Search

Machine learning has found unique applications in nuclear medicine from photon detection to quantitative image reconstruction. Although there have been impressive strides in detector development for time-of-flight positron emission tomography (PET), most detectors still make use of simple signal processing methods to extract the time and position information from the detector signals. Now, with the availability of fast waveform digitizers, machine learning techniques have been applied to estimate the position and arrival time of high-energy photons. In quantitative image reconstruction, machine learning has been used to estimate various corrections factors, including scattered events and attenuation images, as well as to reduce statistical noise in reconstructed images. Here, machine learning either provides a faster alternative to an existing time-consuming computation, such as in the case of scatter estimation, or creates a data-driven approach to map an implicitly defined function, such as in the case of estimating the attenuation map for PET/MR scans. In this article, we will review the above-mentioned applications of machine learning in nuclear medicine. [ABSTRACT FROM AUTHOR]

In a scintillation detector, the light generated in the scintillator by a gamma interaction is converted to photoelectrons by a photodetector and produces a time-dependent waveform, the shape of which depends on the scintillator properties and the photodetector response. Several depth-of-interaction (DOI) encoding strategies have been developed that manipulate the scintillator’s temporal response along the crystal length and therefore require pulse shape discrimination techniques to differentiate waveform shapes. In this work, we demonstrate how maximum likelihood (ML) estimation methods can be applied to pulse shape discrimination to better estimate deposited energy, DOI and interaction time (for time-of-flight (TOF) PET) of a gamma ray in a scintillation detector. We developed likelihood models based on either the estimated detection times of individual photoelectrons or the number of photoelectrons in discrete time bins, and applied to two phosphor-coated crystals (LFS and LYSO) used in a previously developed TOF-DOI detector concept. Compared with conventional analytical methods, ML pulse shape discrimination improved DOI encoding by 27% for both crystals. Using the ML DOI estimate, we were able to counter depth-dependent changes in light collection inherent to long scintillator crystals and recover the energy resolution measured with fixed depth irradiation (~11.5% for both crystals). Lastly, we demonstrated how the Richardson-Lucy algorithm, an iterative, ML-based deconvolution technique, can be applied to the digitized waveforms to deconvolve the photodetector’s single photoelectron response and produce waveforms with a faster rising edge. After deconvolution and applying DOI and time-walk corrections, we demonstrated a 13% improvement in coincidence timing resolution (from 290 to 254 ps) with the LFS crystal and an 8% improvement (323 to 297 ps) with the LYSO crystal. [ABSTRACT FROM PUBLISHER]

The performance of an 8 \times 8 array of 6.0 \times 6.0~\mm^2 (active area) SiPMs was evaluated for PET applications using crystal arrays with different pitch sizes (3.4, 1.5, 1.35, 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 \times 57.4~\mm^2, 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 to 30.0 V in 0.5 V intervals) and two different temperatures (5^\circ C and 20^\circC). Results show that the best flood histogram was obtained at a bias voltage of 28.0 V and 5^\circC 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^\circC ranged from 12.7 \pm 0.8\% to 14.6 \pm 1.4\% and 1.58 \pm 0.13~\ns to 2.50 \pm 0.44~\ns, for crystal array pitch sizes of 3.4 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. [ABSTRACT FROM AUTHOR]

We are developing detectors for an un-collimated scanner for very high sensitivity single-photon imaging. The scanner consists of two large, thin, closely spaced, pixelated scintillation detectors of either thallium-doped sodium iodide (NaI(Tl)), sodium-doped cesium iodide (CsI(Na)), or bismuth germanate (BGO). The scintillator arrays are interchangeable to optimize performance for a wide range of isotope energies and imaging subjects. For each scintillator, flood histograms were obtained utilizing test gamma sources: ^109Cd, ^241Am, ^57Co, ^133Ba, ^137Cs, ^22Na, and beta source ^32Si. Using a 3 mm thick 21 \times 41 NaI(Tl) pixelated array (2~\mm \times 2~\mm pixels on a 2.2 mm pitch), read out by two adjacent Hamamatasu H8500 multi-channel photomultiplier tubes (MCPMTs), the energy resolution for ^57Co was 12.2% over the center of the MCPMTs and 45.3% for pixels over the gap, with flood histogram peak-to-valley ratios > \!\!\!\!6.1 for the center of the MCPMTs and > \!\!\!\!4.2 for the gap between the two MCPMTs. The NaI(Tl) detector consistently produced peak-to-valley ratios of >\!\!\!\! 1.6 in the center of the MCPMTs for all radioisotopes studied. CsI(Na) and BGO detectors had flood histograms of similar quality to NaI(Tl) at intermediate to high-energies, allowing for optimal imaging across a wide energy range. The measured sensitivity for point sources placed 10 mm from the detector approached 35% for low-energy isotopes and all scintillator materials, which results in an expected system sensitivity (two opposing detectors) of \sim \!\!70\% for these isotopes. The 5 mm thick BGO arrays had the greatest efficiency for high-energy isotopes due to the increased stopping power of this material. The measured spatial resolution for ^99m Tc for the CsI(Na) detector (1.6 mm pitch) was 7.5 mm FWHM at the aluminum entrance window (1.5 mm from scintillator face), decreasing to 15 mm at a distance of 5 mm from the scintillator face. The NaI(Tl) detector showed similar spatial resolution. This spatial resolution is sufficient for imaging well-localized radiotracer distributions in thin objects, and the very high sensitivity can be used for fast dynamic imaging, or imaging using very low injected doses. [ABSTRACT FROM AUTHOR]

We present an analysis of the signal properties of a position-sensitive solid-state photomultiplier (PS-SSPM) that has an integrated resistive network for position sensing. Attractive features of PS-SSPMs are their large area and ability to resolve small scintillator crystals. However, the large area leads to a high detector capacitance, and in order to achieve high spatial resolution a large network resistor value is required. These inevitably create a low-pass filter that drastically slows what would be a fast micro-cell discharge pulse. Significant changes in the signal shape of the PS-SSPM cathode output as a function of position are observed, which result in a position-dependent time delay when using traditional time pick-off methods such as leading edge discrimination and constant fraction discrimination. The timing resolution and time delay, as a function of position, were characterized for two different PS-SSPM designs, a continuous 10~\mm \times 10~\mm PS-SSPM and a tiled 2 \times 2 array of 5~\mm \times 5~\mm PS-SSPMs. After time delay correction, the block timing resolution, measured with a 6 \times 6 array of 1.3 \times 1.3 \times 20~\mm^3 LSO crystals, was 8.6 ns and 8.5 ns, with the 10 mm PS-SSPM and 5 mm PS-SSPM respectively. The effect of crystal size on timing resolution was also studied, and contrary to expectation, a small improvement was measured when reducing the crystal size from 1.3 mm to 0.5 mm. Digital timing methods were studied and showed great promise for allowing accurate timing by implementation of a leading edge time pick-off. Position-dependent changes in signal shape on the anode side also are present, which complicates peak height data acquisition methods used for positioning. We studied the effect of trigger position on signal amplitude, flood histogram quality, and depth-of-interaction resolution in a dual-ended readout detector configuration. We conclude that detector timing and positioning can be significantly improved by implementation of digital timing methods and by accounting for changes in the shape of the signals from PS-SSPMs. [ABSTRACT FROM AUTHOR]

Driven by a need for high and uniform spatial resolution, as well as depth correction to improve time-of-flight PET, the design of PET detectors with depth-of-interaction (DOI) information is of great interest. To combine the advantages of continuous DOI information with a single-ended readout of the crystal, we have proposed a solution based on the use of phosphor-coated crystals. Here we build on our initial work to optimize the coating process and detector design, thus taking a significant step towards making this approach practical. We implemented a new method to coat 20 mm-long unpolished LSO crystals with a mixture of YAG phosphor powder and UV-curable optical cement that provides accurate, robust, and reproducible coatings. Using this new coating process, critical parameters such as coating thickness, coating shape, and crystal size were studied. The performance of different coating configurations was evaluated by recording pulses at different irradiation depths and extracting the decay time of individual pulses. The delayed charge integration method was used to characterize the pulses acquired at various DOI. The coating thickness was decreased from 250~\mum (previous work) to 50~\mu m with minimal degradation in performance. 100~\mu m-thick coating provided the best trade-off between performance and additional volume around the crystal, which will be important to build crystal arrays with high packing fraction. Additionally, varying the coating length on the different crystal faces allowed for better DOI encoding along the crystal length, when compared to previous coating configurations. Crystal face sizes ranging from 2 mm \times 2 mm to 5 mm \times 5 mm were studied; best performance was obtained with the 2 mm \times 2 mm crystals. [ABSTRACT FROM PUBLISHER]

A capacitive charge-division readout method for reading out a 2\,\times\,2 array of 5 mm\,\times\,5 mm position-sensitive solid-state photomultipliers (PS-SSPM) was designed and evaluated. Using this analog multiplexing method, the 20 signals (16 position, 4 timing) from the PS-SSPM array are reduced to 5 signals (4 position, 1 timing), allowing the PS-SSPM array to be treated as an individual large-area PS-SSPM module. A global positioning approach can now be used, instead of individual positioning for each PS-SSPM in the array, ensuring that the entire light signal is utilized. The signal-to-noise ratio (SNR) and flood histogram quality at different bias voltages (27.5 V to 32.0 V at 0.5 V intervals) and a fixed temperature of 0^\circC were evaluated by coupling a 6\,\times\,6 array of 1.3 mm\,\times\,1.3 mm\,\times\,20 mm polished LSO crystals to the center of the PS-SSPM array. The timing resolution was measured at a fixed bias voltage of 31.0 V and a fixed temperature of 0^\circC. All the measurements were evaluated and compared using capacitors with different values and tolerances. Capacitor values ranged from 0.051 nf to 10 nf, and the capacitance tolerance ranged from 1% to 20%. The results show that better performance was achieved using capacitors with smaller values and better capacitance tolerance. Using 0.2 nf capacitors, the SNR, energy resolution and timing resolution were 24.3, 18.2% and 8.8 ns at a bias voltage 31.0 V, respectively. The flood histogram quality was also evaluated by using a 10\,\times\,10 array of 1 mm\,\times\,1 mm\,\times\,10 mm polished LSO crystals and a 10\,\times\,10 array of 0.7 mm\,\times\,0.7 mm\,\times\,20 mm unpolished LSO crystals to determine the smallest crystal size resolvable. These studies showed that the high spatial resolution of the PS-SSPM was preserved allowing for 0.7 mm crystals to be identified. These results show that the capacitive charge-division analog signal processing method can significantly reduce the number of electronic channels, from 20 to 5, while retaining the excellent performance of the detector. [ABSTRACT FROM AUTHOR]

A new approach to depth of interaction (DOI) encoding for a pixelated LSO array using a single multi-channel PMT was investigated. In this method the DOI information was estimated by taking advantage of optical crosstalk between LSO elements and examining the standard deviation (spread) of signals on all channels of the PMT. Unpolished and polished 6 × 6 LSO arrays with a crystal size of 1.3 × 1.3 × 20 mm3 were evaluated on a Hamamatsu H7546 64-channel PMT. The arrays were placed on the center of the PMT and the central 16 channels of the PMT were individually read out and digitized. For the unpolished array, all crystals were resolved in the flood histogram. An average DOI resolution of 8 mm was obtained. The energy resolution was 25% after the signal amplitude was corrected using the measured DOI information. For the polished array, the flood histogram was superior to the unpolished array, however no DOI information could be measured. Using unpolished crystals, this method could be a practical way to achieve limited DOI information in PET detectors. The standard deviation of all PMT channels can be readily obtained using a resistor network. Only five signals (four signals to determine the x-y position and one signal measuring the standard deviation) need to be digitized, and this method only requires a single photon detector to read out the array. Unlike phoswich detectors, the method does not require segmenting the scintillator array into layers. The measured DOI resolution was much worse than that obtained with the dual-ended readout method, however, it was similar to that obtained with a two-layer phoswich detector. [ABSTRACT FROM AUTHOR]

A magnetic resonance (MR) compatible positron emission tomography (PET) insert has been developed in our laboratory for simultaneous small animal PET/MR imaging. This system is based on lutetium oxyorthosilicate (LSO) scintillator arrays with position-sensitive avalanche photodiode (PSAPD) photodetectors. The PET performance of this insert has been measured. The average reconstructed image spatial resolution was 1.51 mm. The sensitivity at the center of the field of view (CFOV) was 0.35%, which is comparable to the simulation predictions of 0.40%. The average photopeak energy resolution was 25%. The scatter fraction inside the MRI scanner with a line source was 12% (with a mouse-sized phantom and standard 35 mm Bruker 1H RF coil), 7% (with RF coil only) and 5% (without phantom or RF coil) for an energy window of 350-650 keV. The front-end electronics had a dead time of 390 ns, and a trigger extension dead time of 7.32 μs that degraded counting rate performance for injected doses above ~0.75 mCi (28 MBq). The peak noise-equivalent count rate (NECR) of 1.27 kcps was achieved at 290 μCi (10.7 MBq). The system showed good imaging performance inside a 7-T animal MRI system; however improvements in data acquisition electronics and reduction of the coincidence timing window are needed to realize improved NECR performance. [ABSTRACT FROM AUTHOR]

Position-sensitive avalanche photodiodes (PSAPDs) are gaining widespread acceptance in modern PET scanner designs, and owing to their relative insensitivity to magnetic fields, especially in those that are MRI-compatible. Flood histograms in PET scanners are used to determine the crystal of annihilation photon interaction and hence, for detector characterization and routine quality control. For PET detectors that use PSAPDs, flood histograms show a characteristic pincushion distortion when Anger logic is used for event positioning. A small rotation in the flood histogram is also observed when the detectors are placed in a magnetic field. We first present a general purpose automatic method for spatial distortion correction for flood histograms of PSAPD-based PET detectors when placed both inside and outside a MRI scanner. Analytical formulas derived for this scheme are based on a hybrid approach that combines desirable properties from two existing event positioning schemes. The rotation of the flood histogram due to the magnetic field is determined iteratively and is accounted for in the scheme. We then provide implementation details of a method for crystal identification we have previously proposed and evaluate it for cases when the PET detectors are both outside and in a magnetic field. In this scheme, Fourier analysis is used to generate a lower-order spatial approximation of the distortion-corrected PSAPD flood histogram, which we call the "template". The template is then registered to the flood histogram using a diffeomorphic iterative intensity-based warping scheme. The calculated deformation field is then applied to the segmentation of the template to obtain a segmentation of the flood histogram. A manual correction tool is also developed for exceptional cases. We present a quantitative assessment of the proposed distortion correction scheme and crystal identification method against conventional methods. Our results indicate that our proposed methods lead to a large reduction in manual labor and indeed can routinely be used for calibration and characterization studies in MRI-compatible PET scanners based on PSAPDs. [ABSTRACT FROM AUTHOR]

Excellent spatial resolution is a requirement for preclinical PET imaging. In order to achieve spatial resolution of significantly better than one millimeter, an appealing possibility is to employ direct detector materials, such as cadmium telluride (CdTe). Prototype thin orthogonal strip detectors have been developed for testing. They have dimensions of 20 mm by 20 mm and are 0.5 mm thick, and have strips of 0.5 mm pitch on one side and 2.5 mm pitch on the other. Results are presented for the energy resolution (3% at 511 keV), intrinsic position resolution (equal to the 0.5 mm strip pitch), and timing resolution (3 ns FWHM in coincidence with an LSO detector, 8 ns FWHM for coincidence of two CdTe detectors) of the detectors. A PET scanner design is proposed using blocks made of the CdTe strip detectors, oriented in the blocks with their thin edges toward the center of the scanner. Simulation results suggest that this scanner, using a threshold of 250 keV, would have a sensitivity of 3.4% for a point source at its center. [ABSTRACT FROM PUBLISHER]

We explore two depth of interaction (DOI) capable detectors with a high effective packing fraction to be used for the PET portion of a next generation cylindrical pendant breast PET/CT scanner. The proposed dual-readout detectors consist of an unpolished (14 14) array of LSO crystals coupled to a position sensitive-PMT (PSPMT) at the back end for position sensing and either (A) a (2 2) matrix of avalanche photodiodes (APDs) or (B) a single large APD at the front end for measuring DOI. Increased effective packing fraction is achieved by choosing the APDs such that the total front surface contact area of the APD matrix or the single large APD approximately matches both the contact area of the crystal array and the active area of the PSPMT. This results in reduced dead space between adjacent detectors when placed in a cylindrical scanner geometry. We propose to recover DOI for the crystals that are partially occluded due to the inactive area of the APD by using the optical crosstalk from neighboring crystals. We conducted preliminary studies utilizing five prototype modules: two of which used a (9 9) array of LSO crystals, and three that used a (14 14) LSO array of crystals. The readout was performed either by a , or a position sensitive-APD (PSAPD) at the front end and a PSPMT at the rear end. At room temperature, the average energy and DOI resolutions degraded only by 2% and 0.3 mm respectively for edge crystals compared to the center crystals for all PSAPDs. The energy resolution and the DOI resolution for the detector design with readout (A) was estimated to be 14.2% and 4.0 mm, and that for the design with readout (B) was estimated at 14.7% and 4.4 mm. After cooling to 10, the performance of the second design was close to that of the first design. We conclude that a dedicated cylindrical breast PET/CT scanner could be operated with either of the proposed DOI detector schemes, thus improving packing fraction, and hence, sensitivity and NEC. The two designs have dif- - ferent advantages and drawbacks: the first design performs well at room temperature, however, the readout is complex, while the second design has a simpler readout scheme but requires a cooling system for it to achieve optimal performance. [ABSTRACT FROM PUBLISHER]

A number of laboratories and companies are currently exploring the development of integrated imaging systems for magnetic resonance imaging (MRI) and positron emission tomography (PET). Scanners for both preclinical and human research applications are being pursued. In contrast to the widely distributed and now quite mature PET/computed tomography technology, most PET/MRI designs allow for simultaneous rather than sequential acquisition of PET and MRI data. While this offers the possibility of novel imaging strategies, it also creates considerable challenges for acquiring artifact-free images from both modalities. This paper discusses the motivation for developing combined PET/MRI technology, outlines the obstacles in realizing such an integrated instrument, and presents recent progress in the development of both the instrumentation and of novel imaging agents for combined PET/MRI studies. The performance of the first-generation PET/MRI systems is described. Finally, a range of possible biomedical applications for PET/MRI are outlined. The Integration of Positron Emission Tomography With Magnetic Resonance Imaging. [ABSTRACT FROM AUTHOR]

The goal of this study was to evaluate the magnitude and origin of scattered radiation in a small-animal PET scanner and to assess the impact of these findings on noise equivalent count rate (NECR) measurements, a metric often used to optimize scanner acquisition parameters and to compare one scanner with another. The scatter fraction (SF) was measured for line sources in air and line sources placed within a mouse-sized phantom (25 mm ϕ × 70 mm) and a rat-sized phantom (60 mm ϕ × 150 mm) on the microPET II small-animal PET scanner. Measurements were performed for lower energy thresholds ranging from 150-450 keV and a fixed upper energy threshold of 750 keV. Four different methods were compared for estimating the SF. Significant scatter fractions were measured with just the line source in the field of view, with the spatial distribution of these events consistent with scatter from the gantry and room environment. For mouse imaging, this component dominates over object scatter, and the measured SF is strongly method dependent. The environmental SF rapidly increases as the lower energy threshold decreases and can be more than 30% for an open energy window of 150-750 keV. The object SF originating from the mouse phantom is about 3-4% and does not change significantly as the lower energy threshold in- creases. The object SF for the rat phantom ranges from 10 to 35% for different energy windows and increases as the lower energy threshold decreases. Because the measured SF is highly dependent on the method, and there is as yet no agreed upon standard for animal PET, care must be exercised when comparing NECR for small objects between different scanners. Differences may be methodological rather than reflecting any relevant difference in the performance of the scanner. Furthermore, these results have implications for scatter correction methods when the majority of the detected scatter does not arise from the object itself. [ABSTRACT FROM AUTHOR]

We report on the development of a new high efficiency detector for small animal PET. The detector is based on a monolithic block of LSO pixelated using laser ablation technique. The laser processing allows pixelation with very narrow, 70 μm wide, inter-pixel gaps resulting in a substantially enhanced sensitivity when the detectors are operated in coincidence mode. This paper presents the first results of a detector module fabricated using this approach. Preliminary imaging data at 511 keV obtained by coupling the pixelated LSO to a position sensitive photomultiplier tube (PSPMT) and a position sensitive avalanche photodiode (PSAPD) are presented. [ABSTRACT FROM AUTHOR]

Discusses a study which designed a small animal magnetic resonance (MR) compatible photon emission tomography (PET) (McPET) scanner for simultaneous PET and MR imaging of mice. Surface treatment of lutetium oxyorthosilicate (LSO) crystals; Result of the crystal coupling experiments; Significant difference between connectors made of different materials; Description of the proposed scanner.

Deals with the results of a depth of interaction (DOI) measurement of bending optical fibers and the light loss due to fiber bending. Amount of light loss due to fiber bending; Overview of DOI measurements; Discussion on the attenuation and scatter measurements.

Describes a three dimensional (3D) method, used for high resolution reconstruction of images for the Siemens/CTI ECAT EXACT HR+ whole body positron emission tomography (PET) scanner. How Bayesian reconstructions are computed; Exploitation of the sinogram symmetries using a factored system matrix and automoted indexing scheme; Performance of phantom studies to compare 3D MAP algorithm with the 3D OSEM and 3DRP methods.