ESTRO 2020 Abstract Book

S389 ESTRO 2020

regression was used to define the dosimetric parameters associated with the risk of cognitive decline at one year following the completion of chemoradiotherapy. Statistical significance was reached when p-level was below 0.05. Results A cohort of 40 surviving patients having performed the neuropsychological assessment at one year was considered for analyses. Median PTV was 240 cc (89-440). The main PTV locations were the right frontal lobe (27.5%) and the right temporal lobe (22.5%). At least one cognitive function was altered in 34 patients (85%) at baseline and at one year. Median doses to both hippocampi, the left subventricular zone and the left frontal lobe were associated with a significant risk of impairment of the general cognitive assessment, the verbal episodic memory and the processing speed at one year (p<0.05 in the three cases). Median doses of 20 Gy and 25 Gy to the normal brain exposed to a risk of 20% and 40% of alteration of the general cognitive assessment, respectively (p=0.009). Maximal respective doses of 42 Gy and 53 Gy to the left temporal lobe were correlated with probabilities of 10% and 20% of the verbal episodic memory decline (p=0.014). In multivariate analysis, a PTV upper than 320 cc exposed to a significant risk of alteration of the general cognitive assessment at one year (p=0.048). Conclusion This prospective study finds statistically significant correlations observed between dose distributions to various brain areas and late cognitive impairment. These results have to be confirmed in a higher number of patients in order to validate new relevant dose constraints to the healthy tissues implied in cognitive functions. OC-0697 Particle therapy: Assessing clinical benefit of direct stopping-power prediction from dual-energy CT N. Peters 1,2 , P. Wohlfahrt 1,2,3 , C. Hofmann 4 , C. Möhler 4 , S. Makocki 5 , C. Richter 1,2,5,6 1 OncoRay – National Center for Radiation Research in Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus- Technische Universität Dresden- Helmholtz-Zentrum Dresden - Rossendorf, Dresden, Germany ; 2 Helmholtz-Zentrum Dresden - Rossendorf, Institute of Radiooncology - OncoRay, Dresden, Germany ; 3 Now with Massachusetts General Hospital and Harvard Medical School, Department of Radiation Oncology, Boston, USA ; 4 Siemens Healthineers, Forchheim, Germany ; 5 Department of Radiotherapy and Radiation Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus- Technische Universität Dresden, Dresden, Germany ; 6 German Cancer Consortium DKTK- partner site Dresden, Germany and German Cancer Research Center DKFZ, Heidelberg, Germany Purpose or Objective Together with the first clinical implementation of direct stopping-power prediction (DirectSPR) from dual-energy CT (DECT), the resulting range uncertainty was comprehensively quantified and clinically applied. Moreover, the dosimetric impact was evaluated on the first patients treated with this new approach. Material and Methods The implemented DirectSPR approach is characterised by patient-specific determination of tissue parameters and a patient-size-dependent model calibration. Uncertainties Proffered Papers: Proffered papers 36 - Managing uncertainties in proton therapy

in the prediction of the stopping-power-ratio (SPR) with DirectSPR were comprehensively quantified for different tissue groups (soft tissue, bone and lung tissue) and then propagated to an overall range uncertainty depending on the body region. The resulting safety margin was compared to that of the current state-of-the-art heuristic CT-to-SPR conversion (HLUT), which was 3.5%+2mm in our institution. Moreover, the impact of the reassessed range uncertainty on dosimetric treatment outcome was analysed for 10 prostate-cancer as well as 10 brain-tumour patients (CTV in close proximity to the brainstem). For comparison, the clinical treatment plans, using DirectSPR with the reassessed uncertainty, were reoptimised with the HLUT approach and its respective range uncertainty. The two approaches were compared concerning the change in integral normal tissue dose and DVH parameters of relevant organs at risk. In addition, the difference in maximal proton range per field was analysed for the first 70 patients treated with DirectSPR (50 brain-tumour, 20 prostate-cancer patients). Results For DirectSPR, SPR uncertainties (1σ) of 1.1%-1.3%, 1.6% and 1.3% for soft tissue, bone and lung tissue were determined, resulting in a clinical range uncertainty (2σ) of (1.7%+2mm) for the head and (2.0%+2mm) for the pelvic and lung region. This corresponds to a safety margin volume reduction of ≈35%. DirectSPR together with the smaller range uncertainties were clinically implemented in our institution in April 2019. Thereby, integral dose was reduced on average by 4.4% for prostate- and 0.8% for brain-tumour patients based on our dosimetric evaluation (Fig.1). Mean dose to bladder and rectum was decreased by 1.6% on average for prostate-cancer treatments, with bladder ΔV 65Gy/ V 65Gy =3.5% and rectum ΔV 60Gy/ V 60Gy =2.3%. In the brainstem, the mean dose was reduced by 3.3%. For all 70 patients treated so far, the maximum proton range was decreased by 3.7mm and 2.6mm for treatments in the pelvis and head, respectively. This directly corresponds to a better sparing of healthy tissue in beam direction. Conclusion Enabled by the first clinical implementation of DECT-based DirectSPR, range prediction accuracy was substantially increased in routine clinical practice. For the first patients treated, the clinical benefits regarding dosimetric outcome and normal tissue sparing have been demonstrated. The impact on clinical outcome, side effects and tumour control, should be addressed in future studies.