Single Scatter Simulation Algorithm Research Articles

PET image reconstruction using the Origin Ensemble algorithm and geometric constraints

Abstract The Origin Ensemble method allows image reconstruction of photon-limited emission tomography data to be performed entirely in the image domain. This offers attractive perspectives such as including scatter events for image reconstruction in Positron Emission Tomography. In this work, the probability of single Compton scatter along a line-of-response is estimated by the Single Scatter Simulation algorithm; for every event a decision is made whether this event is reconstructed along a line or an area confined by two circular arcs holding potential scatter points. First results of 2D simulations show visual agreement with the reference and locally increased contrast recovery coefficient values.

Incorporation of a two metre long PET scanner in STIR

The Explorer project aims to investigate the potential benefits of a total-body 2 metre long PET scanner. The following investigation incorporates this scanner in STIR library and demonstrates the capabilities and weaknesses of existing reconstruction (FBP and OSEM) and single scatter simulation algorithms. It was found that sensible images are reconstructed but at the expense of high memory and processing time demands. FBP requires 4 hours on a core; OSEM: 2 hours per iteration if ran in parallel on 15-cores of a high performance computer. The single scatter simulation algorithm shows that on a short scale, up to a fifth of the scanner length, the assumption that the scatter between direct rings is similar to the scatter between the oblique rings is approximately valid. However, for more extreme cases this assumption is not longer valid, which illustrates that consideration of the oblique rings within the single scatter simulation will be necessary, if this scatter correction is the method of choice.

A Hybrid Scatter Correction Method Combining the SSS Algorithm and Beam Stopper Method for 3D PET

Scattered events represent a major cause of inaccurate quantification and image degradation in three-dimensional (3D) positron emission tomography (3D PET) scanners. The single scatter simulation (SSS) method is the most widely used scatter correction method. However, this method relies on tail data to scale the estimated scatter to the measured data. Consequently, this method might fail when tail data are insufficient; for example, when scanning obese patients. Unfortunately, the global prevalence of obesity has increased substantially in recent decades; hence, developing alternative solutions to this problem is necessary. The beam stopper (BS) method is used to measure the actual scatter at locations blocked by the BS device. This method can then be applied to estimate the entire scatter sinogram by using an interpolation. However, this interpolation might not reflect the true condition if the BS device has been set up inappropriately. This study presents a hybrid scatter correction method (SSS-BS) based on the SSS and BS methods. In the SSS-BS method, the scatters measured by applying the BS method are used to adjust the scatter distribution calculated by employing the SSS method, thereby obtaining a more accurate scatter distribution. The performance of this hybrid method was investigated by using Monte Carlo simulations. A thorax Zubal phantom (55.8 cm in width and 37.8 cm in height) and a Jaszczak-like phantom (55.2 cm in diameter and 22.75 cm in length) were used to investigate cases involving obese patients. In contrast to the SSS method, the hybrid method does not suffer from the scatter-distribution inaccuracies caused by insufficient tails. The results from the Zubal phantom show that the SSS-BS, SSS, and BS methods yield similar scatter-fraction (SF) estimates, whereas the SSS-BS and SSS methods have a slightly smaller root mean square error (RMSE) of reconstructed images than that of the BS method. For the Jaszczak-like phantom, the SSS-BS method generates a more accurate estimate of SF than both the SSS and BS methods do, reducing the RMSE by nearly half. This study demonstrates that the hybrid method enables the accurate scatter-distribution estimation of 3D PET in cases involving obese patients.

Correction Technique for Cascade Gammas in I-124 Imaging on a Fully-3D, Time-of-Flight PET Scanner

It has been shown that I-124 PET imaging can be used for accurate dose estimation in radio-immunotherapy techniques. However, I-124 is not a pure positron emitter, leading to two types of coincidence events not typically encountered: increased random coincidences due to non-annihilation cascade photons, and true coincidences between an annihilation photon and primarily a coincident 602 keV cascade gamma (true coincidence gamma-ray background). The increased random coincidences are accurately estimated by the delayed window technique. Here we evaluate the radial and time distributions of the true coincidence gamma-ray background in order to correct and accurately estimate lesion uptake for I-124 imaging in a time-of-flight (TOF) PET scanner. We performed measurements using a line source of activity placed in air and a water-filled cylinder, using F-18 and I-124 radio-isotopes. Our results show that the true coincidence gamma-ray backgrounds in I-124 have a uniform radial distribution, while the time distribution is similar to the scattered annihilation coincidences. As a result, we implemented a TOF-extended single scatter simulation algorithm with a uniform radial offset in the tail-fitting procedure for accurate correction of TOF data in I-124 imaging. Imaging results show that the contrast recovery for large spheres in a uniform activity background is similar in F-18 and I-124 imaging. There is some degradation in contrast recovery for small spheres in I-124, which is explained by the increased positron range, and reduced spatial resolution, of I-124 compared to F-18. Our results show that it is possible to perform accurate TOF based corrections for I-124 imaging.

Optimization of a fully 3D single scatter simulation algorithm for 3D PET

We describe a new implementation of a single scatter simulation (SSS) algorithm for the prediction and correction of scatter in 3D PET. In this implementation, out of field of view (FoV) scatter and activity, side shields and oblique tilts are explicitly modelled. Comparison of SSS predictions with Monte Carlo simulations and experimental data from uniform, line and cold-bar phantoms showed that the code is accurate for uniform as well as asymmetric objects and can model different energy resolution crystals and low level discriminator (LLD) settings. Absolute quantitation studies show that for most applications, the code provides a better scatter estimate than the tail-fitting scatter correction method currently in use at our institution. Several parameters such as the density of scatter points, the number of scatter distribution sampling points and the axial extent of the FoV were optimized to minimize execution time, with particular emphasis on patient studies. Development and optimization were carried out in the case of GSO-based scanners, which enjoy relatively good energy resolution. SSS estimates for scanners with lower energy resolution may result in different agreement, especially because of a higher fraction of multiple scatter events. The algorithm was applied to a brain phantom as well as to clinical whole-body studies. It proved robust in the case of large patients, where the scatter fraction increases. The execution time, inclusive of interpolation, is typically under 5 min for a whole-body study (axial FoV: 81 cm) of a 100 kg patient.

Assessment of the impact of model-based scatter correction on [18F]-FDG 3D brain PET in healthy subjects using statistical parametric mapping

It is recognized that scatter correction can supply more accurate absolute quantification, and that iterative reconstruction results in better noise properties and significantly reduces streak artefacts; however, it is not entirely clear whether they produce significant changes in [18F]-FDG distribution of reconstructed 3D brain PET images relative to not scatter corrected images and analytic reconstruction procedures. The current study assesses the effect of model-based scatter correction using the single-scatter simulation algorithm and iterative reconstruction in 3D brain PET studies, using statistical parametric mapping (SPM) analysis. The study population consisted of 14 healthy volunteers (6 males, 8 females; age 63–80 years). PET images were reconstructed using an analytic 3DRP reprojection algorithm with (SC) and without explicit scatter correction (NSC), as well as using an iterative ordered subset–expectation maximization (OSEM) algorithm. Calculated attenuation correction was performed assuming uniform attenuation (μ = 0.096 cm−1) for brain tissues when data are precorrected for scatter. The broad-beam attenuation coefficient (μ = 0.06 cm−1) determined from phantom studies was applied to NSC images. The images were coregistered and normalized using the default [15O]-H2O template supplied with SPM99 and an [18F]-FDG template. A t statistic image for the contrast condition effect was then constructed. The contrast comparing SC to NSC images suggest that regional brain metabolic activity decreases significantly in the frontal gyri, in addition to the middle temporal and postcentral gyri. On the other hand, activity increases in the cerebellum, thalamus, insula, brainstem, temporal lobe, and the frontal cortex. No significant changes were detected when comparing images reconstructed using analytic and iterative algorithms. It is concluded that, for some cerebral areas, significant differences in [18F]-FDG distribution arise when images are reconstructed with and without explicit SC. This needs to be considered when interpreting [18F]-FDG 3D brain PET images after applying SC.

Fast implementation of the single scatter simulation algorithm and its use in iterative image reconstruction of PET data

In positron emission tomography (PET), scatter correction is usually performed prior to image reconstruction using a more or less exact model of the scatter processes. These models require estimates of the true activity and object density distributions of the imaged object. The problem is that these estimates are computed from measured data and, therefore, already contain scattered events. The purpose of this work was to overcome this problem by incorporating scatter characteristics directly into the process of iterative image reconstruction. This could be achieved by an optimized implementation of the single scatter simulation (SSS) algorithm, which results in a significant speed-up of the scatter estimation procedure. The scatter simulation was then included in the forward projection step of maximum likelihood image reconstruction. The results demonstrate that this approach leads to a more exact estimation of the scatter component which cannot be obtained by a simple sequential data processing strategy.

Model‐based scatter correction for positron emission tomography (in German)

Positron emission tomography (PET) images include a significant scatter component. The presence of the scatter manifests itself as a loss of spatial resolution and as an apparent migration of activity from hot to cold regions. Monte Carlo (MC) based approaches have proved to be very useful in simulating the image‐forming process in the PET scanner. However, fully 3D MC simulations are computationally too intensive to be applied in clinical routine. Consequently, many efforts have been undertaken to develop approximate scatter correction techniques for PET. Scatter correction is generally performed prior to image reconstruction using an appropriate model of the scatter process. These models require estimates of the correct emission and attenuation distribution in the imaged object. The problem is that these estimates are computed from measured data and, therefore, already contain scattered events. The purpose of this work was to overcome this problem by incorporating scatter characteristics directly into the process of iterative image reconstruction. This was achieved by an optimized implementation of the Single Scatter Simulation (SSS) algorithm resulting in a significant speed‐up of the scatter estimation procedure (from about 10 min to 30 s). The computationally improved SSS algorithm was then included in the forward projection step of maximum likelihood image reconstruction. Results obtained from phantom measurements demonstrate that this approach leads to a better estimation of the scatter component than can be obtained by a simple sequential data processing strategy.