ESTRO 2022 - Abstract Book

S1579

Abstract book

ESTRO 2022

harmonized features they were [0.65 − 0.89], [0.83 − 0.96] and [0.73 − 0.95]. For the external validation dataset, the AUC results were: [0.59 − 0.77], [0.70 − 0.85] and [0.40 − 0.75] for harmonized features; while for non-harmonized features they were [0.33 − 0.71], [0.39 − 0.70] and [0.38 − 0.76]. Conclusion In conclusion, our proposed harmonization applied to the quadratic SVM model allows for higher AUC for both the training dataset and the external validation dataset. This result confirms the strength of our harmonization method, highlighted by the smaller difference between training and validation AUC values.

PO-1774 Oxygen distribution in the microenvironment and RT treatment outcome: a modeling study on metastasis

L. Possenti 1,2 , A. Gallo 2 , A. Cicchetti 1 , T. Rancati 1 , P. Zunino 3

1 Fondazione IRCCS Istituto Nazionale dei Tumori, Prostate Cancer Program, Milan, Italy; 2 Politecnico di Milano, LaBS, CMIC Dept., Milan, Italy; 3 Politecnico di Milano, MOX, Department of Mathematics, Milan, Italy Purpose or Objective Oxygen (O2) distribution is a known determinant of radiosensitivity, especially for low-LET ionizing radiation. It is delivered to the tissues (normal and tumoral) by the microvasculature and then depleted by cells. The work aims to evaluate different treatments (in terms of cell surviving fraction (SF) after RT) using a computational model of the microenvironment describing the O2 distribution and radiation damage. We analyze different RT sources and fractionations. We benchmark the model in clinical scenarios of lung metastasis grown in tissues with different vascularisation. Materials and Methods To analyze two different metastasis scenarios, we investigated two possible vascular architectures: a high and low level of vascularisation (HV – network lateral surface 14.4 mm 2 /mm 3 tissue , LV – 3.3 mm 2 /mm 3 tissue ). We computed the O2 distribution using the 3D1D computational model (Ann Biomed Eng 2021), which includes O2 consumption and accounts for the presence of red blood cells. We applied the linear-quadratic model following the Alper-Howard- Flanders formulation, which describes the SF dependence on O2 and LET. We tuned the parameters to reflect the radiosensitivity of lung metastasis (a = 0.08 Gy -1 , a/b= 4 Gy). The model was complemented by the repopulation term, including a kick-off time of 14 days. Three different fractionation schemes were analyzed (1x30 Gy[RBE], 3x10 Gy[RBE] – 1fr/w, 6x5 Gy[RBE] – 2fr/w) considering photons (LET = 2 keV/ µ m), protons (12 keV/ µ m, RBE = 1.1), and carbon ions (Cions, 100 keV/ µ m, RBE = 3). Results We obtained a poorly oxygenated scenario (LV – average pO2 14 mmHg) and an oxygen-rich one (HV - average pO2 48 mmHg). Our model describes the spatial distribution of tissue O2 (Figure 1a) and the 3D local map of the SF in the tumor tissue. We report the LV case irradiated with photons in the 3 x10Gy setting (Figure 1b). The heterogeneity of the SF is evident, with greater damage close to the vessels. After the 3 fractions and considering even repopulation occurring between the 2 nd and the 3 rd one, the difference in SF due to oxygen spans 3 orders of magnitude (9.5e-4 vs 0.19). Conversely, in the HV case, the same treatment scenario efficiently killed the tumor cells (max SF = 4.7e-5, Figure 1c). Comparing all the cases (Figure 2), the high dependence of photons irradiation on O2 is evident, determining different SF distributions. With more O2 available (HV cases) the SF distribution is less heterogeneous (6x5Gy photons – SF in LV: 1e-2 - > 2e-1, SF in HV: 5.2 -> 5.8e-3). The O2 effect is lower when greater LET is involved, such as in protons, but even more in