In this paper, the authors review the field of resolution modeling in positron emission tomography (PET) image reconstruction, also referred to as point-spread-function modeling. The review includes theoretical analysis of the resolution modeling framework as well as an overview of various approaches in the literature. It also discusses potential advantages gained via this approach, as discussed with reference to various metrics and tasks, including lesion detection observer studies. Furthermore, attention is paid to issues arising from this approach including the pervasive problem of edge artifacts, as well as explanation and potential remedies for this phenomenon. Furthermore, the authors emphasize limitations encountered in the context of quantitative PET imaging, wherein increased intervoxel correlations due to resolution modeling can lead to significant loss of precision (reproducibility) for small regions of interest, which can be a considerable pitfall depending on the task of interest.

Resolution modeling in PET imaging: Theory, practice, benefits, and pitfalls

We evaluated the performance characteristics of the eXplore VISTA dual-ring small-animal PET scanner, a stationary, ring-type, depth-of-interaction (DOI) correcting system designed to simultaneously maximize sensitivity, resolution, and resolution uniformity over a field of view sufficient to image rodent-sized animals. Methods: We measured the intrinsic spatial resolution response of the VISTA detector modules, spatial and volume resolution throughout a representative portion of the field of view, and imaged several common resolution phantoms to provide a qualitative picture of resolution performance. We obtained an axial sensitivity profile and measured central point source sensitivity, scatter fractions and noise equivalent count (NEC) rates for rat- and mouse-sized objects using different energy windows, and count rate linearity. In addition, we measured the energy and timing resolution of both of the crystal layers (cerium-doped gadolinium orthosilicate and cerium-doped lutetium–yttrium orthosilicate) that give VISTA machines a DOI compensation capability. We examined the effectiveness of this DOI compensation by comparing spatial resolution measurements with and without the DOI correction enabled. Finally, several animal studies were included to illustrate system performance in the field. Results: Spatial and volume resolutions averaged approximately 1.4 mm and 2.9 mm3, respectively (with 3-dimensional Fourier rebinning and 2-dimensional filtered backprojection image reconstructions and an energy window of 250–700 keV), along the central axis of the scanner, and the spatial resolution was better than 1.7 mm and 2.1 mm at 1 and 2 cm off the central axis, respectively. Central point source sensitivity measured approximately 4% with peak NEC rates of 126.8 kcps at 455 kBq/mL and 77.1 kcps at 141 kBq/mL for mouse- and rat-sized uniform, cylindric phantoms, respectively. The radial spatial resolution at 2.8 cm off axis with DOI compensation was 2.5 mm but degraded (by 56%) to 3.9 mm without DOI compensation (as would be the case with a geometrically identical scanner without DOI correction capability). Conclusion: These results indicate that the VISTA small-animal PET scanner is well suited to imaging rodent-sized animals. The combination of high spatial resolution, resolution uniformity, sensitivity, and count rate performance, made possible in part by the novel use of phoswich detector modules, confers significant technical advantages over machines with similar geometry but without DOI correction capability.

https://jnm.snmjournals.org/content/jnumed/47/11/1891.full.pdf

Performance Evaluation of the GE Healthcare eXplore VISTA Dual-Ring Small-Animal PET Scanner

This review concentrates on the latest advances in molecular imaging technology, including PET, MRI, and optical imaging. In PET, significant improvements in tumor detection and image resolution have been achieved by introducing new scintillation materials, iterative image reconstruction, and correction methods. These advances enabled the first clinical scanners capable of time-of-flight detection and incorporating point-spread-function reconstruction to compensate for depth-of-interaction effects. In the field of MRI, the most important developments in recent years have mainly been MRI systems with higher field strengths and improved radiofrequency coil technology. Hyperpolarized imaging, functional MRI, and MR spectroscopy provide molecular information in vivo. A special focus of this review article is multimodality imaging and, in particular, the emerging field of combined PET/MRI.

https://jnm.snmjournals.org/content/jnumed/49/Suppl_2/5S.full.pdf

Latest Advances in Molecular Imaging Instrumentation

Accurate system modeling in tomographic image reconstruction has been shown to reduce the spatial variance of resolution and improve quantitative accuracy. System modeling can be improved through analytic calculations, Monte Carlo simulations, and physical measurements. The purpose of this work is to improve clinical fully-3-D reconstruction without substantially increasing computation time. We present a practical method for measuring the detector blurring component of a whole-body positron emission tomography (PET) system to form an approximate system model for use with fully-3-D reconstruction. We employ Monte Carlo simulations to show that a non-collimated point source is acceptable for modeling the radial blurring present in a PET tomograph and we justify the use of a Na22 point source for collecting these measurements. We measure the system response on a whole-body scanner, simplify it to a 2-D function, and incorporate a parameterized version of this response into a modified fully-3-D OSEM algorithm. Empirical testing of the signal versus noise benefits reveal roughly a 15% improvement in spatial resolution and 10% improvement in contrast at matched image noise levels. Convergence analysis demonstrates improved resolution and contrast versus noise properties can be achieved with the proposed method with similar computation time as the conventional approach. Comparison of the measured spatially variant and invariant reconstruction revealed similar performance with conventional image metrics. Edge artifacts, which are a common artifact of resolution-modeled reconstruction methods, were less apparent in the spatially variant method than in the invariant method. With the proposed and other resolution-modeled reconstruction methods, edge artifacts need to be studied in more detail to determine the optimal tradeoff of resolution/contrast enhancement and edge fidelity.

/pdf/application-and-evaluation-of-a-measured-spatially-variant-4fhkvoye03.pdf

Application and Evaluation of a Measured Spatially Variant System Model for PET Image Reconstruction

We present design studies of a multi-layer PET detector module that uses an 8/spl times/8 array of 3 mm square PIN photodiodes to both identify the crystal of interaction and measure the depth of interaction. Each photodiode is coupled to one end of a 3/spl times/3/spl times/30 mm BGO crystal, with the opposite ends of 64 such crystals attached to a single 1" square photomultiplier tube that provides a timing signal and energy discrimination. Each BGO crystal is coated with a lossy reflector, so the ratio of light detected in the photodiode and photomultiplier tube depends on the interaction depth in the crystal, and is used to determine this depth of interaction on an event by event basis, A test module with one 3/spl times/3/spl times/30 mm BGO crystal, one 3 mm square PIN photodiode, and one photomultiplier tube is operated at -20/spl deg/C with an amplifier peaking time of 4 /spl mu/s, and a depth of interaction resolution of 5 to 8 mm fwhm measured. Simulations predict that this virtually eliminates radial elongation in a 60 cm diameter BGO tomograph. The photodiode signal corresponding to 511 keV energy deposit varies linearly with excitation position, ranging from 1250 electrons (e/sup -/) at the end closest to the photodiode to 520 e/sup -/ at the opposite end. The electronic noise is a position independent 330 e/sup -/ fwhm, so the signal to noise ratio is sufficient to reliably identify the crystal of interaction in a 64 element module. >

/pdf/design-studies-for-a-pet-detector-module-using-a-pin-16gyjskel3.pdf

Design studies for a PET detector module using a PIN photodiode to measure depth of interaction

A novel type of positron camera achieving nearly isotropic spatial resolution and uniform sensitivity throughout the useful field of view is proposed. The design is based on a quad scintillator dual-channel detector module using two avalanche photodiodes as photodetectors. Original features include bilayered scintillators to measure the depth of interaction, tilted detectors relative to the radial direction to ensure proper sampling in a stationary mode, multiring capability, and an independent acquisition channel for every discrete detector. Significantly better resolution is obtained toward the edges of the camera field, and the sampling uniformity near the center is improved. Multiring systems with a slice thickness of 5 mm can be implemented to achieve true volume imaging with symmetrical 3-D spatial resolution. The performance of the proposed design is analyzed in terms of resolution, sensitivity, contrast, and field uniformity. >

Design of a high resolution positron emission tomograph using solid state scintillation detectors

The problem of extracting point spread functions from detector aperture functions in high-resolution PET is addressed. In the limit of very small size detectors relative to the ring dimensions, assumptions are made that lead to a fast and simple computation model yielding point spread functions with negligible errors due to the reconstruction algorithm. The methods allows one to assess accurately the intrinsic performance of a PET tomograph, and it appears to be adequate to relate the imaging capabilities in every point of the camera reconstruction field to the geometric and physical characteristics of the detection system. The method was developed as an investigation tool to help design the next generation of very-high-resolution PET tomographs. >

Fast point spread function computation from aperture functions in high-resolution positron emission tomography

Silicon avalanche photodiodes (APD) can be used as photodetectors in combination with scintillation crystals for the detection of radiation. These devi

Status of BGO-avalanche photodiode detectors for spectroscopic and timing measurements

The major loss of signal in ideal rectangular parallelepipedic scintillators results from the presence of permanent modes which cannot escape the crystal volume. When the photodetector is coupled with an index of n d =1.55, this value can be as high as 46% in BGO. The purpose of this study is to investigate some alternatives to this geometry with the aid of Monte Carlo simulations and direct measurements on a set of crystals. The study is focused on long narrow-shaped BGO scintillators designed to be used in conjunction with avalanche photodiodes in positron emission tomography. Results show that slightly unpolishing the crystal surfaces efficiently eliminates trapped modes but that very-high-reflectivity reflectors are needed to transport photons to the detection plane. Minor modifications to the geometry eliminating all symmetry from the crystal also efficiently eliminate trapped modes and generally lead to the highest light output.

Effect of geometrical modifications and crystal defects on light collection in ideal rectangular parallelepipedic BGO scintillators

This work proposes a theoretical model of the light transport in rectangular parallelepipedic scintillators The model is applied to the evaluation of light collection from BGO scintillators used in high-resolution positron emission tomography Results show that the major limitation of signal is due to the presence of trapped photons in the crystal volume Covering the plane parallel to the detection surface with a high-refractive-index diffuse reflector or unpolishing of this surface appear as the best solutions to detrap photons to the profit of the detection plane Some modifications to the rectangular parallelepipedic configuration which would yield similar or better results are also proposed, but can only be evaluated by a Monte Carlo simulation

C. Carrier

Papers

Design of a high resolution positron emission tomograph using solid state scintillation detectors

Fast point spread function computation from aperture functions in high-resolution positron emission tomography

Status of BGO-avalanche photodiode detectors for spectroscopic and timing measurements

Effect of geometrical modifications and crystal defects on light collection in ideal rectangular parallelepipedic BGO scintillators

Theoretical modelling of light transport in rectangular parallelepipedic scintillators