IEEE 2017 NSS/MIC/RTSD ControlCenter

Online Program Overview Session: M-19

To search for a specific ID please enter the hash sign followed by the ID number (e.g. #123).

Simulation & Modeling of Systems

Session chair: Adam M. Alessio; Georges El Fakhri
 
Shortcut: M-19
Date: Saturday, October 28, 2017, 08:00
Room: Centennial IV
Session type: MIC Session

Contents

8:00 am M-19-1

Simulation Study of the Generation of Cerenkov Photons in Bismuth Germanate for PET (#2161)

S. I. Kwon1, S. R. Cherry1

1 University of California, Davis, Biomedical Engineering, Davis, California, United States of America

Content

Bismuth Germanate (BGO) was a very attractive scintillator in early-generation positron emission tomography (PET) scanners. However, its lower scintillation light yield and longer rise and decay time compared to lutetium–yttrium oxyorthosilicate (LYSO) result in degradation of coincidence timing resolution. Recently, it was reported that the coincidence timing resolution of BGO can be much improved by detecting promptly generated Cerenkov photons, while scintillation photons are still used for determining energy information. In BGO, the estimated number of generated Cerenkov photons produced per photoelectric interaction can be calculated using the relevant equations. However, it is difficult to estimate the number of Cerenkov photons generated by Compton interactions. Further, the direction of the emitted Cerenkov photons depends upon the momentum of the energetic electrons, while scintillation photons are emitted isotropically. The directionality and transport of the electrons and subsequent Cerenkov photons have not been well characterized. In this study, we perform Monte Carlo simulations using Geant4 v10.3 to interrogate the short time period between the release of energetic electrons following a 511 keV interaction and the production of Cerenkov photons. Distributions of initial kinetic energy and the travel length of energetic electrons were obtained. The number and directional dependency of Cerenkov photons were carefully studied for different interaction types. The number of Cerenkov photons produced by photoelectric interactions was ~16 over a wavelength range of 320–800 nm. Cerenkov photons were also produced by energetic electrons released by Compton scattering. The directions of the generated Cerenkov photons was dependent on the incident direction of the 511 keV photons, especially for the earliest generated Cerenkov photons. The directional patterns of Cerenkov photons were different dependent on the type of interaction.

Keywords: Bismuth Germanate (BGO), Cerenkov, Simulation, positron emission tomography (PET)
8:18 am M-19-2

First Studies of Tracking Cerenkov Light in Scintillation Crystals with a New Light Transport Model in GATE (#1344)

E. Roncali1, S. I. Kwon1, M. Stockhoff1, E. Berg1, S. R. Cherry1

1 University of California Davis, Department of Biomedical Engineering, Davis, California, United States of America

Content

Improving scintillation detector timing resolution for time-of-flight (TOF) positron emission tomography (PET) is a major focus in the field of PET detector research. Although not optimal for traditional time triggering on the early scintillation photons due to its slow decay time and low light yield, bismuth germanate (BGO) is an excellent candidate to generate a prompt timing trigger based on Cerenkov light produced in an extremely short time by energetic electrons from a photoelectric or Compton interaction. Triggering detectors on this extremely fast Cerenkov light has been recently demonstrated to achieve excellent timing resolution in short BGO crystals, but is complicated by the very low (~15) number of Cerenkov photons produced per interaction. More detailed understanding of the transport and collection of Cerenkov photons is needed to optimize their use for effective triggering of the detectors. 

In this work, we use advanced optical Monte Carlo simulation to study the transport and detection of Cerenkov and scintillation light simultaneously. We generate and track Cerenkov and scintillation photons with GATE, using a new model of light transport in crystals that we have recently published. This crystal reflectance model allows accurate description of the light propagation in the scintillator, including the detected photon time stamps subject to the travel time. Several surface finishes and reflectors can be used to study photon detection and timing properties.

We demonstrate the feasibility of the complete simulation of Cerenkov and scintillation photons, from generation to detection and compare the behavior of Cerenkov and scintillation photons at many levels including detection time stamps, travel time in the scintillator, and the number of reflections on the crystal side.

To our knowledge this is the first complete simulation of the generation, transport, and detection of Cerenkov photons combined with scintillation photons for TOF detectors.

Keywords: Cerenkov light; time-of-flight; optical simulation; light transport modeling; BGO
8:36 am M-19-3

Improved Woodcock tracking on Monte Carlo simulations for medical applications (#1849)

A. Behlouli1, J. Bert1, D. Visvikis1

1 LaTIM, INSERM UMR1101, CHRU Brest, Brest, France

Content

This paper present a new variance reduction technique called Super Voxel Woodcock (SVW), which combines Woodcock tracking technique with the super voxel concept, used in computer graphics. It consists in grouping the voxels of the volume in a super voxel grid (pre-processing step) by associating to each of the super voxels a local value of the most attenuate medium which will later serve to the interaction distances sampling. SVW allows reducing the sampling of the particle path while a high-density material is present within the simulated phantom. In order to evaluate the performance of the SVW method compare to both standard and woodcock tracking methods, algorithms were implemented within the same GPU MCS framework GGEMS. A MCS of prostate brachytherapy was used as benchmark, showing that SVW method provide the same dosimetry results than the Woodcock tracking method or the standard particle navigation approach. Calculation of the efficiency factor has shown that the proposed method improves significantly the performance of the Woodcock tracking method by a factor of 2.7. The next step will consist to estimate the improvement of this method considering different medical applications context.

Keywords: Monte Carlo Simulation, Variance reduction techniques, Woodcock tracking
8:54 am M-19-4 Download

GATE simulations to study extended axial FOVs for the PennPET Explorer scanner (#2179)

V. Viswanath1, M. E. Daube-Witherspoon1, M. E. Werner1, S. Surti1, A. Trindade3, P. Rodrigues3, A. E. Perkins2, J. S. Karp1

1 University of Pennsylvania, Department of Radiology, Philadelphia, Pennsylvania, United States of America
2 Philips Healthcare, Highland Heights, Ohio, United States of America
3 Philips Research, Oncology Solutions, Eindhoven, Netherlands

Content

Clinically, PET/CT is a widely used technology in oncology and cardiology. In the realm of research, dynamic PET imaging provides valuable information about ongoing biologic processes in the body. Currently, all commercial PET scanners have an axial field of view (AFOV) less than 25 cm, and whole-body scans are taken in multiple static, step-and-shoot bed positions.  The increased sensitivity from extending the AFOV will allow for lower dose imaging, improved count statistics, and simultaneous whole-body static and dynamic imaging. To quantify the effect of extending the axial FOV of a PET scanner, we have created GATE simulations of the 16.4-cm AFOV Philips Vereos PET/CT scanner, and extended this to 23-cm (E23) and 70-cm (E70) AFOV PET scanners. The simulation model is based on the 1:1 coupled digital detectors used in the Vereos scanner that has a timing resolution of 320 ps. NEMA NU2-2012 sensitivity, spatial resolution, count rate, and image quality standards were run on each of the three simulated scanners and corroborated with measured data from the Vereos. While there was little variation in CRC values across the three geometries, there were notable differences in sensitivity, spatial resolution, and count rate. Sensitivity increased from 5.5 kcps/MBq on the Vereos to 10.8 kcps/MBq on the E23 and 90.5 kcps/MBq on the E70. Transaxial spatial resolution at the scanner center (FWHM: 4.0 mm) did not degrade with increased AFOV, while the axial resolution at the center degraded slightly from 3.78 mm (Vereos) to 4.31 mm (E70). We will continue to study the axial resolution as a function of AFOV. At 55 kBq/cc - the measured peak NECR (155 kcps) for the Vereos - the NECR were 551 kcps and 4443 kcps for the E23 and E70 scanners, respectively. Moving forward, these simulated scanners, along with scanners with even longer AFOVs, will assist in planning both static and dynamic patient studies on the PennPET E70 prototype once operational.

Keywords: PET Imaging, GATE, NEMA, PennPET Explorer
9:12 am M-19-5 Download

Simulataneous Generation of X-ray and Range Images using XCAT under Motion (#2435)

O. Rybakov1, B. Bier1, J. Maier1, M. Unberath1, A. Maier1

1 Friedrich-Alexander University Erlangen-Nürnberg, Erlangen, Germany

Content

Novel algorithms in the field of X-ray imaging are commonly evaluated on simulation software first, before they are implemented on a clinical scanner in order to test their performance in a very controlled setup. This reduces patient dose and facilitates the development of new approaches and methods. In recent years, range imaging applications were established in the field of medical imaging. Range imaging showed its potential in radiotherapy, augmented reality, and, recently, also for motion correction in cone-beam computed tomography. In this work, we present an open-source software tool that allows to generate X-ray projections and surface information in one step completely on the graphics processing unit. To this end, we extend a state-of-the-art rendering method based on an append buffer. Arbitrary imaging geometries for the X-ray source and the range imaging camera can be selected. We test the proposed method on the XCAT phantom in two fundamentally different scenarios: a weight-bearing acquisition showing a squat, and a supine acquisition of a breathing patient.

Keywords: range imaging, x-ray imaging, append buffer, weight-bearing, open-source, XCAT
9:30 am M-19-6 Download

Optimizing molecular breast tomosynthesis with slit-slat collimators (#3265)

J. van Roosmalen1, F. Beekman1, 2, M. C. Goorden1

1 Delft University of Technology, Biomedical Imaging, Delft, Netherlands
2 MILabs BV, Utrecht, Netherlands

Content

Planar imaging of 99mTc labelled tracers is gaining popularity for detecting breast tumors. Recently, we proposed a novel design for molecular breast tomosynthesis (MBT) based on two sliding focusing multi-pinhole collimators that scan a lightly compressed breast. In simulations, we showed that MBT improved the tumor-to-background contrast-to-noise ratio significantly over state-of-the-art planar molecular breast imaging. We are currently considering different types of collimators for MBT and we are calculating optimized collimator-detector design parameters. The aim of the present work is to optimize the MBT geometry for slit-slat collimation.

Using analytical models, we optimized sensitivity of MBT equipped with a slit-slat collimator at different fixed system resolutions (ranging from 5 to 10 mm) by tuning the distance between slats, the slit diameters and the distance between breast and detector for a whole series of automatically generated slit-slat designs. For each fixed system resolution, we determined the maximum sensitivity that could be reached and compared this with sensitivity that could be obtained by optimized pinhole and fan beam collimators from earlier work. We found that at equal system resolution, slit-slat collimation always gave a higher sensitivity than optimized pinhole collimation with improvements ranging from 5% (at 5 mm system resolution) up to 23% (10 mm system resolution). On the contrary, slit-slat collimation resulted in a lower sensitivity than optimized fan beam collimation at the same system resolution, with sensitivity being lower by 12% and 38% at 10 mm and 5 mm system resolution respectively. Realistic full system simulations with breast phantoms will have to be performed to decide which collimator type and which system resolution are optimal for breast tumor detection with MBT.

Keywords: SPECT, tomosynthesis, breast imaging