Abstract Book

S262

ESTRO 37

vascular damage on tumours. This would produce reduced blood perfusion and indirect tumour cell death due to severe hypoxia, what has been suggested as one of the possible causes for the increased tumour response in hypofractionated SRT. The purpose of this work is to develop a model to calculate the microscopic distribution of oxygen in tumour volumes considering realistic 3D vascular architectures (VA) and to assess the possible role of vascular damage in tumour response. Material and Methods VA are generated in 1 mm 3 subvolumes using the open source Vascusynth tool. A diffusion-consumption equation is numerically solved for pO2 leading to tumour-like oxygen distributions. To assess vascular damage, a serial-parallel model is implemented to deactivate vascular segments considering published endothelial radiosensitivity parameters and the LQ model. This vascular damage produces a decrease of tumour oxygenation. Tumour cell death due to severe hypoxia is then simulated in a conservative way, by killing 5% of cells in the most hypoxic range (pO2<0.25 mmHg). Additionally, enhanced radioresistance for the surviving hypoxic cells is considered using the OER modified LQ model. Several vascular fractions (VF) and dose levels are simulated. The impact of these mechanisms on tumour cell killing is studied for a hypofractionated scheme. Three simulations cases are compared: a) LQ model considering OER; b) LQ model considering OER and vascular damage; and c) LQ model considering OER, vascular damage and indirect death due to severe hypoxia. Results Figure 1 shows an example of vascular damage to a VA with 2.2% VF and the impact on the pO2 distribution. For a simulated 3 fractions (fx) SRT treatment of 7Gy/fx, the number of surviving cells (SC) in case b) is 89% larger than that in case a). On the other hand, the SC decreases in 14% for case c) compared to case a). These survival curves are in accordance with previously proposed explanations for the increased response under SRT.

vessels would help to determine the quality of our model assumptions such as the endothelial cells radiosensitivity and the severe hypoxic cells killing. Other considerations, like the role of vascular network hierarchy on deactivation and time dependent repair mechanisms, are being currently studied OC-0507 The effect of incidental dose on out-of-field failure probability in anal cancer M. Van Herk 1 , H. Sekhar 2 , E. Vasquez Osorio 1 , T. Marchant 3 , R. Kochhar 4 , A. Renehan 2 1 Manchester Cancer Research Centre- Division of Cancer Science- School of Medical Sciences- Faculty of Biology- Medicine and Health- University of Manchester- Manchester Academic Health Sciences Centre, Radiotherapy Related Research, Manchester, United 2 Christie NHS Trust, Surgery, Manchester, United Kingdom 3 Christie NHS Trust, Radiotherapy, Manchester, United Kingdom 4 Christie NHS Trust, Radiology, Manchester, United Kingdom Purpose or Objective Elective field sizes in radiotherapy are mostly defined from clinical experience. In rare cancers, such as anal cancer, statistical knowledge of out-of-field failure rates is limited. We aim to develop a TCP and tumour presence probability model for out-of-field failures, to eventually develop an evidence-based CTV definition. In particular, we aim to evaluate the effect of incidental dose on failure probability. Material and Methods We analysed data from 265 anal cancer patients treated by chemoradiotherapy through a centralised multi- disciplinary team, using MR staging. In follow-up MR scans, 13 out-of-field nodal failures (those at unknown locations) and 18 in-field failures were located in 15 patients. We assume the out-of-field failures are due to occult disease. By registering the follow-up MR to the planning scan, the dose in the centre of failures was evaluated, accounting for treatment interruptions. We recorded these failures on a template patient (Fig. 1). Dose distributions of all 265 patients were mapped to the template patient (accuracy ~6 mm SD), and a population- based estimate of the incidental dose on the failure locations was determined allowing TCP modelling. By combining the TCP model with a simple tumour presence probability model, a failure model was obtained that was fitted to the observed failure pattern.

Figure 1: Vascular architecture with a 2.2% VF before and after a fraction of 10 Gy, inducing a 70% vascular deactivation (black coloured segments), top left and top right respectively. Bottom panels show pO2 distributions corresponding to each case. Hypoxic fraction (% of cells with PO2<5 mmHg) increase in 18% due to radiation induced damage. Conclusion Our model shows that vascular damage may, at least partially, explain the increased radiation response in SRT. A better understanding of the radiobiology of blood