ESTRO 38 Abstract book

S1049 ESTRO 38

functionality of tissue/structures/organs neighboring the treated region. EP-1928 Radiomic features and local response in Lung Cancer treated with Stereotactic Radiation Therapy S. Chiesa 1 , R. Gatta 2 , A. Martino 1 , A. Piras 1 , D. Cusumano 1 , L. Boldrini 1 , C. Masciocchi 1 , M. Massaccesi 3 , N. Dinapoli 3 , F. Cellini 1 , A. D'aviero 1 , G.C. Mattiucci 1 , G. Mantini 1 , M. Balducci 1 , V. Valentini 1 1 Fondazione Policlinico A. Gemelli IRCCS- Università Cattolica Sacro Cuore, Dipartimento Scienze Radiologiche- Radioterapiche ed Ematologiche- Istituto di Radiologia, Rome, Italy ; 2 Università Cattolica del Sacro Cuore- Roma Italia, Dipartimento Scienze Radiologiche- Radioterapiche ed Ematologiche- Istituto di Radiologia, Rome, Italy ; 3 Fondazione Policlinico A. Gemelli IRCCS, Dipartimento Scienze Radiologiche- Radioterapiche ed Ematologiche- Istituto di Radiologia, Rome, Italy Purpose or Objective Stereotactic ablative radiation therapy (SABR) is a treatment option for early-stage lung cancer. This study investigates the prognostic performance of radiomic features for disease recurrence in early stage non-small cell lung cancer (NSCLC) patients treated with stereotactic ablative radiation therapy (SABR). Material and Methods Patients with early stage NSCLC, treated with SABR were considered for this study. The lung lesions received a total dose of 42 Gy in 3 fractions or 50 Gy in 5 fractions according to the site. A 4D-CT was performed to delineate the internal target volume (ITV) and the radiomic features were extracted from the ITV reported on average intensity projection (AIP), through the use of a dedicated software. The Wilcoxon Mann Whitney test was applied to evaluate the significance of the radiomic features; afterwards a Logistic Regression model was built for each significant feature. We considered as main outcome the response to SABR according to RECIST Criteria. The clinical response was evaluated during the follow up through the CT and PET-CT scans. Results Forthy-two early stage NSCLC patients and 47 lesions, receiving SABR, were considered for this analysis. Thirteen lesions presented a local recurrence. Ninety-four features were extracted using a dedicated software. The correlation between the radiomic features and the local recurrence at 12-24 and 36 months was investigated. Seventeen features showed a significant correlation (p<0.01) with local recurrence at 12 and 24 months at Wilcoxon Mann Whitney test: 3 intensity based statistical features; 2 Grey level co-occurrence based features–Texture features (GLCM); 6 Grey level run length based features–Texture features (GLRLM); 6 Grey level size zone based features–Texture features (GLDZM), according to Image biomarker standardization initiative v.7 ( https://arxiv.org/abs/1612.07003). No correlation was found between radiomic features and local recurrence at 36 months. At the kernel density estimation of Skewness (p < 0.01 at the Mann Whitney test) is shown on the picture on the left for both, the positive (blue) and negative (red) lesions, on Figure 1. In the middle the distribution of positive and negative outcomes (same colors) depending on the value of Skewness and on the right the ROC curve of the associated Logistic Regression model (AUC = 0.847).

suitable to describe the effects of radiation on the fluid balance at the microvascular level. We will discuss different scenarios that illustrate this dependence. Material and Methods The model describes the tumor microenvironment in a three-dimensional setting representing a tissue slab of about 0.5 mm size(Fig.2). The model consists of a system of coupled partial differential equations that are solved numerically, by means of an in house code based on the finite element method. This approach has been already successfully applied to model drug delivery and hyperthermia ( Int J Numer Meth Biomed Engng 2014).

Results We study interplay between radiotherapy and microcirculation in the early phase, namely within 24 hours form treatment. The effect of radiation on microvasculature is primarily represented in the model by means of an increase of vascular permeability and degradation of the mechanical properties of the vascular walls (in terms of thinning and decrease of the Young modulus). We show that the combination of these effects entails a cascade of phenomena involving interstitial hypertension, capillary constriction and finally occlusion of the weakest channels, which correspond to the observed outcomes of radiotherapy, such as decrease of the volumetric density of functional capillaries. More precisely, the fundamental governing equations (Fig.2) are solved with numerical simulations. Possible effects of radiation on the functional parameters are also shown. The exact tuning of the parameters describing damage as a function of dose/dose fractionation (for example the constitutive relation for L_p increasing with increasing RT dose) will be performed by measuring damage to microvessels on microfluidic chips as a function of dose, using a clinical linear accelerator.

Conclusion The results of this study represent a first step towards the challenging objective of understanding, and describing in a mechanistic way, the effect of radiation on the vascular microenvironment. Combining such mechanistic causality laws with patient specific data about the vascular environment, will enable prescription of radiotherapy treatment based on additional/patient specific guidelines that improve the treatment efficacy and preserve the