To ensure the required accuracy and prevent from misadministration, cancer treatments, by external radiation therapy are simulated on Treatment Planning System or TPS before radiation delivery in order to ensure that the prescription is achieved both in terms of target volumes coverage and healthy tissues protection. The TPS calculates the patient dose distribution and the treatment time per beam required to deliver the prescribed dose. TPS is a key system in the decision process of treatment by radiation therapy. It is therefore essential that the TPS be subject to a thorough check of its performance (quality control or QC) and in particular its ability to accurately compute dose distributions for patients in all clinical situations that be met. The "traditional" methods recommended to carry out dosimetric CQ of algorithms implemented in the TPS are based on comparisons between dose distributions calculed with the TPS and dose measured in physical test objects (PTO) using the treatment machine. In this thesis we propose to substitute the reference dosimetric measurements performed in OTP by benchmark dose calculations in Digital Test Objects using PENELOPE Monte-Carlo code. This method has three advantages: (i) it allows simulation in situations close to the clinic and often too complex to be experimentally feasible; (ii) due to the digital form of reference data the QC process may be automated; (iii) it allows a comprehensive TPS CQ without hindering the use of an equipment devoted primarily to patients treatments. This new method of CQ has been tested successfully on the Eclipse TPS from Varian Medical Systems Company.
