Evaluation of Matrix9 silicon photomultiplier array for small-animal PET

PET Detectors Based on Multi-Resolution SiPM Arrays

Almost all high spatial resolution positron emission tomography (PET) detectors based on pixelated scintillator arrays utilize crystal arrays with smaller pitches than photodetector arrays, leading to challenges in resolving edge crystals. To address this issue, this paper introduces a novel multi-resolution silicon photomultiplier (SiPM) array design aimed at decreasing the number of readout channels required while maintaining the crystal resolvability of the detector, especially for edge crystals. The performance of a pseudo 9 × 9 multi-resolution SiPM array, consisting of 6.47 × 6.47 mm2, 6.47 × 3.07 mm2, and 3.07 × 3.07 mm2 SiPMs, was compared to those of a pseudo 8 × 8 SiPM array with a 6.8 mm pitch, and a 16 × 16 SiPM array with a 3.4 mm pitch using a 36 × 36 LYSO array with a pitch of 1.5 mm. The large-size pseudo SiPMs were implemented by digitally grouping multiple 3.07 × 3.07 mm2 SiPMs. The flood histograms show that the edge crystal resolvability of the pseudo 9 × 9 multi-resolution SiPM array is comparable to that of the 16 × 16 SiPM array and is significantly better than that of the 8 × 8 SiPM array.

A high resolution and high detection efficiency depth-encoding detector for brain positron emission tomography based on a 0.75 mm pitch scintillator array

The quantitative accuracy and precision of brain positron emission tomography (PET) studies can be considerably improved using dedicated brain PET scanners with a uniform high resolution and a high sensitivity across the brain volume. One approach to building such a system is to construct the PET scanner using depth-of-interaction (DOI) encoding detectors with finely segmented and thick crystal arrays. In this paper, the performance of a DOI PET detector based on two 16 × 16 arrays of 2 × 2 mm2 SiPMs coupled to both ends of a 44 × 44 array of 0.69 × 0.69 × 30 mm3 polished LYSO crystals was evaluated at different temperatures (-9°C, 0°C, 10°C, and 20°C) for brain PET applications. The pitch size of the LYSO array is 0.75 mm. The flood histograms show that all the crystal elements in the LYSO array can be resolved except some edge crystals, due to the limited light sharing. The average energy resolution, average DOI resolution, and average timing resolution across crystal elements are 21.1 ± 3.0%, 3.47 ± 0.17 mm, and 1.38 ± 0.09 ns, respectively, which were obtained at a bias voltage of 56.5 V and a temperature of 0°C.

Energy spectra due to the intrinsic radiation of LYSO/LSO scintillators for a wide range of crystal sizes

PURPOSE

Most detectors in current positron emission tomography (PET) scanners and prototypes use lutetium oxyorthosilicate (LSO) or lutetium yttrium oxyorthosilicate (LYSO) scintillators. The aim of this work is to provide a complete set of background energy spectra, due to the scintillator intrinsic radioactivity, for a wide range of crystal sizes.

METHODS

An analytical model, developed and validated in a previous work, was used to obtain the background energy spectra of square base cuboids of different dimensions. The model uses the photon absorption probabilities of the three gamma rays (88, 202, and 307 keV) emitted following the beta decay of ¹⁷⁶ Lu to ¹⁷⁶ Hf excited states. These probabilities were obtained for each crystal size considered in this work from Monte Carlo simulations using the PENELOPE code. The probabilities are then used to normalize and shift the beta spectrum to the corresponding energy value of the simultaneous detection of one, two, or three gamma rays in the scintillator. The simulated cuboids had side lengths of 5, 10, 20, 30, 40, 50, and 60 mm and crystal thickness T = 2.5, 5, 10, 15, and 20 mm. From these results a complete set of energy spectra, including intermediate dimensions, were obtained. In addition, LYSO and LSO were compared in terms of their analytical background energy spectra for two crystal sizes. The analytical spectra were convolved using a variable Gaussian kernel to account for the energy resolution of a typical detector. A parameterization of the photon absorption probabilities for each gamma ray energy as a function of the cuboid volume to surface area ratio was obtained.

RESULTS

A data set of L(Y)SO background energy spectra was obtained and is available for the reader as 2D histograms. The model accurately predicts the structure of the energy spectra including the relative peak and valley intensities. The data allow visualizing how the structure evolves with increasing crystal length and thickness. Lutetium yttrium oxyorthosilicate and LSO present very similar background energy spectra for the range of sizes studied in this work and therefore the data generated can be confidently used for both scintillator materials. The filtered spectra showed a variable shift in the main peaks, depending on crystal size, alerting that to achieve a correct detector calibration using the background spectrum is not straight forward and requires precise data analysis and measurements. In addition, we found that square base L(Y)SO cuboids with same volume to surface area ratio have background spectra with the same structure.

CONCLUSIONS

We present the energy spectra of L(Y)SO crystal of different sizes which will be very useful for industry and research groups developing and simulating detectors for positron imaging applications in terms of calibration, quality assurance, crystal maps, detector fine gain tuning, background reduction and other applications using the long-lived ¹⁷⁶ Lu source. We analyzed the data produced in this work and found that crystal cuboids with equal volume to surface area ratio produce the same background energy spectra, a conclusion that simplifies its calculation and clarifies why the same energy spectrum is observed under different experimental setups.

Offset compensated baseline restoration and computationally efficient hybrid interpolation for the brain PET

The acquisition of positron emission tomography (PET) pulses introduces artifacts and limits the performance of the scanner. To minimize these inadequacies, this work focuses on the design of an offset compensated digital baseline restorer (BLR) along with a two-stage hybrid interpolator. They respectively treat the incoming pulse offsets and limited temporal resolution and improve the scanner performance in terms of calculating depth of interactions and line of responses. The offset of incoming PET pulses is compensated by the BLR and then their interesting parts are selected. The selected signal portion is up-sampled with a hybrid interpolator. It is composed of an optimized weighted least-squares interpolator (WLSI) and a simplified linear interpolator. The processes of calibrating the WLSI coefficients and characterizing the BLR and the interpolator modules are described. The functionality of the proposed modules is verified with an experimental setup. Results have shown that the devised BLR effectively compensates a dynamic range of bipolar offsets. The signal selection process allows focusing only on the relevant signal part and avoids the unnecessary operations during the post-interpolation process. Additionally, the hybrid nature allows improving the signal temporal resolution with an appropriate precession at a reduced computational complexity compared to the mono-interpolation-based arithmetically complex counterparts. The component-level architectures of the BLR and the interpolator modules are also described. It promises an efficient integration of these modules in modern PET scanners while using standard and economical analog-to-digital converters and field-programmable gate arrays. It avoids the development of high-performance and expensive application-specific integrated circuits and results in a cost-effective realization.

A depth-of-interaction encoding PET detector module with dual-ended readout using large-area silicon photomultiplier arrays

The performance of a depth-of-interaction (DOI) encoding PET detector module with dual-ended readout of LYSO scintillator arrays using large-area SiPM arrays was evaluated. Each SiPM array, with a surface area of 50.2  ×  50.2 mm², consists of 12  ×  12 C-series SiPMs from SensL (SensL, Inc). The LYSO array, with a total size of 46  ×  46 mm² and a pitch size of 1.0 mm, consists of a 46  ×  46 array of 0.945  ×  0.945  ×  20 mm³ polished LYSO crystals, separated by Toray reflector. Custom front-end electronics were designed to reduce the 288 SiPM signals of one detector module to nine signals, eight for position information and 1 for timing information. Schottky diodes were used to block noise from SiPMs that did not detect a significant number of scintillation photons following a gamma interaction. Measurements of noise, signal, signal-to-noise ratio, energy resolution and flood histogram quality were obtained at different bias voltages (26.0 to 31.0 V in 0.5 V intervals) and at two temperatures (5 °C and 20 °C). Clear acrylic plates, 2.0 mm thick, were used as light guides to spread the scintillation photons. Timing resolution, depth of interaction resolution, and the effect of event rate on detector performance were measured at the bias voltage determined to be optimal for the flood histograms. Performance obtained with and without the noise-blocking Shottky diodes was also compared. The results showed that all crystals in the LYSO array can be clearly resolved, and performance improved when using diodes to block noise, and at the lower temperature. The average energy resolution, flood histogram quality, timing resolution and DOI resolution were 23.8%  ±  2.0%, 1.54  ±  0.17, 1.78  ±  0.09 ns and 2.81  ±  0.13 mm respectively, obtained at a bias voltage of 30.0 V and a temperature of 5 °C using the diode readout method. The event rate experiments showed that the flood histogram and energy resolution of the detector were not significantly degraded for an event rate of up to 150 000 counts s^-1.

Performance of a high-resolution depth-encoding PET detector module using linearly-graded SiPM arrays

The goal of this study was to exploit the excellent spatial resolution characteristics of a position-sensitive silicon photomultiplier (SiPM) and develop a high-resolution depth-of-interaction (DOI) encoding positron emission tomography (PET) detector module. The detector consists of a 30  ×  30 array of 0.445  ×  0.445  ×  20 mm3 polished LYSO crystals coupled to two 15.5  ×  15.5 mm2 linearly-graded SiPM (LG-SiPM) arrays at both ends. The flood histograms show that all the crystals in the LYSO array can be resolved. The energy resolution, the coincidence timing resolution and the DOI resolution were 21.8  ±  5.8%, 1.23  ±  0.10 ns and 3.8  ±  1.2 mm, respectively, at a temperature of -10 °C and a bias voltage of 35.0 V. The performance did not degrade significantly for event rates of up to 130 000 counts s-1. This detector represents an attractive option for small-bore PET scanner designs that simultaneously emphasize high spatial resolution and high detection efficiency, important, for example, in preclinical imaging of the rodent brain with neuroreceptor ligands.

A Time-Walk Correction Method for PET Detectors Based on Leading Edge Discriminators

The leading edge timing pick-off technique is the simplest timing extraction method for PET detectors. Due to the inherent time-walk of the leading edge technique, corrections should be made to improve timing resolution, especially for time-of-flight PET. Time-walk correction can be done by utilizing the relationship between the threshold crossing time and the event energy on an event by event basis. In this paper, a time-walk correction method is proposed and evaluated using timing information from two identical detectors both using leading edge discriminators. This differs from other techniques that use an external dedicated reference detector, such as a fast PMT-based detector using constant fraction techniques to pick-off timing information. In our proposed method, one detector was used as reference detector to correct the time-walk of the other detector. Time-walk in the reference detector was minimized by using events within a small energy window (508.5 - 513.5 keV). To validate this method, a coincidence detector pair was assembled using two SensL MicroFB SiPMs and two 2.5 mm × 2.5 mm × 20 mm polished LYSO crystals. Coincidence timing resolutions using different time pick-off techniques were obtained at a bias voltage of 27.5 V and a fixed temperature of 20 °C. The coincidence timing resolution without time-walk correction were 389.0 ± 12.0 ps (425 -650 keV energy window) and 670.2 ± 16.2 ps (250-750 keV energy window). The timing resolution with time-walk correction improved to 367.3 ± 0.5 ps (425 - 650 keV) and 413.7 ± 0.9 ps (250 - 750 keV). For comparison, timing resolutions were 442.8 ± 12.8 ps (425 - 650 keV) and 476.0 ± 13.0 ps (250 - 750 keV) using constant fraction techniques, and 367.3 ± 0.4 ps (425 - 650 keV) and 413.4 ± 0.9 ps (250 - 750 keV) using a reference detector based on the constant fraction technique. These results show that the proposed leading edge based time-walk correction method works well. Timing resolution obtained using this method was equivalent to that obtained using a reference detector and was better than that obtained using constant fraction discriminators.

Performance Comparison of Different Readouts for Position-Sensitive Solid-State Photomultiplier Arrays

A thorough comparison of five different readouts for reading out a 2 × 2 array of 5 mm × 5 mm position-sensitive solid-state photomultipliers (PS-SSPM) was undertaken. The five readouts include reading out the 20 signals (16 position and 4 timing) individually, two signal multiplexing readouts, and two position decoding readouts. Flood histogram quality, signal-to-noise ratio (SNR) and energy resolution were compared at different bias voltage (27.0 V to 32.0 V, at 0.5 V intervals) and at a fixed temperature of 0 °C by coupling a 6 × 6 array of 1.3 mm × 1.3 mm × 20 mm polished LSO crystals to the center of the PS-SSPM array. The timing resolution was measured at a bias voltage of 31.0 V (optimal bias voltage in terms of flood histogram quality). The best flood histogram quality value and signal-to-noise were 7.3 ± 1.6 and 33.5 ± 3.1, respectively, and were obtained by shaping and digitizing the 16 position signals individually. The capacitive charge-division readout is the simplest readout among the five evaluated but still resulted in good performance with a flood histogram quality value of 3.3 ± 0.4 and a SNR of 18.3 ± 1.3. The average energy resolution and the average timing resolution were 15.2 ± 1.2 % and 8.4 ± 1.6 ns for individual signal readout and 15.9 ± 1.2 % and 8.8 ± 1.3 ns by using the capacitive charge-division readout method. These studies show that for an ultra-high spatial resolution applications using the 2 × 2 PS-SSPM array, reading out the 20 signals individually is necessary; whilst the capacitive charge-division readout is a cost-effective readout for less demanding applications.

Characterization of Large-Area SiPM Array for PET Applications

The performance of an 8 × 8 array of 6.0 × 6.0 mm² (active area) SiPMs was evaluated for PET applications using crystal arrays with different pitch sizes (3.4 mm, 1.5 mm, 1.35 mm and 1.2 mm) and custom designed five-channel front-end readout electronics (four channels for position information and one channel for timing information). The total area of this SiPM array is 57.4 × 57.4 mm², and the pitch size is 7.2 mm. It was fabricated using enhanced blue sensitivity SiPMs (MicroFB-60035-SMT) with peak spectral sensitivity at 420 nm. The performance of the SiPM array was characterized by measuring flood histogram decoding quality, energy resolution, timing resolution and saturation at several bias voltages (from 25.0 V to 30.0 V in 0.5 V intervals) and two different temperatures (5 °C and 20 °C). Results show that the best flood histogram was obtained at a bias voltage of 28.0 V and 5 °C and an array of polished LSO crystals with a pitch as small as 1.2 mm can be resolved. No saturation was observed up to a bias voltage of 29.5 V during the experiments, due to adequate light sharing between SiPMs. Energy resolution and timing resolution at 5 °C ranged from 12.7 ± 0.8% to 14.6 ± 1.4 % and 1.58 ± 0.13 ns to 2.50 ± 0.44 ns, for crystal array pitch sizes of 3.4 mm and 1.2 mm respectively. Superior flood histogram quality, energy resolution and timing resolution were obtained with larger crystal array pitch sizes and at lower temperature. Based on our findings, we conclude that this large-area SiPM array can serve as a suitable photodetector for high-resolution small-animal PET or dedicated human brain PET scanners.