Improving confidence for IMRT and helical tomotherapy treatments using accurately benchmarked Monte Carlo simulations

(2008)

Files

These_Edmond_Sterpin.pdf
  • Open Access
  • Adobe PDF
  • 3.49 MB

Details

Authors
Supervisors
Vynckier, Stefaan
Abstract
(en) The radiotherapy community's dream has always been to have available technologies being able to deliver a perfectly accurate high dose to tumor volumes without irradiating surrounding healthy tissues. This dream will never come true, but all the efforts of physicists, physicians and manufacturers must be combined to bring reality as close as possible to this utopia. Since the beginning of the 60s and the advent of electron linear accelerator mounted on treatment units, the technology in photon therapy has evolved tremendously. Nowadays, the state of art in photon treatments involves intensitymodulated radiotherapy (IMRT) with the help of highly sophisticated imaging devices. IMRT is a complex technique which requires a good control in the accuracy of all the steps involved in a radiotherapy treatment. Those may be summarized into 3 main categories: the calibration and stability of the treatment unit, the positioning and the quality of the data associated to the patient and the dose calculation process during treatment planning. To improve the overall uncertainty of a given treatment, research efforts are needed in the three categories. This thesis focuses on the dose calculation process. During the last decades, the complexity and accuracy of dose calculation algorithms increased with the huge improvements in computer technology. Despite this, the vast majority of the algorithms perform important approximations regarding the physics of the particles transport. However, it is possible to design algorithms with minor approximations based on Monte Carlo methods (MC), which faithfully simulate the physical reality and are considered as the most accurate to compute dose in human tissue. Unfortunately, until recently, the speed of MC simulations was too slow to be compatible with the time constraints associated with clinical routine. But the continuous improvements in computer power combined with the introduction of pertinent simplifications in the MC codes allowed for the introduction of MC into clinical practice, which is already available in several commercial treatment planning systems. The aim of this thesis was to evaluate the added-value of MC compared to modern analytical algorithms for complex IMRT treatments of tumors surrounded by an important amount of inhomogeneities. Those evaluations were performed for two IMRT delivery techniques: step-and-shoot IMRT and helical tomotherapy. For stepand- shoot IMRT (delivered by an Elekta treatment unit), MC simulations with BEAMnrc were compared to an algorithm released recently by Varian called the Anisotropic Analytical Algorithm (AAA). For helical tomotherapy, a similar study "Improving confidence for IMRT and helical tomotherapy treatments using accurately benchmarked Monte Carlo simulations" was performed but for the collapsed-cone convolution/superposition code implemented in the tomotherapy system provided by Tomotherapy Incorporated. During this second study, the MC modeling of tomotherapy based on the MC code PENELOPE was an important aspect since this was the first time that a complete MC model of tomotherapy was designed with all the machine details fully disclosed by the manufacturer. Moreover, the MC model, called TomoPen, was designed for future implementation in the clinical system. Therefore, the speed of the simulation was a critical issue. The simulation strategy adopted in TomoPen, which mainly consisted in symplifying the transport through the multi-leaf collimator, allowed the calculation of dose distributions in a bilateral head-and-neck tumor in around 6 hours on a single 2.5 GHz processor, without losing significant accuracy. Using the cluster delivered with each tomotherapy treatment unit, this simulation time could be reduced by a factor of at least 32. Generally, a good agreement was achieved between analytical algorithms and MC for the clinical cases selected during this study. However, differences could be observed for critical situations...
Affiliations

Citations

Sterpin, E. (2008). Improving confidence for IMRT and helical tomotherapy treatments using accurately benchmarked Monte Carlo simulations. https://hdl.handle.net/2078.5/129786