Study On Independent Dose Verification Method And System Of Proton Therapy Based On Monte Carlo Algorithm

Proton therapy is an advanced technique compared to conventional photon radiotherapy because of its superior physical and biological characteristics.However,due to the sharp distal dose fall off near the Bragg peak,proton therapy is very sensitive to the possible uncertainties,especially associated with dose calculation.Analytical pencil beam algorithms are now widely used in commercial treatment planning system(TPS),although the analytical algorithms are known to suffer from dosimetric errors when involveing inhomogeneous medium compromising the precision of proton therapy.Monte Carlo method is considered as the "golden standard" in dose calculation because it is more accurate than the analytical algorithms.But the accuracy of Monte Carlo is dependent on the system modelling and commissioning,a task known to be too complex for routinely clinical practice.The purpose of this thesis is to explore and investigate the implementations of Monte Carlo method in dose verification in proton therapy clinical practice.The specific aim is to develop an independent dose verification tool for proton therapy based on the TOPAS Monte Carlo model,which can then be used for clinical patient specific quality assurance(PSQA)and three dimensions(3D)dose verification for treatment plans.To achieve this goal,4 tasks will be accomplished:(1)To build the Monte Carlo model in TOPAS based on the proton treatment system at Univerisy of Florida Health Proton Therapy Institute(UFHPTI),including both the passively double scattering(DS)and pencil beam scanning(PBS)modality.(2)To validate and commission the model to develop the specific parameter library for this model.(3)To implement this model in PSQA,3D dose verification and LET calculation for patient treatment cases with PBS.(4)To develop an interface between TOPAS Monte Carlo model and clinical TPS,so that TOPAS input files can be created automatically using the in-house scripts.In this dissertation,a thourough description of the modelling of UFHPTI's system,including the DS and PBS mode,was presented.For commissioning purpose,verificaitons including proton pristine Bragg peaks,integrated depth dose,in-air spot profiles,absolute dose calibrations,spread out Bragg peaks(SOBPs)and lateral profiles were performed by comparing with the measured and TPS calculated data.The results showed that TOPAS calculated depth and lateral dose distributions in DS model agreed well with measurements,the maximum difference was under 5%,most of the regions agreed within 2%.Due to the inherent inability,TPS(Eclipse)analytical algorithm predicted the SOBP and lateral profiles with a maximum dose difference up to 10%.If highly inhomogeneous materials were encountered in the beam path,TPS can overestimate the dose up to 25%.For PBS mode,TOPAS calculated depth and lateral doses agreed within 3%with measured data and that by TPS(RayStation)simplified Monte Carlo dose engine.In some cases,TPS analytical algorithm can overestimate the dose in SOBP and lateral profile by as much as 8%.For the 2D planar dose calculation for the 8 PSQA cases including 21 fields,TOPAS calculated dose distributions also agreed well with measurements and TPS simplified Monte Carlo dose engine,with the Gamma passing rates(criteria of 3mm/3%)being higher than 98%,for most fields.For 3D dose calculation in patient treatment plan verification,the Gamma passing rates with criteria of 3mm/3%for TOPAS calculation and TPS simplified Monte Carlo dose engine was found to be 97.1%.TOPAS calculated LET distributions showed a high LET region outside and behind the PTV,with maximum value up to 14keV/?m.Finally this commissioned model has been incorporated into the interface with two TPSs(RayStation and DeepPlan),in order to automatically generate the specific TOPAS input files from TPS.To summarize,in this work,we have successfully developed a method of accurate Monte Carlo modelling based on UFHPTI's proton therpy system,and this method can be used as an independent dose verification tool for clinical practice.Interfaces between the commissioned Monte Carlo model and TPS have been successfully developed to facilitate the creation of TOPAS input files automatically,thus enhancing the clinical applicability of the Monte Carlo model developed in this project.