|
In this paper, we deepen the R&D program named DTO-DC (Digital Object Test and Dosimetric Console),
which goal is to develop an efficient, accurate and full method to achieve dosimetric quality control (QC) of radiotherapy
treatment planning system (TPS). This method is mainly based on Digital Test Objects (DTOs) and on Monte Carlo
(MC) simulation using the PENELOPE code [1]. These benchmark simulations can advantageously replace experimental
measures typically used as reference for comparison with TPS calculated dose. Indeed, the MC simulations rather than
dosimetric measurements allow contemplating QC without tying treatment devices and offer in many situations (i.p.
heterogeneous medium, lack of scattering volume...) better accuracy compared to dose measurements with classical
dosimetry equipment of a radiation therapy department. Furthermore using MC simulations and DTOs, i.e. a totally
numerical QC tools, will also simplify QC implementation, and enable process automation; this allows radiotherapy
centers to have a more complete and thorough QC. The program DTO-DC was established primarily on ELEKTA
accelerator (photons mode) using non-anatomical DTOs [2]. Today our aim is to complete and apply this program on
VARIAN accelerator (photons and electrons mode) using anatomical DTOs.
First, we developed, modeled and created three anatomical DTOs in DICOM format: 'Head and Neck', Thorax
and Pelvis. We parallelized the PENELOPE code using MPI libraries to accelerate their calculation, we have modeled in
PENELOPE geometry Clinac head of Varian Clinac 2100CD (photons mode). Then, to implement this method, we
calculated the dose distributions in Pelvis DTO using PENELOPE and ECLIPSE TPS. Finally we compared simulated
and calculated dose distributions employing the relative difference proposed by Venselaar [3].
The results of this work demonstrate the feasibility of this method that provides a more accurate and easily achievable
QC. Nonetheless, this method, implemented on ECLIPSE TPS version 8.6.15, has revealed large discrepancies (11%)
between Monte Carlo simulations and the AAA algorithm calculations especially in equivalent air and equivalent bone
areas. Our work will be completed by dose measurement (with film) in the presence of heterogeneous environment to
validate MC simulations.
|
