ESTRO 38 Abstract book

S255 ESTRO 38

in Head and Neck Cancer, Munich, Germany ; 17 Ludwig- Maximilians-Universität, Department of Radiation Oncology, Munich, Germany ; 18 Technische Universität München, Department of Radiation Oncology, Munich, Germany ; 19 Helmholtz Zentrum München, Institute of Innovative Radiotherapy iRT, Oberschleißheim, Germany ; 20 German Cancer Research Center DKFZ, Heidelberg and German Cancer Consortium DKTK partner site, Tübingen, Germany ; 21 Faculty of Medicine and University Hospital Tübingen- Eberhard Karls Universität Tübingen, Department of Radiation Oncology, Tübingen, Germany ; 22 Helmholtz-Zentrum Dresden - Rossendorf, Institute of Radiooncology – OncoRay, Dresden, Germany ; 23 German Cancer Research Center DKFZ, Scientific Director, Heidelberg, Germany Purpose or Objective In order to improve radiotherapy outcomes, further treatment personalisation is considered beneficial. Radiomics analyses aim to predict treatment outcomes based on medical imaging data. Commonly, hand-crafted imaging features are used that require domain knowledge and further feature selection steps. This may cause relevant information to be lost. Deep convolutional neural networks (CNNs) on the other hand can act as automatic feature detectors and are able to learn highly nonlinear relationships directly from imaging data, thus addressing the drawbacks of conventional radiomics approaches and enabling end-to-end learning. We investigated whether CNNs are capable of quantifying loco-regional tumour control (LRC) based on CT imaging of patients with locally advanced head and neck squamous cell carcinoma (HNSCC). Material and Methods A multicentre cohort consisting of 302 patients with locally advanced HNSCC was collected and divided into an exploratory and a validation cohort (207 and 95 patients, respectively). All patients received a CT scan for treatment-planning and were treated by primary radio(chemo)therapy. 9725 transverse CT slices from the exploratory cohort were used to train a CNN with eight convolutional layers. For every patient (with one exception) we used 23 CT slices cranial and caudal of the slice with the largest tumour area, resulting in 47 slices per patient. Discriminative performance was evaluated using 4465 slices of the validation data set. The hazard of loco-regional recurrence was estimated by the CNN maximising the likelihood of the Cox proportional hazards model, which allows for incorporation of nonlinear relationships between the imaging features and the hazard prediction. The final hazard for every patient was obtained by averaging the results of the individual slices. The prognostic value of the model was evaluated by the concordance index (C-Index). Patients were stratified into groups of low and high risk of recurrence using the median hazard in the exploratory cohort. Results The validation of our CNN model revealed a C-Index of 0.68 (95% confidence interval: 0.57-0.79) for the prognosis of LRC. The estimated hazards were used to stratify patients into two risk groups. LRC significantly differed between these groups, both in the exploratory and the validation cohort (log-rank p<0.0001 and p=0.0005, respectively). Compared to previously published results with an average validation C-Index of 0.62 based on conventional radiomics [1], prognostic performance was slightly improved. [1] Leger et al. Sci Rep 7: 13206 (2017).

Conclusion We showed that CNNs are capable of automatically stratifying patients with locally advanced HNSCC into high and low-risk groups for loco-regional tumour recurrence. The obtained results suggest that deep-learning based approaches can become useful for non-invasively evaluating individual recurrence risks encouraging future research in this area. OC-0497 Predictive modelling of risk of breast fibrosis at >10 years after radiotherapy using the RILA assay C. Herskind 1 , P. Seibold 2 , I. Helmbold 2 , E. Sperk 1 , F.A. Giordano 1 , S. Behrens 2 , F. Wenz 1 , J. Chang-Claude 2 , M.R. Veldwijk 1 1 Universitätsmedizin Mannheim- Medical Faculty Mannheim- Heidelberg University, Department of Radiation Oncology, Mannheim, Germany ; 2 German Cancer Research Center, Division of Cancer Epidemiology, Heidelberg, Germany Purpose or Objective The radiation-induced lymphocyte apoptosis (RILA) assay has shown associations with the risk of late adverse reaction in cohorts of radiotherapy patients with mixed and single-entity tumors. However, most published studies have scored late reactions up to three years only. The purpose of the present study was to test the assay in breast cancer patients who had undergone radiotherapy after breast-conserving surgery (BCS) more than 10 years earlier. A particular aim was to compare the predictive value of RILA for fibrosis within and outside the surgical area, as well as for telangiectasia, and to assess the influence of other clinical risk factors. Material and Methods Patients from the German ISE cohort (BCS and adjuvant radiotherapy without chemotherapy) with median 11.6 years of follow-up were included in the analysis. RILA for CD4 + and CD8 + T cells was determined by flow cytometry of peripheral blood cells 48h after irradiation with a dose of 8 Gy (6 MV X-rays). Late reactions scored by LENT-SOMA were dichotomized as moderate-severe (grade 2-3) vs non- mild (grade 0-1). Fibrosis was scored outside and within the surgical area. Multivariate logistic regression model included: Age at surgery, BMI, hypertension, smoking status, total dose (EQD2), and hormonal treatment. Multivariate predictive modelling was performed by bootstrapping using c statistics to evaluate discrimination High CD4 + RILA values were inversely correlated with the risk of fibrosis (p=0.011) and telangiectasia (p<0.001) while CD8 + RILA showed a trend (p=0.06) for telangiectasia only. Univariate ROC analysis and multivariate analysis showed higher AUC and c-stat values outside than within the surgical area. Notably, the improvement by including CD4 + RILA in the multivariate analysis relative to including of risk. Results