M5A1  Imaging in Radiotherapy

Friday, Nov. 6  08:30-10:00  Golden Pacific Ballroom

Session Chair:  Taiga Yamaya, National Institute of Radiological Sciences, Japan; Ramsey Badawi, UC Davis Medical Center, United States

Show/Hide All Abstracts

(08:30) M5A1-1, Short-Lived Positron Emitters in Beam-on PET Imaging During Proton Therapy

P. Dendooven1,2, T. Buitenhuis1, F. Diblen1,3, P. Heeres1, A. Biegun1, F. Fiedler4, M.-J. van Goethem5, E. van der Graaf1, S. Brandenburg1

1KVI - Center for Advanced Radiation Technology, University of Groningen, Groningen, Netherlands
2Helsinki Institute of Physics, Helsinki, Finland
3MEDISIP, Department of Electronics and Information Systems, Ghent University, Ghent, Belgium
4Institute of Radiation Physics, Helmholtz-Zentrum Dresden - Rossendorf, Dresden, Germany
5Department of Radiation Oncology, University Medical Center Groningen, Groningen, Netherlands
Positron emission tomography is so far the only method for in-vivo dose delivery verification in hadron therapy that is in clinical use. PET imaging during irradiation maximizes the number of detected counts and minimizes washout. In such a scenario, also short-lived positron emitters will be observed. As very little is known on the production of these nuclides, we determined which ones are relevant for proton therapy treatment verification by measuring the production of short-lived positron emitters in the stopping of 55 MeV protons in water, carbon, phosphorus and calcium. Experiments were performed using the AGOR superconducting cyclotron at KVI – Center for Advanced Radiation Technology at the University of Groningen. The most copiously produced short-lived nuclides are: 12N (T1/2 = 11 ms) on carbon, 29P (T1/2 = 4.1 s) on phosphorus and 38mK (T1/2 = 0.92 s) on calcium. No short-lived nuclides are produced on water. The experimental production rates are used to calculate the production on PMMA and a representative set of 4 tissue materials. The number of decays per 55 MeV proton stopped in these materials, integrated over an irradiation, is calculated as function of the duration of the irradiation. The most noticeable result is that for an irradiation in (carbon-rich) adipose tissue, 12N will dominate the PET image up to an irradiation duration of 70 s. On bone tissue, 15O dominates over 12N after 8-15 s (depending on the carbon-to-oxygen ratio). Considering nuclides created on phosphorus and calcium, the short-lived ones provide 2.5 times more decays than the long-lived ones during a 70 s irradiation. Bone tissue will thus be better visible in beam-on PET compared to PET imaging after an irradiation. The results warrant detailed investigations into the energy-dependent production of 12N, 29P and 38mK and their effect on PET imaging during proton irradiations.

(08:45) M5A1-2, Prompt Gamma Imaging of a Pencil Beam with a High Efficiency Compton Camera at a Clinical Proton Therapy Facility

F. Hueso-Gonzalez1, J. Petzoldt2, K. E. Roemer3, S. Schoene3, F. Fiedler3, C. Golnik2, T. Kormoll2, G. Pausch2, W. Enghardt1

1Helmholtz-Zentrum Dresden - Rossendorf, Institute of Radiooncology, Dresden, Germany
2OncoRay - National Center for Radiation Research in Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, Technische Universität Dresden, Dresden, Germany
3Helmholtz-Zentrum Dresden - Rossendorf, Institute of Radiation Physics, Dresden, Germany
Protons are excellent particles for tumor treatment due to the increased ionization density close to their stopping point. In practice, the uncertainty on the particle range compromises the achievable accuracy. Compton cameras imaging prompt gammas, a by-product of the irradiation, have been proposed for indirect range verification years since. At UniversitaetsProtonenTherapieDresden, two BGO block detectors (from PET scanners) arranged as Compton camera are deployed for imaging tests with high energy prompt gamma rays produced in PMMA by a proton pencil beam. Target shifts, thickness increase and beam energy variation experiments are conducted. Each measurement lasts about 15 minutes at a low proton beam current. The effect of one centimeter proton range deviations on the backprojected images is analyzed. In conclusion, the experimental results highlight the potential application of Compton cameras for high energy prompt gamma imaging of pencil beams, as a real-time and in vivo range verification method in proton therapy.

(09:00) M5A1-3, A Compton Telescope for Ion-Beam Therapy Monitoring: from Monte Carlo Modeling to First on-Beam Tests

P. Solevi1, E. Muñoz1, C. Solaz1, M. Trovato1, J. Barrio1, P. Dendooven2,3, A. Etxebeste1, J. E. Gillam1,4, C. Lacasta1, J. F. Oliver1, M. Rafecas1,5, I. Torres-Espallardo1,6, G. Llosá1

1IFIC (CSIC/UV), Valencia, Spain
2KVI-Center for Advanced Radiation Technology, University of Groningen, Groningen, The Netherlands
3now at Helsinki Institute of Physics, Helsinki, Finland
4now at Brain & Mind Research Institute, University of Sydney, Sydney, Australia
5now at Institute of Medical Engineering, University of Lübeck, Lübeck, Germany
6now at Nuevo Hospital La Fe, Valencia, Spain
Ion-beam therapy is an effective tumor treatment modality that requires non invasive in-vivo monitoring techniques in order to be fully exploited in the clinical setting. A verification technique is provided by the detection of the prompt gammas exiting the patient body. For this purpose we developed a three-layer Compton Telescope based on Lanthanium Bromide (LaBr3) crystals and Silicon Photo-Multipliers (SiPMs). The first two layers have been extensively characterized in different experimental setups employing Na22 and Am-Be sources. An intrinsic detector resolution of about 1 mm (FWHM) and an energy resolution of 6.4% and 7.4% (FWHM at 511 keV, first and second layer, respectively) have been measured. The dynamic range of the detector was tested up to several MeV by measuring an Am-Be source, providing a few percents deviation from linearity at 3.44 MeV. The Compton Telescope was tested on-beam at the irradiation facility of the AGOR cyclotron at KVI- Center for Advanced Radiation Technology, University of Groningen, where different targets (graphite and PMMA) have been irradiated with a proton beam at 150 MeV energy. The set-up was simulated in order to inspect the measured spectra and optimize the background rejection mostly due to neutron detection.

(09:15) M5A1-4, Patient Data-Based Monte Carlo Simulation of in-Beam Single-Ring OpenPET Imaging

H. Tashima1, C. Kurz2,3, E. Yoshida1, J. Debus3, K. Parodi2,3, T. Yamaya1

1Molecular Imaging Center, National Institute of Radiological Sciences, Chiba, Japan
2Ludwig-Maximilians University, Munich, Germany
3Heidelberg University Hospital, Heidelberg, Germany
We are developing the OpenPET, which can provide an open space to make the patient accessible during positron emission tomography (PET) measurements. The OpenPET provides PET measurements during particle therapy such as carbon ion beam treatment, enabling visualization of the irradiation field of the patient as well as direct tumor tracking with a real-time imaging system. In the carbon ion beam treatment, positron emitters are produced by the fragmentation reactions. Therefore, by means of PET, it can be confirmed whether or not the dose distribution in the treatment plan and that in the actual irradiation are matched. In this study, we conducted Monte Carlo simulations based on clinical patient data for scanned 12C ion beam irradiation to investigate the effectiveness of the OpenPET geometry compared with offline PET measurements. Production rates of the positron emitters were calculated based on the patient CT image and physical dose information in the treatment plan by the FLUKA code. The activity distribution averaged over the measurement time was calculated from the production rates by an analytical model including activity decay and biological washout effects. Annihilation gamma ray events from the activity distribution were simulated by the Geant4 Monte Carlo toolkit. Activity distribution images were reconstructed by the list-mode maximum a posteriori ordered subset expectation maximization. As a result, we found the online OpenPET with a measurement time of 10 minutes could detect a similar number of list-mode data events as an offline PET with a measurement time of 30 minutes. Reconstructed images for the online OpenPET contained activity distribution not only around the distal edge but also at the entrance area of the irradiation field. The offline PET showed strong noise before the Bragg peak position. The simulation results showed that online OpenPET could achieve good quality of images with a short measurement time.

(09:30) M5A1-5, 3D Dose Reconstruction in External Beam Radiotherapy Using Portal Imaging

L. Autret1, J. Bert1, Y. Lemaréchal1, L. Desbat2, D. Visvikis1

1LaTIM, INSERM UMR 1101, CHRU Brest, Brest, France
2TIMC-IMAG, CNRS and University Joseph Fourier (UMR5525), Grenoble, France
Only few methods exist for in-vivo dosimetry allowing the determination of an accurate and fully personalized 3D dose distribution for the most current state of the art irradiation techniques. In this work, we propose a 3D dose reconstruction method based on Monte Carlo simulations allowing an accurate and fully personalized dose estimation from an Electronic Portal Imaging Device (EPID). This previously proposed method supposes that the cumulative dose distribution inside the patient can be related to the cumulative dose deposition in the EPID using kernel matrices. These matrices represent patient and EPID dose responses to a set of impulsional sources and they can be estimated using Monte Carlo (MC) simulations. To enable its use in a clinical context, a previously proposed methodology has been adapted, taking advantage of the sparsity of the matrix system linking the patient cumulative dose distribution to the EPID cumulative dose deposition. In addition, Graphic Processing Units (GPUs) were used in kernel matrix estimations in order to reduce the associated estimation computational time. The validation of this proposed sparse dose reconstruction was first realized on simulated datasets. An EPID image was simulated and recorded to perform a 3D dose reconstruction for a head and neck patient CT. Simulated and reconstructed dose distributions within the patient show good agreement. In conclusion, the proposed approach allows a clinically feasible 3D dose reconstruction in external beam radiotherapy using portal imaging.

(09:45) M5A1-6, Feasibility Study on Real-Time Tumor Tracking Using a Parallel Plane PET

M. Ishikawa1, S. Tanabe2, S. Yamaguchi3, N. Ukon4, N. Miyamoto5, R. Suzuki5, N. Katoh4, K. Yasuda4, H. Shirato4

1Department of Biomedical Science and Engineering, Graduate School of Health Science, Hokkaido University, Sapporo, Japan
2Department of Radiotherapy, Niigata University Hospital, Niigata, Japan
3Department of Radiology, Iwate Medical University, Morioka, Japan
4Department of Radiation oncology, Graduate School of Medicine, Hokkaido University, Sapporo, Japan
5Department of Medical Physics, Hokkaido University Hospital, Sapporo, Japan
Objective
Molecular imaging is one of the important modalities in delineating tumors particularly in radiotherapy treatment planning. If the real-time tumor position can be detected using molecular imaging during radiotherapy, it may be helpful for gated irradiation. The purpose of this study to investigate a feasibility of a beam gating system for radiotherapy using real-time molecular imaging which was conducted by simulating a parallel plane PET system.
Materials and Methods
Assuming that the motion of the positron source is constrained to the central plane, the source position can be calculated from a cross point of the Line of Response (LOR) and the central plane between detector surfaces. If a positron source is located at the ISO center, distribution of the cross points might be blurred due to random/scattered coincidence. Center Located Ratio (CLR) was defined as a ratio of LORs passing through the ISO center divided by the entire LORs. When dislocation for perpendicular direction is occurred, a distribution of cross points will be spread out and associated decrease of CLR value will be expected.
Results and discussions
The behavior between real measurement and simulation was similar on proto-type experiments, however, the result from simulation for demonstrator might be different from actual measurement. RTRT system recognizes the position of a gold marker in the rate of 30 fps using two X-ray television systems. It is shown that 15,000 events per second will be needed for an appropriate gating irradiation to recognize discrepancy over 2mm of time resolution in the parallel plane PET system demonstrator.
Conclusions
A feasibility study was carried out to verify the potential for gating irradiation of tumors with real-time molecular imaging using a parallel plane PET system. For an parallel plane PET system demonstrator, the possibility of detecting the tumor position with an accuracy of 2 mm from the ISO center with 500 events.

(10:00) M5A1-7, Student Paper Award Presentation

A. M. Alessio

University of Washington, Seattle, WA, USA
Description: For outstanding student poster or oral papers as desired by each of the technical committees of NPSS that organizes a conference. The purpose of these awards is to encourage both outstanding student contributions and greater student participation as principal or sole authors of papers as well as to acknowledge the importance of student contributions to the fields embraced by the NPSS. Prize: The two best papers (two awards) will receive cash awards of $500 each and a Certificate. The two runners-up will receive a certificate only. Funding: Funded by the budget of each conference, as determined by each of the individual conferences sponsored by IEEE NPSS. Eligibility: Any student who is the principal or sole author/researcher and the presenter of either a poster or oral paper at any IEEE NPSS conference that has chosen to provide outstanding student awards, and who has been identified as an eligible student author, will be eligible. If there is a tie, preference will be given 1) to IEEE NPSS members; 2) to IEEE members; or 3) to non-IEEE members. Basis for Judging: All candidates for selection must have identified themselves either at the time of abstract submittal or no later than registration. The on-site awards committee will rank the papers for technical content and originality first. Other criteria such as graphic display and clarity of data presentation may be considered.