Performance evaluation of the mindview PET using GATE and STIR

The scintillation light distribution produced by photodetectors in positron emission tomography (PET) provides the depth of interaction (DOI) information required for high-resolution imaging. The goal of positioning techniques is to reverse the photodetector signal’s pattern map to the coordinates of the incident photon energy position. By considering the DOI information, monolithic crystals offer good spatial, energy, and timing resolution along with high sensitivity. In this work, a supervised deep neural network was used for the approximation of DOI and to assess through Monte Carlo (MC) simulations the performance on a small-animal PET scanner consisting of ten 50 × 50 × 10 mm3 continuous Lutetium-Yttrium Oxyorthosilicate doped with Cerium (LYSO: Ce) crystals and 12 × 12 silicon photomultiplier (SiPM) arrays. The scintillation position was predicted by a multilayer perceptron neural network with 256 units and 4 layers whose inputs were the number of fired pixels on the SiPM plane and the total deposited energy. A GEANT4 MC code was used to generate training and test datasets by altering the photons’ incident position, energy, and direction, as well as readout of the photodetector output. The calculated spatial resolutions in the X-Y plane and along the Z-axis were 0.96 and 1.02 mm, respectively. Our results demonstrated that using a multilayer perceptron (MLP)-based positioning algorithm in the detector modules, constituting the PET scanner, enhances the spatial resolution by approximately 18% while the absolute sensitivity remains constant. The proposed algorithm proved its ability to predict the DOI for depth under 7 mm with an error below 8.7%.

High spatial resolution Positron Emission Tomography (PET) imaging provides functional data about a patient's body and provides us with a range of new possibilities, including medicine development, new diagnostic methods, and others. The use of Depth of Interaction (DOI) detectors is one effective technique for the improvement of whole-body PET imaging, as is evident in the fact that most small animal PET images are taken using a DOI system. In this study, we developed a prototype whole-body DOI- equipped PET system and evaluated its performance. This scanner's DOI detectors consist of dual-layer (front and rear) 9 x 10 array of GSO/GSO crystals, a light guide and two rectangular double-anode photo multiplier tubes (PMTs). The DOI layers are discriminated according to decay time differences, which are controlled by Ce concentrations. The DOI detectors are arranged in a circular detector ring, with a diameter of 664 mm. In order to avoid parallax errors, coincidental events, including rear layer detection, were addressed to the nearest neighbor LOR front-layer coincidence pair. To evaluate the spatial resolution of this PET system at various radial positions, 22Na point sources at 1, 10, 20, and 25 cm offset from the center were reconstructed, both with and without DOI information. Results showed that spatial resolution degradation at 25 cm improved from 195 % to 154 % as a result of using DOI data. During a visual evaluation using a Derenzo phantom, image reconstruction using DOI data clarified the image and corrected hot rod shapes. The results presented here show that DOI detectors were useful in creating an effective whole-body PET system.

This contribution shows the feasibility of using thick monolithic scintillation crystals, in particular of the type LYSO, for the development of brain PET systems. We described the characterization of two main designs. The first is used in a PET alone system where the main field of application is Alzheimer Disease. In this case a 15 mm thick block was used with $50 \times 50$ mm $^{\mathbf {2}}$ size. The second design described in this work is a thicker 20 mm LYSO crystal also with $50 \times 50$ mm $^{\mathbf {2}}$ size. Both designs make use of an optical layer at the crystal entrance, so-called retroreflector, which bounces back the light to the source of emission. This allows us to preserve the light distribution while transferring a larger number of scintillation photons to the photosensor. Both approaches make use of a SiPM array with $12 \times 12$ elements $( 3 \times 3$ mm $^{\mathbf {2}}$ dimensions) with about 4.2 mm pitch. SiPMs with low dark count rates of about 30 kcps/mm $^{\mathbf {2}}$ at the breakdown voltage have been selected.We have characterized both main detector blocks namely 20 mm and 15 mm thick. We characterized their measured spatial and energy resolution as a function of the photon depth of interaction (DOI). The DOI resolution was also evaluated. Average values for the whole scintillation volume of 1.9 mm spatial resolution (FWHM) was obtained, together with about 13% energy resolution. DOI performance showed FWHM values nearing 4 mm. In the case of the 15 mm block, studies showing the system capability to return accurate results at impact positions as close as 0.3 mm from the crystal edge were also carried out.We have also, within the MINDView collaboration for the brain PET insert, and together with Oncovision team, acquired and reconstructed pilot data of a mini Derenzo-like phantom showing for both cases the possibility to visualize the 1.6 mm in diameter rods.

Wavelength discrimination (WLD) detector optimization for time-of-flight positron emission tomography with depth of interaction information

The NeuroEXPLORER (NX) system has recently been proposed for a high-resolution and high-sensitivity PET scanner dedicated for brain research. A long axial field-of-view (FOV) of 50cm was used for maximizing the sensitivity in brain region. Depth-of-interaction (DOI) modelling of NX system is necessary for generating high-resolution images and compensate for the large angle effects caused by long axial FOV.To minimize the modifications on the existing reconstruction algorithm and to minimize computational cost, a listmode-based DOI compensation method (DOI mashing) was developed. For detected event with non-zero DOI value, a new virtual detector pair was calculated along the incidence angle. The virtual detector pair has DOI value 0 and therefore has minimal DOI effect. Monte Carlo simulations were used to calculate the point-spread-functions (PSF) with the DOI effect with and without mashing. Analytical simulations were carried out with the acquired PSFs to simulate reconstructions with and without DOI corrections.The full-width-half-maximum (FWHM) of the PSF in the axial direction was 2.8mm with 0 incidence angle and was 8mm with the maximum incidence angle. With DOI mashing, the maximum FWHM reduced to 4mm. The residual blurring was found to be caused by inter-crystal scatters. Similar trend was found in the transaxial PSFs, where a large FWHM was observed in off-center locations. Image reconstruction without DOI mashing showed that without DOI mashing, image resolution degrades in off-center locations. With DOI mashing, image resolution was preserved in off-center locations.In conclusion, we proposed a DOI mashing approach that can compensate the DOI effects on the listmode data with minimal modifications of existing reconstruction methods.

We propose a depth-of-interaction (DOI)-encoding method to extract continuous DOI information using a single-layer scintillation crystal array with single-ended readout for cost-effective high-resolution positron emission tomography (PET). DOI information is estimated by different light dispersions along the x- and y-directions tailored by the geometric shape of reflectors around the crystals. The detector module comprised a 22 × 22 array of unpolished LGSO crystals (2.0 × 2.0 × 20 mm3). A multi-anode photomultiplier tube with 64 anodes measured light dispersion in the crystal array. Gain non-uniformity of each anode was corrected by an analogue gain compensation circuit. DOI information was determined from peaks in the x and y anode-signal distributions normalized by the total energy of the distribution. Average DOI resolution (full width at half maximum, FWHM) over all crystals and depths was estimated to be 4.2 mm. Average energy resolution from the 2 to 18 mm DOI positions was 11.3% ± 0.79%, with 13% difference in photo-peak positions. Average time resolutions (FWHM) were 320–356 ps. Energy, time and DOI resolutions were uniform over all crystal positions except at the array's edge. This DOI-PET detector shows promise for applications that require high resolution and sensitivity at low cost.

Absence of scintillation crystal in the region between adjacent PMTs, PMT blind region, will cause loss of data sampling and less sensitivity in a PET or another system. We propose design of a depth of interaction (DOI) detector to fill up the PMT blind region with crystals. In the design, a 3-dimensional (3D) array of single kind crystals stands over two PMTs. The crystal array is optically coupled to the PMTs directly and materials between crystal elements are chosen among optical film, coupling material and just remained as air gap to control scintillation light path in the crystal array instead of coupling a light guide on the bottom of the array as is generally done. Not only reflector but translucent film is also utilized as optical film in the DOI detector design. In this paper, first, we seek for an available DOI crystal array design in simple setting with a PMT, and then we demonstrate its DOI detection capability in the setting with two PMTs. The DOI detector is developed with the intention of the use in jPET-RD system; the PET system we have developed for small animals. In the system, there is about 6 mm interval in axial direction between outer anodes of adjacent two PMTs. The cross section of the crystal elements is 1.46 mm /spl times/ 1.46 mm in jPET-RD detector so that 2-layers of four crystals are involved in the PMT blind region. We introduce the design of the 2-layer DOI crystal array for the region and the results of its performance evaluation.

Depth-of-interaction enhanced Compton camera using pixelated LYSO(Ce) scintillator arrays and dual-ended SiPMs.

Organ-dedicated PET scanners are becoming more prevalent because of their advantages in higher sensitivity, improved image quality, and lower cost. Detectors utilized in these scanners have finer pixel size with depth of interaction (DOI) capability. This work presents a LYSO(Ce) detector module with DOI capability which has the potential to be scaled up to a high-resolution small animal or organ-dedicated PET system. For DOI capability, a submodule with one LYSO block detector utilizing PETsys TOFPET2 application-specific integrated circuit (ASIC) was previously developed in our lab. We scaled up the submodule and optimized the configuration to allow for a compact housing of the dual-readout boards in one side of the blocks by designing a high-speed dual-readout cable to maintain the original pin-to-pin relationship between the Samtec connectors. The module size is 53.8 × 57.8 mm2. Each module has 2 × 2 LYSO blocks, each LYSO block consists of 4 × 4 LYSO units, and each LYSO unit contains a 6 × 6 array of 1 × 1 × 20 mm3 LYSO crystals. The four lateral surfaces of LYSO crystal were mechanically ground to W14, and the two end surfaces were polished. Two ends of the LYSO crystal are optically connected to SiPM for DOI measurement. Eight LYSO blocks performance including energy, timing, and DOI resolution is characterized with a single LYSO slab. The in-panel and orthogonal-panel spatial resolution of the two modules with 107.4 mm distance between each other are measured at 9 positions within the field of view (FOV) with a 22Na source. Results show that the average energy, timing, and DOI resolution of all LYSO blocks are 16.13% ± 1.01% at 511 keV, 658.03 ± 15.18 ps, and 2.62 ± 0.06 mm, respectively. The energy and timing resolution of two modules are 16.35% and 0.86 ns, respectively. The in-panel and orthogonal-panel spatial resolution of the two modules at the FOV center are 1.9 and 4.4 mm respectively.

Imaging performance evaluation in depth-of-interaction PET with a new method of sinogram generation:A Monte Carlo simulation study

We are exploring a scintillator-based PET detector with potential of high sensitivity, depth of interaction (DOI) capability, and timing resolution, with single-side readout. Our design combines two previous concepts: (1) multiple scintillator arrays stacked with relative offset, yielding inherent DOI information, but good timing performance has not been demonstrated with conventional light sharing readout. (2) Single crystal array with one-to-one coupling to the photodetector, showing superior timing performance compared to its light sharing counterparts, but lacks DOI. The combination, where the first layer of a staggered design is coupled one-to-one to a photodetector array, may provide both DOI and timing resolution and this concept is here evaluated through light transport simulations. Results show that: (1) unpolished crystal pixels in the staggered configuration yield better performance across all metrics compared to polished pixels, regardless of readout scheme. (2) One-to-one readout of the first layer allows for accurate DOI extraction using a single threshold. The number of multi pixel photon counter (MPPC) pixels with signal amplitudes exceeding the threshold corresponds to the interaction layer. This approach was not possible with conventional light sharing readout. (3) With a threshold of 2 optical photons, the layered approach with one-to-one coupled first layer improves timing close to the MPPC compared to the conventional one-to-one coupling non-DOI detector, due to effectively reduced crystal thickness. Single detector timing resolution values of 91, 127, 151 and 164 ps were observed per layer in the 4-layer design, to be compared to 148 ps for the single array with one-to-one coupling. (4) For the layered design with light sharing readout, timing improves with increased MPPC pixel size due to higher signal per channel. In conclusion, the combination of straightforward DOI determination, good timing performance, and relatively simple design makes the proposed concept promising for DOI-Time-of-Flight PET detectors.

Depth of interaction (DOI) errors in commonly employed crystal detectors cause severe degradation of off-axis resolution in current positron emission tomography (PET) scanners used for small animal or single organ imaging. We propose a PET camera design that uses low cost, easily fabricated, lead-walled straw (LWS) detectors. Photons converted in Pb generate photoelectrons that escape into the straw interior producing an avalanche whose position can be readily measured with high precision. Straws can be of considerable length (up to 50 cm), but small in diameter (2 mm), and can be packed to form scanners with high sensitivity and large fields of view (FOV). The rigorous three-dimensional position encoding eliminates DOI errors encountered in crystal detectors, and provides submillimeter resolution and sensitivity response that is uniform over the full extent of the FOV. We present preliminary results of the operation of a prototype 200-straw module. We then propose a PET scanner design that consists of two rotating trapezoidal detector panels, each panel containing fourteen 200-straw modules (a total of 5600 straws). The FOV has a diameter of 5 cm and extends axially to 15 cm. This unique geometry affords a long field of view, with near constant axial sensitivity of 4.7%, and uniform resolution of 0.9 times 0.9 times 1.0 mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> maintained over the entire extent of the FOV, as determined in Monte Carlo simulations.

Small-animal PET systems are now striving for sub-millimetre resolution. Current systems based upon PSPMTs and finely pixellated scintillators can be pushed to higher resolution, but at the expense of other performance parameters and a rapidly escalating cost. Moreover, depth of interaction (DOI) information is usually difficult to assess in such systems, even though this information is highly desirable to reduce the parallax error, which is often the dominant error for such high-resolution systems. In this study we propose a high-resolution detector head for a small-animal PET imaging system with intrinsic DOI information. Instead of a pixellated scintillator, our design is based upon the classic Anger camera principle, i.e. the head is constructed of modular layers each consisting of a continuous slab of scintillator, viewed by a new type of compact silicon photodetector. The photodetector is the recently developed silicon photomultiplier (SiPM) that as well as being very compact has many other attractive properties: high gain at low bias voltage, excellent single-photoelectron resolution and fast timing. A detector head of about 4 × 4 cm2 in area is proposed, constructed from three modular layers of the type described above. We perform a simulation study, using the Monte Carlo simulation package Geant4. The simulation results are used to optimize the geometry of the detector head and characterize its performance. Additionally, hit estimation algorithms are studied to determine the interaction position of annihilation photons correctly over the whole detector surface. The resulting detector has a nearly uniform efficiency for 511 keV photons of ∼70% and an intrinsic spatial resolution of less than ∼0.4 mm full width at half maximum (fwhm).

Depth-of-interaction (DOI) detectors are considered to be advantageous to time-of-flight positron emission tomography (TOF-PET) because they can observe timing error arising in the scintillation crystals due to a propagation speed difference between annihilation radiation and scintillation photons. We experimentally measured this timing error in our four-layer DOI detector. The upper layer exhibited a larger timing delay due to the longer path length after conversion from annihilation radiation into scintillation photons that traveled with the zigzag paths at a decreased speed by a factor of the refractive index (n). The maximum timing delay between the first and fourth layers was evaluated as 164 ps when n=1.47. Because this delay can be corrected using three-dimensional crystal address, thick detector can be employed for TOF-PET to achieve a higher sensitivity without degrading the timing resolution. This demonstration was the first step for a DOI-TOF-PET.

The Silicon Photomultiplier for application to high-resolution Positron Emission Tomography