One of the main rationales of proton therapy is the lower exit dose, resulting in reduced dose to healthy tissue distal to the treatment volume and increased organ sparing. However, high energy protons produce neutrons in the treatment head and in the patient that can distribute dose well outside the treatment field. The variable biological effectiveness of neutrons along with the current inability of treatment planning systems (TPS) to account for out-of-field dose means that there is a gap in knowledge and documentation regarding the potential risk of developing a second cancer for patients undergoing proton therapy.Computational phantoms provide one way of calculating the dose within a patient for a given treatment plan. A new polygon mesh computational phantom was here adopted for scaling to match individual patient measurements and treatment positions. By combining computed tomography data representing the in-field and a customized computational phantom derived from the mesh phantom covering the out-of-field, we aim to improve estimates of neutron dose in far field organs of interest accumulated during proton therapy. This computational framework has been integrated in Geant4 for coupling to an in-house TPS research engine. The Monte Carlo-based dose estimation in a more realistic whole-body patient anatomy will eventually be coupled to various risk models to estimate the risk of secondary cancer due to the therapy, and this in turn will be used to further optimize treatment planning. Initial results of the project, including validation of the proposed methodology, will be presented.

Abstract: Background Quantitative image analysis based on radiomic feature extraction is an emerging field for survival prediction in oncological patients. 18F-Fluorethyltyrosine positron emission tomography (18F-FET PET) provides important diagnostic and grading information for brain tumors, but data on its use in survival prediction is scarce. In this study, we aim at investigating survival prediction based on multiple radiomic features in glioblastoma patients undergoing radio(chemo)therapy. Methods A dataset of 37 patients with glioblastoma (WHO grade 4) receiving radio(chemo)therapy was analyzed. Radiomic features were extracted from pre-treatment 18F-FET PET images, following intensity rebinning with a fixed bin width. Principal component analysis (PCA) was applied for variable selection, aiming at the identification of the most relevant features in survival prediction. Random forest classification and prediction algorithms were optimized on an initial set of 25 patients. Testing of the implemented algorithms was carried out in different scenarios, which included additional 12 patients whose images were acquired with a different scanner to check the reproducibility in prediction results. Results First order intensity variations and shape features were predominant in the selection of most important radiomic signatures for survival prediction in the available dataset. The major axis length of the 18F-FET-PET volume at tumor to background ratio (TBR) 1.4 and 1.6 correlated significantly with reduced probability of survival. Additional radiomic features were identified as potential survival predictors in the PTV region, showing 76% accuracy in independent testing for both classification and regression. Conclusions 18F-FET PET prior to radiation provides relevant information for survival prediction in glioblastoma patients. Based on our preliminary analysis, radiomic features in the PTV can be considered a robust dataset for survival prediction

Dual-Energy computed tomography (DECT) is expected to allow for more accurate proton therapy treatment planning by improving the estimation of relative stopping power (SPR) and other tissue properties.In this study, we investigated the accuracy of SPR prediction and tissue segmentation based on dual- and Single-Energy CT (SECT). For this purpose, fresh animal tissue samples were irradiated in a clinical proton therapy facility and high spatial-resolution three-dimensional proton dose distributions were obtained using dosimetric polymer gel downstream to the samples. The accuracy of this setup was benchmarked against depth-dose measurements obtained with an ionization chamber behind an adjustable water column (peakfinder, PTW, Germany). The predicted SPR values showed good consistency for both methods with deviations below 1%.DECT (90/150 kVp) and SECT (120kVp) images were acquired and converted to SPR and tissue compositions as input for MC- simulations. DECT-to-SPR conversion yielded mean errors of 0.5%, outperforming the SECT calibration with 1.1% deviation from peakfinder results.A detailed comparison of measured dose distributions to MC-simulations for highly inhomogeneous samples will be presented.In summary, dosimetric gel was used to obtain 3D high-resolution proton ranges and compared to peakfinder measurements and MC-simulations, in order to quantify the accuracy of DECT and SECT based tissue characterization.