms has a significant influence on the performance of PET block detector. Monte Carlo simulation plays an important role in PET detector design and optimization. In this work, GATE is used to simulate the transportation processes of gamma rays and scintillated optical photons in a PMT-based 9×9 L YSO block detector of a self-developed animal PET system. The aim of this work is to optimize the optical parameters of the block detector design. Scintillation photons are generated and transported in the L YSO crystal array and finally detected by the 4 PMTs. Based on 2D Anger logic, the position of the crystal from which optical photons originate is calculated from the number of optical photons detected by the 4 PMTs. Different simulation parameters such as crystal side and bottom roughness and optical coupling pattern between crystals and reflecting films are tested in the simulation and 2D-MAPs of the detector are obtained. By comparing the 2D-MAPs acquired, we conclude that crystal side surface roughness has a significant influence on PET detector performance and a relatively rougher crystal side surface is recommended for applications while the detector performance is not sensitive to crystal bottom surface roughness and optical coupling pattern between crystals and reflecting films is key parameter for PET detector design as it determines the amounts of light sharing. The developed method is feasible for facilitating PET detector design and performance prediction.

Dedicated brain PET system of PET/MR for brain research

This work is to replace PET ring in human brain PET/MR system with a dedicated wearable PET insert, aimed at improving both patient feasibility and system performance for brain imaging. The designed PET/MR system includes two parts: the inside parts, including a radio frequency (RF) coil and PET ring, are mounted on patient’s head, and the outside part, a MR imager, is dependent of patient. The RF coil is the innermost layer, surrounded by an outer PET-ring layer. They are supported by a MRcompatible structure. And both RF coil and PET detectors are placed inside a standard clinical 3-T MR imager. From the design of the system we can infer that some advantages can be achieved. First, high sensitivity will be achieved with the same amount crystals as the PET ring is more close to region-of-interest area, at a reduced cost. Second, by using a 2-layer depth of interaction (DOI) detector, the parallax effect can be minimized. The resolution will benefit from short positron range caused by magnetic field and smaller ring diameter will also reduce the effect of non-collinearity. Thirdly, as the PET ring is mounted on head, impact of patient motion will be reduced.

Development of an MR-compatible DOI-PET detector module

Silicon Photomultiplier (SiPM) is a promising sensor for MR-compatible PET systems. In this paper, we developed a compact 2-layer DOI-PET detector. The top layer is a 15A—15 LYSO array, and the crystal size is 2x2x7mm3. The bottom layer is a 16A—16 array with the same size crystals. There is half-crystal offset between two layers in both transverse directions. The detector is coupled to an 8A—8 SiPM array (MicroFB-30035-SMT, Sensl). Sixty-four channels of SiPMs are read out by an ASIC chip with in-chip multiplexing resistor networks in the form of two position and one energy analog signals, and are then converted to wave-form digital signals with 80 MHz 12-bit ADC chips. The energy is calculated by averaging the 3 points around the peak of the pulse. Flood images with two 22Na point sources irradiated on the top and at the bottom of the detector module were acquired. The results show that the detector module achieves good crystal identification capability in both layers with an average energy resolution of 17.1% at 511 keV.

SiPM based PET detector modules with air-gapped pixelated LYSO

Stticon Photomultiplier (SiPM) is a promising sensor for PET with a number of advantages. A traditional design for PET detector module is using a pixelated scintillator array coupled to a SiPM array, and it is critical to design reflectors between crystals to decode photon event location. Unlike PMT based detectors, the pixelated scintillator array may not necessarily require complex reflector schemes for optimal light sharing. An intuitive design method is to fill the gaps between adjacent crystals with reflective materials such as ESR or BaSO4. However, the manufacture of the array with reflectors is still time-consuming and costly. In this paper, we evaluate the performances of pixelated LYSO arrays separated by only air-gap. A mathematical calculation shows that there is not much difference between reflector based and air-gap based detector module design. Three groups of experiments were conducted based on a home-made 8×8 SiPM array, including a single LYSO crystal with three types of reflector process, three 8×8 LYSO arrays and two 15×15 LYSO arrays with different crystal surface treatments and with air or reflective-material gaps. The results show that the air-gap design for polished LYSO has similar event decoding capability and energy resolution compared with the reflector-filled design. In conclusion, the scintillator arrays with air-gap are simple and feasible for SiPM based detector, which would greatly reduce the cost and manufacture time of detector modules without performance deterioration.

Crystal identification for a dual-layer-offset LYSO based PET system via Lu-176 background radiation and mean shift algorithm

Crystal identification is a critical procedure for PET systems. In this paper, we proposed a crystal identification method for a dual-layer-offset LYSO based animal PET system via Lu-176 background radiation. Single event data of Lutetium background radiation were acquired in list-mode for one hour to generate a single photon flood histogram (SPH). Coincidence events were retrieved from the same data using time information to generate a coincidence flood histogram (CFH). The coincidence flood map was employed to identify the peaks of the inner layer responses using average peaks calculation and mean-shift algorithm. The response of the inner layer was removed from the SPH using the CFH and then the peaks of the outer layer were also identified using mean-shift algorithm. At last, the crystal position map was generated using these peaks with the distance criteria. Results show that the proposed method can be employed for the dual-layer offset PET system to implement crystal identification instead of using external radiation sources.

Development of a MR-compatible DOI-TOF detector module for PET imaging systems

PET systems are widely used in clinical and preclinical molecular imaging applications. Various PET systems, including whole-body human and small animal PET, and dedicated systems such as breast and brain PET have been developed. A Standardized PET detector module with good performance will make it much easier and more efficient for different kind of PET scanner instrumentation. In this paper, we developed a compact MR-compatible PET detector with DOI and TOF capability. The detector module has staggered 15×15 +16 ×16 LYSO crystal array with single crystal size of 2×2×7mm3. The detector is coupled to SiPM array (MicroFJ-30035-TSV, SensL). The output signals of SiPMs are multiplexed by 64-channel ASIC chips and then digitized by 80MHz 12-bit ADC chips and TDC implemented inside FPGA to generate energy and timing information. The performance of the detector was preliminarily evaluated. A flood image was acquired and all crystals were clearly identified with an average energy resolution of 12.7% for 15×15 array and 14.2% for 16×16 array. About 300 ps coincidence timing resolution was achieved for a pair of single crystal. More accurate energy resolution and coincidence timing resolution for the detector module will be evaluated in the future. And the MR compatibility evaluation of this module is underway. Large quantities of detector blocks will be made and tested to verify the performance stability and reproducibility. Furthermore, we are developing standardized power and signal interface for better usability.

3D positioning and readout channel number compression methods for monolithic PET detector

Monolithic PET detector design is promising for low cost, high packing fraction, intrinsic depth-of-interaction (DOI) capability as well as potentially better energy and timing resolution compared to the conventional segmented crystal-based detector design. In this work, we developed a monolithic detector with 3D positioning capability and readout channel number compression. The detector contains a monolithic LYSO crystal (size: 25 mm × 25 mm × 14 mm) coupled to a 6 × 6 SiPM array (SensL FC30035). Two 3D positioning algorithms including statistical-based positioning (SBP) and artificial neural network (ANN) were developed to estimate gamma ray interaction positions inside the crystal. To simplify detector electronics design, two readout channel number compression strategies respectively based on Anger and principal component analysis (PCA) were proposed. With Anger, the readout channel number was compressed from 36 to 12 while with PCA the readout channel number was respectively compressed to 12 (PCA12), 6 (PCA6) and 4 (PCA4) corresponding to using different number of principal component vectors. The results demonstrate that 3D positioning capability is achievable with both SBP and ANN and the positioning resolution in x, y and z direction are respectively 2.58 mm, 2.56 mm and 4.79 mm with SBP, and 2.02 mm, 2.03 mm and 3.96 mm with ANN. Both strategies can effectively reduce the readout channel number without inducing degradations in detector positioning performance. With Anger and PCA12, almost no degradation is observed in positioning performance compared to no compression. With PCA6 or PCA4, deteriorations in positioning performance in z direction are observed, especially with PCA4, although there is no significant degradation in × or y direction. We conclude that the developed monolithic detector is promising for its 3D positioning capability as well as engineering application value with compressed readout channel number in PET system designs.

Optimization of dual mode readout of Sensl SiPM for TOF PET detectors

A novel silicon photomultipliers (SiPM) has been developed with both a standard and a fast output, aiming at providing good energy and timing performance simultaneously for time of flight (TOF) PET application. Optimal readout method of this SiPM is investigated in this study. The standard output can be taken from either anode terminal or cathode terminal and different dual readout mode has different impact on timing resolution of fast output. Three dual mode readout methods were evaluated and we found the optimized way was to readout the standard output on the cathode side with a small capacitance in parallel. This readout method can preserve good timing performance using relatively few electronic components. We developed a detector module by directly coupling a 16×16 LYSO crystal block to an 8×8 Sensl FC30035 SiPM array. Through multiplexing and summing scheme and charge division resistor network, 64 channels of fast output and 64 channels of standard output are reduced to one fast timing signal and four energy signals, making the detector highly compact and scalable. The performance of the detector module was evaluated. 16×16 LYSO crystals were well resolved in the flood image and the average energy resolution were 12.4% ± 0.8% at 511 KeV. The average coincidence resolving timing (CRT) of the block against a single LYSO crystal coupled to a FC30035 pixel were 286 ± 13 ps. The average block-to-block CRT were estimated to be 311 ps. The proposed dual readout design has demonstrated excellent timing performance for TOF PET detectors.