ESTRO 37 Abstract book

S966

ESTRO 37

need for manual data entry or time consuming export to third party systems. The results show that this method has the precision required to detect gross errors and anatomical changes to help ensure safe and accurate [1] Bojechko et al., A quantification of the effectiveness of EPID dosimetry and software-based plan verification systems in detecting incidents in radiotherapy, Medical Physics 42, 5363 (2015) ; [2] Piermattei et al., Application of a practical method for the isocenter point in vivo dosimetry by a transit signal, Phys. Med. Biol. 52, 16, 5101, (2007). EP-1798 Local confidence limits in VMAT pre- treatment QA with COMPASS based on AAPM-TG119 and DVH analysis M. Sutto 1 , A.F. Monti 2 , C. Carbonini 2 , D. Canonico 3 , G. Rinaldin 3 , L. Bindoni 3 , L. Begnozzi 3 , A. Torresin 2 1 Università degli Studi di Milano, Scuola di Specializzazione in Fisica Medica, Milano MI, Italy 2 ASST Grande Ospedale Metropolitano Niguarda, Medical Physics Dept., Milano MI, Italy 3 AULSS2 Marca Trevigiana - Ospedale di Treviso, Medical Physics Dept., Treviso TV, Italy Purpose or Objective The aim of this study was to establish local confidence limits (CL) for pre-treatment QA of VMAT with COMPASS, using AAPM-TG119 tests on a cylindrical phantom. CL were then reviewed by analysing pre-treatment QA of clinical plans. Material and Methods COMPASS QA system, with the MatriXX ionization chamber detector array (IBA Dosimetry) was implemented for pre- treatment QA of VMAT plans optimized with Monaco TPS and delivered by an Elekta Synergy LINAC. Open fields on homogeneous (RW3) and inhomogeneous (Gammex) phantoms were tested to verify the discrepancies of beam modelling and the differences between the dose calculation algorithms of COMPASS and Monaco (Figure 1). Then, as proposed by AAPM TG119, local CL on the 3%,3mm gamma passing rate (γ PR ) for VMAT QA were established using TG119 test plans implemented on a cylindrical phantom, as previously done for VMAT QA on a Delta4 phantom. Mean dose (D mean ) to the PTVs and to the OARs were also compared, and local CL on the dose differences were set. The TPS dose distribution was compared with the dose distribution calculated in COMPASS on the patient/phantom anatomy from the RT- PLAN (CC) and with the dose reconstructed from measurements performed with MatriXX (CR). CC was also compared with CR (Figure 2). CL were finally reviewed by analysing QA results of 41 H&N and 10 prostate VMAT plans. treatment. References

Purpose or Objective In-vivo transit dosimetry (IVD) can be a useful tool as a final safety check of the dose delivered to a patient [1]. A commercial system for EPID IVD has been in clinical use at the centre since 2013 for radical treatments. This commercial system requires a full 3D treatment plan therefore it cannot be used on plans without a volumetric dose calculation (usually palliative treatments). For these plans, an in-house system was developed to complement the use of the commercial software. The initial implementation as a spreadsheet required time consuming manual data entry with risk of errors. This work involved employing the Varian scripting interfaces to develop an improved version of the in-house system, which is fully integrated within the ARIA environment, removing the need for manual data entry or any export of data to an external system, and performs analysis automatically. Material and Methods An in-house IVD system has been implemented clinically as a plugin script for Varian Eclipse and Portal Dosimetry software, using the C# application programming interfaces (APIs) provided by Varian. The system calculates the delivered dose at a reference point using a measurement from an integrated EPID image. The calculation is based on a table of transmitted intensity per unit midplane dose for a range of water phantom separations and field sizes, similar to the approach described by Piermattei et al [2]. The concept is illustrated in Figure 1. Clinical IVD is routinely performed on all patients where technically possible by the treating radiographers.

Results Testing on solid water and anthropomorphic phantoms showed the in-house system to have a measurement accuracy of -0.4%±2.5%, compared to -2.3%±4.1% for the commercial system (1SD, comparison with TPS (Varian Eclipse AAA)). Clinical results categorised by treatment site are shown in Table 1. The new integrated system saves radiographer time in analysing IVD results, reducing the time taken to approximately 5-10 minutes per patient to less than 1 minute. Measurement corrections that were previously applied manually are now automated, also reducing the frequency of and time taken for physicists’ investigation into out-of-tolerance results.

Conclusion An in-house IVD system has been implemented which allows EPID transit dosimetry to be fully integrated within the ARIA Oncology Information System, removing the