Simulation plays a critical role in the development of new emission tomography scanners and innovative applications of such scanners. These applications include the detection and monitoring response to therapy of a wide variety of diseases. For the past 19 years, our group has been developing a Monte Carlo simulation system specifically designed to simulate positron emission tomography (PET) and single emission tomography (SPECT) scanners. The work has been in support of our funded research projects in instrument development and optimization. From the first, we designed the code to be easily implemented on most computer systems. Recognizing that others might want to use the code as well, we adapted formal software engineering approaches in the design and coding of the software. The code has proven to be very useful to a wide user community (over 250 current users that we know about) and has been a key in many research projects. Our goal is now to extend parts of the code to make it more useful for researchers working on projects that have different needs than our own work, and to adopt a more formal open source release so that the work of others to extend the code can be incorporated in a controlled and well documented fashion to the existing code base. The proposed additions to the code include more acceleration techniques, dynamic scans, improved collimator simulations, and more efficient user interfaces and documentation. To address the increasing need of researchers for multiple realizations (e.g. thousands) of data sets, we also propose to recode our internal analytic simulation code (ASIM) and provide both ASIM and SimSET with a common interface and object descriptor meta language to improve consistency, error checking, and ease of use. The codes will be re-designed so that SimSET can be used to determine the parametric models ASIM requires, making the two tools a synergistic solution to many compute-bound problems faced by us and other users of SimSET. The codes fill an important role in emission tomograph research. While other Monte Carlo systems such as GATE can provide more detailed simulations of tomographs, they do so at a significant cost in speed. SimSET and ASIM allow the user to more finely tune the level of detail desired and then to exploit the many acceleration techniques in SimSET and ASIM to obtain an efficient simulation tool for the medical imaging task at hand. Establishing both SimSET and ASIM as formal open source projects will also expedite the enhancement of these two tools for a wide variety of research applications.
|Harrison, Robert L (2010) Monte Carlo Simulation Of Emission Tomography And Other Medical Imaging Techniques. AIP Conf Proc 1204:126-132|
|Hunter, William C J; Harrison, Robert L; Gillispie, Steven B et al. (2009) Parametric Design Study of a Long Axial Field-of-View PET Scanner using a Block-Detector Tomograph Simulation of a Cylindrical Phantom. IEEE Nucl Sci Symp Conf Rec (1997) 2009:3900-3903|
|Harrison, Robert L; Kinahan, Paul E; Lewellen, Thomas K (2007) A generalized simulation description language. IEEE Nucl Sci Symp Conf Rec (1997) 5:4012-4014|