Validation of two different Monte Carlo engines for secondary independent dose calculation.
PO-1397
Abstract
Validation of two different Monte Carlo engines for secondary independent dose calculation.
Authors: Hofer|, Stefan(1)*[stefan.hofer@sabes.it];Ferrari|, Paolo(1);Maffei|, Martin(2);Haller|, Markus(1);
(1)Südtiroler Sanitätsbetrieb, Dienst für medizinische Strahlenphysik, Bozen, Italy;(2)Südtiroler Sanitätsbetrieb, Dienst für onkologische Strahlentherapie, Bozen, Italy;
Show Affiliations
Hide Affiliations
Purpose or Objective
Secondary independent dose calculations ( SIDC ) have the potential to identify several problems (heterogeneity calculations, data corruptions, system failures) with the primary TPS, which generally are not identificable with a measurement based approach.
The aim of this work is to test two Monte Carlo systems for independent dose calculation in terms of accuracy and to confront them with the Raystation (v 7.0, Raysearch Laboratories) treatment planning system with collapsed cone algorithm.
The first Monte Carlo engine (Monaco v 5.11.02 , Elekta Oncology System) is a complete Planning system, which can also be used for SIDC. The second ProSoma Core (v4,1 , MedCom) is only for SIDC and has no accreditation as a TPS. The ProSoma Core Monte Carlo engine has on average, with the same parameters and on similar machines, a ten times shorter calculation time respect to the Monaco TPS.
Material and Methods
The three dose engines were commissioned for the same Elekta Linac with Agility MLC and a 6MV Photon beam.
Twelve Raystation Vmat plans, with two arcs each, for head and neck and pelvic cancer were evaluated.
The plans were recalculated on the PTW Octavius 4D QA phantom with a uncertainty of 0.5% and a voxel size of 2mm.
Measurements were done with the Octavius 4D phantom and the Octavius 1500 measurement array.
Furthermore the plans were calculated with all three dose engines on patients anatomy and evaluated with Gamma Analysis.
Results
With a global gamma of 2%/2mm the three planning systems showed a mean agreement of 95.1%+-1.5% (Raystation), 96.1%+-1.8% (Prosoma) and 96.4%+-1.5%. (Monaco) with measurement. The Prosoma and Monaco Monte Carlo algorithms showed a significantly better agreement (paired bilateral T test - p < 0.05) with measurement than the Raystation calculations.
Between the two Monte Carlo engines the difference was not significant.
Regarding the dose to the isocenter the mean deviation between calculation and measurement is the following: -0.8% +-1.7% (Raystation), -0.7% +-1.3% (Prosoma), -1.5% +-1.3% (Monaco).
The calculation on the patients CT showed no significant difference using a global gamma of 2%/2mm ( mean Gamma value Raystation-Monaco: 97.3%+-2.4% Raystation-Prosoma: 97.1%+-1.2% and Monaco-Prosoma: 98.2%+-1.9% ).
Using a local gamma of 2%/2mm, the agreement between Raystation and Monaco is significantly better than between Raystation and Prosoma.
Conclusion
In the homogenous medium of the phantom the Monte Carlo engines showed a significantly better agreement with measurement. The two Monte Carlo systems were equivalent in terms of accuracy. The behavior of the algorithms in the patient anatomy i.e.inhomogenous media has to be subject of further investigations since there were no reference measurements available. The calculations among themselves showed a better agreement between Raystation-Monaco than between Raystation-Prosoma.