ESTRO 38 Abstract book

S1154 ESTRO 38

dominated by 10%-20% dose areas, where passing rates were lower. Mean setup error was 0.21, 0.27 and 0.39 cm in LAT, INF- SUP and AP directions, averaged over all isocenters. First fraction delivery took 112 minutes, while remaining sessions had a duration between 90-100 minutes. Conclusion Proposed TMI method allows achieve planning, delivery and patient setup accuracy similar to conventional radiotherapy, even managing large volumes where two CT scans in different patient orientation are needed. OARs doses values are lower than in conventional TBI as was demonstrated by previous studies. Showed case values are slightly higher than in those studies, probably due to higher PTV volume. EP-2090 Helical tomography radiation therapy for multiple brain lesions: in-phantom accuracy assessment M. Zani 1 , L. Marrazzo 1 , S. Calusi 1 , C. Talamonti 1 , S. Scoccianti 2 , D. Greto 2 , L. Livi 2 , S. Pallotta 1 1 University of Florence, Medical Physics Unit- Azienda Ospedaliera Universitaria Careggi, Firenze, Italy ; 2 University of Florence, Radiotherapy Unit- Azienda Ospedaliera Universitaria Careggi, Firenze, Italy Purpose or Objective This study aimed to demonstrate the possibility of safely using helical tomography radiation therapy (HT) for the treatment of multiple brain metastases. Imaging and delivery accuracy of HT systems were investigated by performing End to End (E2E) tests on two different phantoms. Material and Methods A two step analysis was followed in order to assess the reliability of the HT for the treatment of multiple brain metastasis. At first the imaging accuracy was tested. In HT systems target localization is provided by registering MegaVoltage CT (MVCT) and CT images and its precision depends on MVCT acquisition protocols as well as on image registration algorithms. These two aspects were investigated, for different registration options and MVCT slice thickness, by applying known misalignments to an ad- hoc developed phantom. The rotation detection of the imaging technique was also studied. E2E tests were then performed to assess the delivery accuracy in phantoms containing multiple targets by using radiochromic films. On the coronal plane 6 target were simulated on a phantom which consists of a PMMA plate 1cm thick with eight embedded glass spheres (GS-phantom) (fig.1a). On the axial plane 3 thin plastic sheets were cut thus to create 5 approximately circular holes, and then they were inserted and fixed between adjacent slices of an Alderson Rando phantom head (fig.1b): 5 targets in total, with the longest axis ranging from 4mm to 8mm (volumes ranging from 0.47cc to 1.79cc). Measured dose distribution centroids were compared with physical and calculated target positions. Moreover, along the axial plane, planned and measured dose maps were compared in terms of absolute values. Results First Step: The algorithm focusing on bone and tissues with the fine MVCT reconstruction grid gave the best results among the automatic options. The most accurate registration modality resulted to be the manual one with a sub-voxel accuracy shifts and a capability in the detection of rotations within 0.3º. Second Step: In fig.2a the results for the centroid shifts along the coronal plane (6 targets) are shown: the mean deviation between measured dose distribution centroids and physical barycenters is of 0.6 mm (range 0.1 mm-1.3 mm). Along the axial plane (5 targets), a mean deviation of 1.2 mm (range 0.7 mm-2.1 mm) was found for the centroid shifts. In fig.2b one dose profile is reported, corresponding to the dotted line in the EBT3 radiation

Patient’s anatomy was splited in 7 sectors: head, thorax, abdomen, pelvis and femoral using HFS patient orientation, and knee and tibial in FFS. One isocenter were used in each anatomic region except in abdomen and pelvis, where two coplanar isocenter were settled with a lateral shift between them. Two VMAT arcs were used in each anatomic area, except in tibial region where AP/PA field in field segmentation was performed. Rigid and deformable registration between HFS and FFS scans were performed to map doses and calculate/optimize total dose in overlapping region of two CT studies. QA was performed using with Octavius 4D system using 3%- 2.5mm 3D gamma index analysis. Patient setup was assured using one CBCT per isocenter with a 20 cm FOV in head-feet direction. Results Total body volume of 55.8 litres and 43.8 cm body radius. Total PTV volume was 12.9 litres. This gives a ratio PTV/Body of 0.23. Plan was normalized to 95% of PTV be covered by 12 Gy. PTV Dmean, D98% and D2% were 12.8 Gy, 11.6 Gy and 14.1 Gy respectively. Table 1 show dose statistics of main OARs.

Figure 1 shows dose distribution in coronal and sagittal planes of HFS, FFS and composite plans.

Total gamma <1 passing rates are 92.4%, 97.4%, 96.3%, 95.7%, 93.5% and 96.8% in head, thoracic, abdominal, pelvic, femoral and knees areas, respectively. In 3D gamma analysis was observed that total results were