ESTRO 37 Abstract book

S982

ESTRO 37

older inaccurate dose algorithms, requiring an adjustment of the currently used prescription dose. We investigated the consequences of the transition from a type-a dose calculation algorithm (pencil beam, PBC) to more advanced type-b (collapsed cone convolution, CCC) and type-c (MonteCarlo, MC) for SBRT of lung lesions treated with VMAT technique. Material and Methods Twenty-seven lung SBRT patients that underwent SBRT for lung metastases were re-calculated with CCC and MC. This last is the XVMC dose engine implemented in the Monaco TPS (Elekta). Historical PBC plans were optimized using a full 6MV coplanar VMAT arc. The prescribed dose (PD) to the PTV ranged from 20Gy to 30 Gy in a single fraction. A 1mm MLC block margin was adopted in order to increase the dose fall-out, the dose heterogeneity and intensify the dose within the GTV. For each PBC plan, the prescription isodose was selected as the greatest fulfilling the two criteria: 95% of the PTV reached 100% of PD (V100%≥95%) and 99% of the PTV reached ≥90% (V90%≥99%) of PD. The PTV was also separated into components in tissue (GTV) and air (PTVair) to better understand the impact of air amount in PTV on dose distributions. Plans were compared using V100, D95, D98 and Dmean for the PTV and the GTV. For lung, Dmean and various Vx% were calculated. Based on changes in D95 and D98, the prescription dose was converted from PBC to CCC and MC. Results The median PTV size was 26.3 cc (range, 5.6–101.3). Plans calculated with PBC caused large PTV underdosage, mainly at the target periphery. Average V100 significantly decreased from 96.1% to 42.3% and 53.2% for PTV, and from 100% to 83.9% and 92.1% for GTV when PBC plans were re-calculated with CCC and MC (p < 0.01). Similarly, D95 significantly decreased from 101.7% to 84.7% and 87.0% for PTV, and from on 114.2% to 101.2% and 103.6% for GTV. Compared with MC and CCC, PBC overestimated D98% by an average of 15.5% and 18.0% to the PTV, and by an average of 11.1% and 13.6% to the GTV. Compared with MC and CCC, the mean dose differences were 12.9% and 14.5% for PTV, and 7.7% and 9.2% for GTV. Deviations from MC doses were found strongly correlated with the percentage amount of air in PTV. Dmean, V5, V10 and V20 to lung strongly correlated among the three algorithms (R 2 >0.99). With respect to MC, CCC shows deviations <3% for all metrics, with a tendency to an over-correction. Based on the mean reduction in D95 and D98, the PBC dose in our last escalation level (1x30Gy) should be converted to a MC prescription dose equal to 1x25Gy. Conclusion Type-a SBRT lung plans considerably overestimate target coverage for all patients. Our aim was to adjust the MC prescription dose for lung metastases in order to approach our current dose level, because our clinical outcomes provided an high local tumor control and low toxicities with PBC planning. EP-1821 Evaluation of dose calculation accuracy at lung-tissue interface in presence of 0.35T magnetic field D. Cusumano 1 , S. Teodoli 1 , F. Greco 1 , A. Fidanzio 1 , L. Boldrini 2 , M. Massaccesi 2 , F. Cellini 2 , V. Valentini 2 , M. De Spirito 1 , L. Azario 1 1 Fondazione Policlinico Universitario A.Gemelli, Unità Complessa di Fisica Sanitaria, Roma, Italy 2 Fondazione Policlinico Universitario A.Gemelli, Area di Radioterapia Oncologica, Roma, Italy Purpose or Objective In the framework of the adaptive RT, the new MR-guided hybrid machines represent a promising resource, allowing to change the dose distribution in accordance with daily organs position. One of the main concerns linked to this

technology regards the influence of the magnetic field (B) on the dose distribution and the TPS ability in modelling the effects due to the presence of magnetic field. Aim of this study is to investigate the accuracy of the MRIdian Montecarlo algorithm (ViewRay) in calculating the dose distribution in presence of magnetic field at tissue-lung interface, where the electron return effect plays a primary role. Material and Methods The investigation was realised using Gafchromic EBT3 films: a dose calibration step was planned on MRIdian to perform absolute dose measurements. Films were cut into 3x3 cm 2 pieces and arranged in a water slab phantom. The detectors were placed at SAD =105 cm, SSD=100 cm. Thirteen film pieces were exposed perpendicularly to the beam axis at different 60 Co source exposure times and a dose measurement with ion chamber was contextually performed. Films were scanned one day after the irradiation and the red channel net optical densities were calculated and correlated with the measured dose values. Figure 1 shows the experimental setup adopted for the evaluation of the dose distribution at lung-tissue interface. Two EBT3 films were inserted in proximity of the interfaces of a non-homogeneous sandwich phantom composed by a 5 cm lung-equivalent slab (ρ= 0.39 g/cm 3 ) embedded inside two 5 cm water-equivalent slabs. A CT scan of the phantom was acquired and 3 treatment plans were calculated and delivered: • Plan A: one single 4.2x4.2 cm 2 beam delivered at 0 degree • Plan B: three single 6.3x6.3 cm 2 beams delivered at 330, 0 and 30 degree • Plan C: IMRT complex plan constituted by 3 triplets (0,120,240; 40,160,280; 80;200;320) A validation step of the established dosimetric workflow was carried out delivering plan A on a homogeneous phantom. The dose distributions measured by the EBT3 films were then compared to those calculated by the TPS in terms of 1%/1mm, 3%/1mm, 3%/3mm and 25cGy/1m m γ analysis.

Results Table 1 summarises the results obtained for γ value <1 for the different delivered plans.

All the tested plans show a γ -passing rate (% of γ<1 points) higher than 90% for the 3%/3 mm tolerance criteria, which are the values suggested by the ESTRO