Premium
Po‐Poster ‐ 07: Commissioning of virtual linacs for Monte Carlo simulations by optimizing photon source characteristics
Author(s) -
Bush K,
Popescu T
Publication year - 2005
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.2030986
Subject(s) - linear particle accelerator , monte carlo method , imaging phantom , phase space , quality assurance , range (aeronautics) , computer science , monitor unit , shields , physics , optics , electromagnetic shielding , mathematics , nuclear medicine , engineering , beam (structure) , aerospace engineering , statistics , external quality assessment , operations management , quantum mechanics , thermodynamics , medicine
In conjunction with rapidly expanding clinical Monte Carlo (MC) implementation, MC users are faced with the difficult and time consuming commissioning process of their virtual linac. After accurately configuring the treatment head according to manufacturer specifications, a rigorous and extensive benchmarking process is required to ensure that the MC virtual linac produces beams of essentially the same quality as those of the real treatment unit being modeled. Often, even after systematically varying the input parameters over a suitably chosen range, it is found that the shape of the measured profiles cannot be exactly matched. This limitation is attributed to the lack of accurate knowledge of the geometry and materials of some linac components, especially, the flattening filter. We have developed an automatic optimization method that allows a user with an arbitrary linear accelerator to commission a MC dose calculation engine that accurately reproduces the measured output of the accelerator in water. Using a simulated annealing optimization algorithm our method converges MC dose distributions to experimental measurements by optimizing the weights of particles in a phase space. To achieve this, a phase space is divided up using LATCH variable assignment, each beamlet is transported into a water tank phantom, and the dose deposition is scored separately. Individual beamlet weights are then optimized such that the weighted sum of beamlet dose depositions converge toward our target dose distribution. The resulting beamlet weights are then assigned to all particles in the original phase space where they are incorporated into all future simulations.