Zheng Gu

MARS: a mouse atlas registration system based on a planar x-ray projector and an optical camera

This paper introduces a mouse atlas registration system (MARS), composed of a stationary top-view x-ray projector and a side-view optical camera, coupled to a mouse atlas registration algorithm. This system uses the x-ray and optical images to guide a fully automatic co-registration of a mouse atlas with each subject, in order to provide anatomical reference for small animal molecular imaging systems such as positron emission tomography (PET). To facilitate the registration, a statistical atlas that accounts for inter-subject anatomical variations was constructed based on 83 organ-labeled mouse micro-computed tomography (CT) images. The statistical shape model and conditional Gaussian model techniques were used to register the atlas with the x-ray image and optical photo. The accuracy of the atlas registration was evaluated by comparing the registered atlas with the organ-labeled micro-CT images of the test subjects. The results showed excellent registration accuracy of the whole-body region, and good accuracy for the brain, liver, heart, lungs and kidneys. In its implementation, the MARS was integrated with a preclinical PET scanner to deliver combined PET/MARS imaging, and to facilitate atlas-assisted analysis of the preclinical PET images.

Design and initial performance of PETbox4, a high sensitivity preclinical imaging tomograph

PETBox4, currently under development at the Crump Institute is a new tomograph dedicated to preclinical imaging of mice. This system presents a significant improvement on sensitivity and spatial resolution compared to the first generation PETBox. We report here on its design and initial performance characteristics.

A DOI Detector With Crystal Scatter Identification Capability for High Sensitivity and High Spatial Resolution PET Imaging

A new phoswich detector is being developed at the Crump Institute, aiming to provide improvements in sensitivity, and spatial resolution for PET. The detector configuration is comprised of two layers of pixelated scintillator crystal arrays, a glass lightguide and a light detector. The annihilation photon entrance (top) layer is a 48×48 array of 1.01 × 1.01 × 7 mm3 LYSO crystals. The bottom layer is a 32 × 32 array of 1.55 × 1.55 × 9 mm3 BGO crystals. A tapered, multiple-element glass lightguide is used to couple the exit end of the BGO crystal array (52 × 52 mm2) to the photosensitive area of the Position Sensitive Photomultiplier Tube (46 × 46 mm2), allowing the creation of flat panel detectors without gaps between the detector modules. Both simulations and measurements were performed to evaluate the characteristics and benefits of the proposed design. The GATE Monte Carlo simulation indicated that the total fraction of the cross layer crystal scatter (CLCS) events in singles detection mode for this detector geometry is 13.2%. The large majority of these CLCS events (10.1% out of 13.2%) deposit most of their energy in a scintillator layer other than the layer of first interaction. Identification of those CLCS events for rejection or correction may lead to improvements in data quality and imaging performance. Physical measurements with the prototype detector showed that the LYSO, BGO and CLCS events were successfully identified using the delayed charge integration (DCI) technique, with more than 95% of the LYSO and BGO crystal elements clearly resolved. The measured peak-to-valley ratios (PVR) in the flood histograms were 3.5 for LYSO and 2.0 for BGO. For LYSO, the energy resolution ranged from 9.7% to 37.0% full width at half maximum (FWHM), with a mean of 13.4 ± 4.8%. For BGO the energy resolution ranged from 16.0% to 33.9% FWHM, with a mean of 18.6 ± 3.2%. In conclusion, these results demonstrate that the proposed detector is feasible and can potentially lead to a high spatial resolution, high sensitivity and DOI PET system.

Optimization of the Energy Window for PETbox4, a Preclinical PET Tomograph With a Small Inner Diameter

Small animal positron emission tomography (PET) systems are often designed by employing close geometry configurations. Due to the different characteristics caused by geometrical factors, these tomographs require data acquisition protocols that differ from those optimized for conventional large diameter ring systems. In this work we optimized the energy window for data acquisitions with PETbox4, a 50 mm detector separation (box-like geometry) pre-clinical PET scanner, using the Geant4 Application for Tomographic Emission (GATE). The fractions of different types of events were estimated using a voxelized phantom including a mouse as well as its supporting chamber, mimicking a realistic mouse imaging environment. Separate code was developed to extract additional information about the gamma interactions for more accurate event type classification. Three types of detector backscatter events were identified in addition to the trues, phantom scatters and randoms. The energy window was optimized based on the noise equivalent count rate (NECR) and scatter fraction (SF) with lower-level discriminators (LLD) corresponding to energies from 150 keV to 450 keV. The results were validated based on the calculated image uniformity, spillover ratio (SOR) and recovery coefficient (RC) from physical measurements using the National Electrical Manufacturers Association (NEMA) NU-4 image quality phantom. These results indicate that when PETbox4 is operated with a more narrow energy window (350-650 keV), detector backscatter rejection is unnecessary. For the NEMA NU-4 image quality phantom, the SOR for the water chamber decreases by about 45% from 15.1% to 8.3%, and the SOR for the air chamber decreases by 31% from 12.0% to 8.3% at the LLDs of 150 and 350 keV, without obvious change in uniformity, further supporting the simulation based optimization. The optimization described in this work is not limited to PETbox4, but also applicable or helpful to other small inner diameter geometry scanners.

Evaluation of the detection limit at low activity levels for three preclinical PET systems

Tomographic systems dedicated to noninvasive imaging of preclinical animal models have become widely available in recent years. Several imaging tasks involve very low levels of injected activity. One of the key advantages of molecular imaging with PET is its capability to detect and measure non-invasively nano-molar to picomolar concentrations of probes in vivo. It is useful therefore to establish the Minimum Detectable Activity (MDA), which is directly related to detectability of PET systems for very low activity distributions. In this work, the detection limits are evaluated for three commercially available imaging systems through ROI analysis in acquired images from well plate experiments. The detection limits at different reconstructed image frames for certain wells were fitted by an exponential. The MDA was calculated by interpolation between the frames above and below the detection limit. To validate the feasibility of using the ROI method to evaluate the detection limit, the theoretical values of Contrast to Noise Ratio (CNR) between two of the systems were compared to the ratio of the MDA values from the image based ROI method. Close agreement between the two methods strengthens the rationale for the methodologies used in this work.

A New Pulse Pileup Rejection Method Based on Position Shift Identification

Pulse pileup events degrade the signal-to-noise ratio (SNR) of nuclear medicine data. When such events occur in multiplexed detectors, they cause spatial misposition, energy spectrum distortion and degraded timing resolution, which leads to image artifacts. Pulse pileup is pronounced in PETbox4, a bench top PET scanner dedicated to high sensitivity and high resolution imaging of mice. In that system, the combination of high absolute sensitivity, long scintillator decay time (BGO) and highly multiplexed electronics lead to a significant fraction of pulse pileup, reached at lower total activity than for comparable instruments. In this manuscript, a new pulse pileup rejection method named position shift rejection (PSR) is introduced. The performance of PSR is compared with a conventional leading edge rejection (LER) method and with no pileup rejection implemented (NoPR). A comprehensive digital pulse library was developed for objective evaluation and optimization of the PSR and LER, in which pulse waveforms were directly recorded from real measurements exactly representing the signals to be processed. Physical measurements including singles event acquisition, peak system sensitivity and NEMA NU-4 image quality phantom were also performed in the PETbox4 system to validate and compare the different pulse pile-up rejection methods. The evaluation of both physical measurements and model pulse trains demonstrated that the new PSR performs more accurate pileup event identification and avoids erroneous rejection of valid events. For the PETbox4 system, this improvement leads to a significant recovery of sensitivity at low count rates, amounting to about 1/4th of the expected true coincidence events, compared to the LER method. Furthermore, with the implementation of PSR, optimal image quality can be achieved near the peak noise equivalent count rate (NECR).

A Digital Pulse Library for the Optimization of Signal Processing in PET

In this work, we developed a comprehensive digital pulse library as a reference data set, against which the performance of pulse processing algorithms can be evaluated. The pulses are directly recorded from real measurements with scintillation detectors multiplexed by a resistor divider network. This way, the pulses better represent systematic variations of detection sensitivity for detector panels due to differences in light sharing, light collection and crystal scatter among other effects. The pulse shapes exactly represent the signals to be processed in the data acquisition and signal processing unit, but for which the ground truth regarding energy, position and timing can be known. A conventional pulse pileup rejection method called leading edge rejection (LER) was evaluated with the help of the pulse library. The simulated pulse library based results show that the LER method can effectively suppress most of the pileup caused mispositioned events, while it also trades off against sensitivity loss. Physical measurements agree with the pulse library simulation results. In conclusion, the proposed evaluation method based on digital pulse library can significantly reduce the time and effort invested in the development and optimization of various signal processing algorithms.

Matched filter for event identification and processing in PET

Positron Emission Tomography (PET) is a technology that can image the spatial and temporal distribution of radio-labeled biomolecules in-vivo. It is dependent on the coincidence detection of photons emitted due to the annihilation of positrons. Most PET scanners use constant fraction discriminators (CFD) to confirm detector triggering. Following this, a blocking time window is used to allow processing of individual events. However, if a second (or more) event interacts with the detectors within this blocking time window, this method will not be capable to accurately identify its timing or energy, while the energy information for the first event is lost. Both energy and timing of all events are necessary for accurate event detection and localization. In these cases and without a method to identify individual events collected within the blocking time window, all events within the blocking window must be rejected, resulting in a real loss in sensitivity. The information loss in these situations is a significant problem when using high activity sources due to the high probability of event pile-up. To address this issue, we examined the performance of a matched filter when applied to raw pulses obtained from a pulse library derived from the PETBox system. Here, we compare the performance of the matched filter to CFD for event detection, positioning and timing. In addition, we show preliminary results from a real-time digital implementation of the matched filter on the PETBox FPGA. Preliminary results indicate that the matched filter can significantly improve the fidelity of the detection of true events at high count rates, as compared to conventional CFD, while event timing is somewhat compromised for the matched filter.

Performance evaluation of G8, a high sensitivity benchtop preclinical PET/CT tomograph

G8 is a benchtop integrated PET/CT scanner dedicated to high-sensitivity and high-resolution imaging of mice. This work characterizes its National Electrical Manufacturers Association NU 4-2008 performance where applicable and also assesses the basic imaging performance of the CT subsystem. Methods: The PET subsystem in G8 consists of 4 flat-panel detectors arranged in a boxlike geometry. Each panel consists of 2 modules of a 26 × 26 pixelated bismuth germanate scintillator array with individual crystals measuring 1.75 × 1.75 × 7.2 mm. The crystal arrays are coupled to multichannel photomultiplier tubes via a tapered, pixelated glass lightguide. A cone-beam CT scanner consisting of a MicroFocus x-ray source and a complementary metal oxide semiconductor detector provides anatomic information. Sensitivity, spatial resolution, energy resolution, scatter fraction, count-rate performance, and the capability of performing phantom and mouse imaging were evaluated for the PET subsystem. Noise, dose level, contrast, and resolution were evaluated for the CT subsystem. Results: With an energy window of 350–650 keV, the peak sensitivity was 9.0% near the center of the field of view. The crystal energy resolution ranged from 15.0% to 69.6% in full width at half maximum (FWHM), with a mean of 19.3% ± 3.7%. The average intrinsic spatial resolution was 1.30 and 1.38 mm FWHM in the transverse and axial directions, respectively. The maximum-likelihood expectation maximization reconstructed image of a point source in air averaged 0.81 ± 0.11 mm FWHM. The peak noise-equivalent count rate for the mouse-sized phantom was 44 kcps for a total activity of 2.9 MBq (78 μCi), and the scatter fraction was 11%. For the CT subsystem, the value of the modulation transfer function at 10% was 2.05 cycles/mm. Conclusion: The overall performance demonstrates that the G8 can produce high-quality images for molecular imaging–based biomedical research.

Lightguides for improving edge crystal identification and energy resolution in pixelated scintillator detectors

A phoswich depth of interaction (DOI) detector composed by two layers of scintillator arrays (LYSO and BGO) has been proposed. To use this design for building practical systems, a tapered multiple-element glass lightguide is necessary to couple the scintillator arrays to a PSPMT. The complete individual detector module offers an overall dimension of 52×52 mm2 that matches the external dimensions of the PSPMT package, allowing continuous positioning of the scintillator arrays for creating flat panel detectors without introducing gaps between detector modules. In this work, we define the elements adjacent to the edge in an array as “edge+” elements. Both of the layers of the phoswich detector and the lightguide have such edge+ elements. The lightguide was modified according to the photodetector geometry, aiming at improving edge and corner crystal identification and energy resolution: In an assembled detector, the optimized edge+ lightguide elements superimpose on top of both edge and edge+ anodes of the PSPMT. In that case, the edge+ lightguide elements distribute light onto both the edge and edge+ anodes, while edge lightguide elements primarily distribute light onto the edge anodes. This different light sharing scheme improves the capability of identifying edge+ crystals from edge crystals. In addition, the edge lightguide elements were enlarged to improve the light collection and corresponding energy resolution for the edge crystals. Flood images were acquired and energy resolution was calculated for individual crystals. All the edge and edge+ crystals were clearly resolved. Energy resolution was improved from 24.4 ± 4.4% to 20.0 ± 1.7% for edge crystals. The new proposed lightguide successfully improved edge crystal identification and energy resolution in the proposed pixelated scintillator detector.