ESTRO 37 Abstract book

S1089

ESTRO 37

Material and Methods To understand the underlying physical mechanisms of radiosensitization and DNA damage induction at the cellular level, a computational cell model with detailed nuclear DNA structure was implemented in the Geant4 Monte Carlo simulation toolkit. A realistic gold nanoparticle distribution was incorporated, allowing for the formation of clusters of vesicles filled with gold nanoparticles. A clinically relevant gold concentration (0.7% wt gold) for the gold nanoparticle size of 6, 15, and 30 nm was considered. Protons with linear energy transfer values found in a spread out Bragg peak (1.3- 26.9 keV/µm) were simulated. The event-by-event models available through the Geant4-DNA project were used for accurate calculation of the DNA damage. To quantify the physical contribution to the DNA damage the formation of single (SSB) and double strand breaks (DSB), and break complexity were scored. The locality of the effect, i.e. the existence of higher damage at a location close to the gold distribution, was also addressed by investigating the DNA damage at a chromosomal territory. For a dose fraction of 2 Gy, each scenario was repeated 1000 times to get an average number of SSB and DSB numbers. Results Our model was able to produce SSB and DSB yields similar to the literature. However, for the combinations of gold nanoparticle size and proton energies studied in the present work, no significant increase in the SSB and DSB formation was observed. Conclusion As gold nanoparticles enhanced proton therapy have been proven experimentally, our results allow hypothesizing contribution from alternative mechanisms of radiosensiti- zation. EP-1998 In silico modelling of the impact of the fractionation for hypofractionated prostate treatments G. Delpon 1,2 , J. N'Guessan 1 , P. Paul-Gilloteaux 3,3 , K. Clément-Colmou 1,2 , V. Potiron 1,2 , S. Supiot 1,2 , F. Paris 1,2 , S. Chiavassa 1,2 1 INSERM, CRCINA U1232, Nantes, France 2 Institut de Cancérologie de l'Ouest, Radiotherapy and Medical Physics, Saint-Herblain, France 3 CNRS UMS 3556- INSERM U16, SFR Santé François Bonamy- Micropicell, Nantes, France Purpose or Objective Due to the improvement of imaging modalities and treatment devices, the concept of hypofractionation is gaining momentum in radiation oncology centres. However the fractionation in terms of total dose and number of fractions (fx) is usually determined according to the linear-quadratic (LQ) model. To derive the hypofractionated schema, the LQ model is used without maintaining the overall duration of the course. In our study, we have used our cellular automata model of tumour response after radiotherapy to study the impact of the number of fractions per week (fx/w) for the different scheme used in international clinical trials. Material and Methods Clinical results published for different moderate or severe hypofractionated prostate international trials were used. For each trial, the automata model was tuned in order to get the Tumor Control Probability (TCP) obtained for the reference arm. a/b was set equal to 3.33Gy. Tuned parameter was the number of cells allowed to divide every day. This methodology allowed to take into account the proliferation rate due to the cancer stage and the associated hormonotherapy when prescribed. Then simulations were run to calculate the TCP for the study arm of each trial according to the number of fx/w, from 2 to 5. For each fractionation, 400 simulations were repeated 5 times to derive an average

TCP value. The impact of the number of fx/w was evaluated in terms of computed TCP variations. Results The dose per fx varied from 1.8Gy to 3.1Gy and 2.5Gy to 7.5Gy respectively in the reference and the study arms. Treatment duration varied from 4 weeks (20x3.1Gy) to 7.8 weeks (39x2Gy) and from 2 weeks (5x7.5Gy) to 5.6 weeks (28x2.5Gy) respectively in the reference and the study arms. In order to obtain published TCP values for the reference arms, the number of cells able to divide ranged from 11% to 19%. For moderate hypofractionated scheme, with a dose per fx up to 3Gy, 5 fx/w were prescribed. With 4 and 3 fx/w, TCP would decrease respectively by 2% and 16%. 2 fx/w would lead to an unacceptable TCP of less than few %. For 3.4Gy per fx, 3 fx/w were prescribed in clinic. Our simulations showed that an increase to 4 or 5 fx would have slightly improved the TCP (2%). However, only 2 fx/w would have dropped by 45% the TCP. Finally, for severe hypofractionation (7.5Gy per fx), 2 fx/w were prescribed. More fx/w would have increased TCP by 4% to 5% but would have probably cause more toxicities. However, the impact of fx/w is clearly less important for such a high dose per fx. Conclusion The impact of the number of fx/w has been investigated. For moderate hypofractionated treatments, the number of fx/w should be at least 4 whereas it could be 2 for very high doses per fx. However those scheme shorten the overall duration of the course by 2 to 4 weeks which is not consistant with the LQ model. More radiobiological investigations are required to define optimized treatment schedules, including studies regarding toxicities. EP-1999 Linear energy transfer and related biological doses in focal prostate boosting with proton therapy P. Bræmer-Jensen 1 , L.P. Muren 2 , J. Pedersen 2 , A.G. Andersen 2 , J.B.B. Petersen 2 , J. Rørvik 3 1 Aarhus University, Physics and Astronomy, Aarhus, Denmark 2 Aarhus University Hospital, Medical Physics, Aarhus, Denmark 3 University of Bergen, Clinical Medicine, Bergen, Norway Purpose or Objective Focal tumour boosting strategies are being explored in radiotherapy of prostate cancer, with the aim of improving local control rates without increasing the risk of normal tissue morbidity. The dose shaping potential of spot scanning proton therapy could be beneficial for focal boosting, but there is emerging evidence of a spatially varying proton relative biological effectiveness (RBE), depending on the linear energy transfer (LET). In particular there is concern of an elevated LET at the distal end of the proton beams, often found near organs at risk. The aim of this study was therefore to investigate dose-averaged LET (LET d ) distributions for focal prostate boosting using spot scanning proton therapy and to derive RBE-weighted dose distributions using LET d -based Spot scanning proton plans were created for six prostate cancer patients in the treatment planning system, PyTRiP. All plans used the conventional two opposing lateral beam configuration (90◦/270◦ gantry angles), with a prescribed dose of 78 Gy(RBE 1.1 ) to the prostate and a total integrated dose of 95 Gy(RBE 1.1 ) to the index prostate tumour. The physical dose and LET d distributions were calculated and subsequently used to derive RBE- weighted dose distributions with three variable RBE models proposed by McNamara (MN), Wedenberg (WB) and Carabe (CB) as well as with the generic RBE=1.1. Results The spot scanning proton plans were highly conformal for all six patients, also for the focal boost (Fig. 1a). The LET d increased towards the end of the trajectory of each variable RBE models. Material and Methods