In-vivo and pre-treatment quality assurance software validation and verification

Abstract

The aim of this study is to test the accuracy of the dose calculation algorithm available in DosimetryCheck (DC, MathResolution®), a patient QA software for both pretreatment QA and in vivo dose verification, that uses the EPID measured fluence of the treatment fields to reconstruct the dose distribution in the CT planning model of the patient. First tests were performed on the IBA “I’mRT” phantom to assess DC performances in the steps of entrance fluence estimation and dose calculation both for pre-treatment and intreatment measurements. The central insert was filled with three different materials (air, bone and homogeneously with RW3) and different square field sizes (FS) were irradiated to evaluate the isocenter dose in the homogeneous phantom. The global accuracy was assessed by comparing OFs and PDDs evaluated with TPS with that reconstructed by DC. Then, DC was run pre-treatment and in-vivo for 15 patients: 1) 7 IMRT prostate cases, 2) 3 abdominal VMAT cases, 3) 5 head VMAT cases. Gamma analysis (3%, 3mm) was used to compare measured with calculated dose distribution. Isocenter dose was equal within 1.75% for pre-treatment and in –treatment measures and for all FS. Due to the use of a pencil beam algorithm, high OF and PDD differences (up to 64.2% and 110%, respectively) were found in the air filled phantom at small FS. In patients, gamma passing rates evaluated on the whole treatment volume were above 98% and 94% for pre-treatment and in vivo prostate cases respectively; above 92% for both pre-treatment and in vivo abdominal VMAT patients, and above 89% and 73% for head VMAT patients. In conclusion, DC is capable of successfully reconstruct the dose distribution in the patient from the EPID measured exit fluences, even if a more accurate algorithm is needed when low density regions are involved.