ESTRO 38 Abstract book

S160 ESTRO 38

Unit, Pavia, Italy; 6 National Centre of Oncological Hadrontherapy CNAO, Diagnostic Imaging Unit, Pavia, Italy Purpose or Objective To estimate a tumour control probability (TCP) model from a group of patients affected by skull-base chordoma and treated with C-ions radiotherapy (CIRT), integrating information from diffusion-weighted (DW-) MRI. Material and Methods From 2013 to 2016, 59 patients were enrolled for CIRT (70.4 GyE total prescribed dose) and 20 of them underwent a pre-treatment harmonized imaging protocol, including DW-MRI scan (b-value=50,400,1000 s/mm 2 ). Local control (LC) was clinically assessed at a median follow-up time of 45.8 months as a binary variable, which was found to relate to D 98 (the maximum dose received by at least 98% of the volume) in the target. Survival fraction was defined by the linear-quadratic (LQ) formalism and the probability of killing was assumed to be Poissonian, thus following a classical analytical formulation of the TCP. The linear LQ parameter, α, was obtained by setting the population TCP to 79%, which corresponds to the LC for all the chordoma patients treated at the facility, and by fitting an average TCP based on D 98 ; β was set to 0 to account for the in-vitro experiments. Apparent diffusion coefficient (ADC) maps were computed from DW-MRI and converted into cellular density (cells/cm 3 ) by using a relationship found in a histology-based study [1]. Therefore, two types of TCP, one based on ADC (TCP ADC ) and one based on parameters from the literature (TCP LIT ), were computed for each patient and their diagnostic performance with respect to LC was evaluated through a ROC curve analysis. Results TCP ADC and TCP LIT were found to agree with the relation between LC and D 98 , showing a good fitting performance (R 2 =0.981 and R 2 =0.976), and the first being more conservative in estimating patients’ TCP values (Fig.1). ROC curves (Fig.2) identified the same sensitivity (0.867) and specificity (0.600) and the same optimal threshold, at which the pointwise confidence bounds at 95% confidence were CI 95% =[0.539 1] and CI 95% =[0.625 1] for TCP ADC and TCP LIT , respectively. This allows to state the significant difference from the random performance but not between the two models.

Conclusion Two TCP models were obtained for a homogeneous group of patients, as for dose prescription, and they were both found to agree with clinically assessed LC. In particular, DW-MRI has confirmed to be a promising tool to predict the outcome in particle therapy, when coupled to dose information. Nevertheless, a more robust relationship between ADC and tissue microstructure is needed before DW-MRI usefulness can be established. Bibliography : [1] D.T. Ginat, R. Mangla, G. Yeaney, M. Johnson, and S. Ekholm, “Diffusion-weighted imaging for differentiating benign from malignant skull lesions and correlation with cell density,” Am. J. Roentgenol. , vol. 198, no. 6, pp. 597– 601, 2012. PV-0312 Distributed learning in radiomics to predict overall survival in head and neck cancer M. Bogowicz 1,2 , A. Jochems 2 , S.H. Huang 3 , B. Chan 3 , J.N. Brakenhoff 6 , I. Nauta 6 , S.E. Gazzani 7 , G. Calareso 8 , K. Scheckenbach 9 , F. Hoebers 10 , S. Barakat 2 , S. Keek 2 , S. Sanduleanu 2 , M. Vergeer 11 , R.C. Leemans 6 , C.H. Terhaard 12 , M.W. Van den Brekel 13 , M. Guckenberger 1 , P. Lambin 2 1 University Hospital Zurich and University of Zurich, Radiation Oncology, Zurich, Switzerland ; 2 GROW – School for Oncology and Developmental Biology- Maastricht University Medical Centre-, The D Lab: Decision Support for Precision Medicine-, Maastricht, The Netherlands ; 3 Princess Margaret Cancer Center- University of Toronto, Department of Radiation Oncology, Toronto- Ontario, Canada ; 4 Kantonsspital Aarau, Center for Radiation Oncology- KSA-KSB-, Aarau, Switzerland ; 5 Cantonal Hospital Lucerne, Institute for Radiation Oncology, Lucerne, Switzerland ; 6 VU University Medical Center, Department of Otolaryngology/Head and Neck Surgery, Amsterdam, The Netherlands ; 7 Parma University Hospital, Radiology department, Parma, Italy ; 8 IRCCS Fondazione Istituto Nazionale dei Tumori, Radiology Department, Milan, Italy ; 9 University Hospital Duesseldorf- Heinrich Heine University, Department of Otorhinolaryngology & Head/Neck Surgery, Duesseldorf, Germany ; 10 GROW- School for Oncology and Developmental Biology- Maastricht University Medical Centre, Department of Radiation Oncology, Maastricht, The Netherlands ; 11 VU University Medical Center, Department of Radiation Waldron 3 , B. O'Sullivan 3 , S. Tanadini-Lang 1 , O. Riesterer 1,4 , G. Studer 1,5 , J. Unkelbach 1 , R.H.