ESTRO 36 Abstract Book

S434 ESTRO 36 _______________________________________________________________________________________________

for the experimental and calculated FWHM. Mean energies calculated from peripheral photon spectra energies ranged from 0.242 MeV in the mediastinum to 0.171 MeV in the pelvis. Based on these results, an average energy correction factor of 10% was applied to TLD readings. A maximum of 15% difference between calculated and measured peripheral doses were obtained. Conclusion The use of Egspp allowed us to model accurately the Gamma Knife. The level of detail achieved in the modeled geometry is essential for the peripheral dose calculation, since it is dependent on the radiation leakage. The agreement between simulations and measurements is good, with higher discrepancies observed in the points located on the limbs of the phantom. A MC methodology for peripheral dose characterization has been validated and therefore, different scenarios regarding patient size and/or beam geometry can be estimated for future references. References [1] Radiother Oncol 2012;10(5):122-126 [2] JACFMP, vol 14, N2, 2013 [3] Phys. Med. Biol. 57 (2012) 6167–6191 [4] Biomed. Phys. Eng. Express 1 (2015) 045205 PO-0812 Dosimetric impact of using Acuros algorithm for stereotactic lung and spine treatments L. Vieillevigne 1,2 , T. Younes 1,2 , A. Tournier 1 , P. Graff Cailleaud 1 , C. Massabeau 1 , J.M. Bachaud 1 , R. Ferrand 1,2 1 Institut Claudius Regaud- Institut Universitaire du Cancer de Toulouse Oncopole, Radiophysique, Toulouse, France 2 Centre de Recherche et de Cancérologie de Toulouse CRCT- UMR1037 INSERM - Université Toulouse 3, Radiophysique- équipe 15, Toulouse, France Purpose or Objective The main aim was to assess the dosimetric impact of calculating with the Acuros (AXB) algorithm instead of Anisotropic Analytical Algorithm (AAA) for stereotactic (SBRT) lung and spine cancer treatments. Material and Methods Ten stereotactic lung patients and ten stereotactic spine patients were selected to investigate the dosimetric impact of using AXB instead of AAA. Dynamic conformal arc was used for SBRT lung patients with a prescription of 50 to 55 Gy in 3 or 5 fractions to the 80% isodose. For the SBRT spine patients, Rapid Arc plans were prepared and 27 Gy was prescribed in 3 fractions to the PTV median dose. The plans were recalculated with the AXB algorithm by using the same beam settings and monitor units as the AAA. Two dose reporting modes of AXB, dose to medium (Dm) and dose to water (Dw) were studied. Relative dose differences between algorithms were calculated for PTV (D98%, D95%, D50% and D2%) and for organs at risk (D2% and mean dose for the ipsilateral lung, the spinal cord or the cauda equina). Results For the 10 SBRT lung patients, the mean lung density was around 0.18 g/cm 3 which corresponded to a normal lung tissue. The dosimetric impact on PTV dose (D98%, D95%, D50%) using AXB instead of AAA was quite small with a maximum underdosage up to 3.1% for D98%. Larger differences were obtained on D2% with a maximum deviation of 7.79%. The average difference on the mean ipsilateral lung dose was 0.22% and 0.44% for AXBDm and AXBDw, respectively. AXBDm and AXBDw presented similar results. For the 10 SBRT spine patients, large relative dose disagreement of up to -5.02% for the D98% of the PTV was observed with AXBDm. On average, for the D50% of the PTV, AXBDm revealed a relative underdosage of -2.36%, whereas AXBDw lead to a relative overdosage of +1.64%. Concerning the spinal cord or the cauda equina, the average mean dose was reduced up to -6.93% and up to -

Figure

1

Conclusion It is shown that the MAASTRO algorithm implementation for gamma analysis based on absolute dose comparison is reliable and provides very good results for both types of plans tested. When compared with the results obtained with PDIP v10.0.28, the MAASTRO algorithm presents at least as good results for the pre-treatment portal dosimetry as the currently available PDIP, while reporting absolute dose results, making it a viable, and even desirable, alternative. PO-0811 Monte Carlo simulation of peripheral dose for Gamma Knife treatments B. Sanchez-Nieto 1 , E. Doerner 1 , A.M. Cardona 1 , F. Bova 2 1 Pontificia U-dad Catolica de Chile, Insitute of Physics, Santiago, Chile 2 Gamma knife Chile, Radiotherapy, Santiago, Chile Purpose or Objective Induction of second cancers after external beam radiotherapy (RT) is associated to the dose deposited outside the treatment field (Peripheral Dose -PD) [1] . New advances in radiation oncology have increased the survival of patients beyond the period of latency of the occurrence of secondary cancer (> 5 years), so that the estimation of PDs has become particularly relevant. Commercial treatment planning systems present a great uncertainty in the dose calculation outside the treatment field (differences up to 50%) [2] ; therefore, alternative methodologies for estimation of PD to radiosensitive organs are needed. There are previous studies [3,4] applicable to external RT with linear accelerators. However, no such a model exists for Gamma Knife. The aim of this study was to estimate the peripheral dose associated to radiosurgery treatments using Monte Carlo (MC) and experimental measurements with TLDs. Material and Methods A Leksell Gamma Knife 4C radiosurgery equipment was modeled using the set of C++ libraries Egspp, part of the MC platform EGSnrc. The model includes the entire set of 201 Cobalt-60 sources, along with their respective beam channel. Validation was performed by comparing profiles and dose deposited in depth during irradiation of the Lucy QA Phantom with all sources opened. Then, the photon spectrum and absorbed dose were calculated and measured with TLD-100 pairs, for the plan above, at 14 points of a pseudo-anthropomorphic phantom. TLD-100 had calibration factors for 6 MV nominal energy. TLD readings were corrected by an energy response correction factor due to the change in response from the 6 MV calibration beam to the softened spectrum at the measurement points. Results The simulated geometry was tested using a raytracing method, included in Egspp, which allowed visualization of geometrical details to be compared with the available technical drawings. Difference of just 3.5 % was obtained