ESTRO 2020 Abstract book

S467 ESTRO 2020

is observed as early as by fraction 10. The principle of adaptive radiotherapy planning relies on adjusting the treatment plan based on observed changes over the course of therapy. The increasing use of magnetic resonance imaging (MRI) for head and neck radiotherapy planning has the advantage of improved soft-tissue visualization, allowing for more confident assessment of anatomical tumor changes during treatment. MR-Linac (MRL) holds promise to facilitate such adaptive MR-guided radiotherapy (MRgRT) workflows by mean of daily on-line MRI during radiation treatment. Daily, MR guided RT adaptation of radiation dose based on change in shape in patients with OPC will be discussed in this presentation. Patients with complete response (CR) following neo- adjuvant chemo-radiotherapy (CRT) for rectal cancer adenocarcinoma are offered organ preservation. Response rates following (CRT) are dose-dependent, with dose escalation of >60 Gy increasing CR rates. MRI has potential to act as a biomarker, identifying good and poorly responding tumours to select patients for dose adaptation in order to improve treatment outcomes. Studies suggest that functional imaging such as diffusion-weighted images (DWI) can predict pathological complete response early during CRT. MRL has the capability of obtaining daily functional imaging, which can potentially be used to identify treatment response early during CRT. This can potentially facilitate tailoring RT dose based on response. Functional image guided RT for rectal cancers and other tumours will be discussed. 1 Department of Radiation Oncology MAASTRO- GROW School for Oncology and Developmental Biology- Maastricht University Medical Centre+, MAASTRO, Maastricht, The Netherlands Abstract text For almost every cancer patient, images are being used for diagnosis, treatment planning, QA, treatment verification and follow-up. But most of the times, these images are only used by humans for the specific task at hand. Radiomics is all about automated, high-throughput quantitative image analysis. The adagium of Radiomcs is "Images are more than pictures, they are data". Some might say, what is new? For them Radiomics is simply an extension of something we do everyday in radiation oncology. But with more data and computing power now available and better algorithms to "mine" images, we are now able to see much more and different things in images with Radiomics than humans can. Radiomics has already led to outcome prediction models which are better than those based solely on clinical variables and certainly better than humans. With the move towards Deep Radiomics - based on deep learning - we can expect to see many routine and often mundane tasks performed by care professionals being taken over by Radiomics leading to higher quality care at reduced cost and more stimulating work for our profession. Perhaps more importantly, Radiomics will be crucial for patients and cancer outcomes as it will power adaptive radiotherapy and decision making in combined modality treatments, especially with immunotherapy. Debate: This house believes that radiomics will improve predictive models in RT SP-0770 For the motion: A. Dekker 1

SP-0771 Against the motion T. Rancati 1

1 Fondazione IRCCS Istituto Nazionale dei Tumori, Prostate Cancer Program, Milan, Italy

Abstract text In the last six/seven years great attention and enthusiasm raised around radiomics, with a considerable amount of literature proposing to use radiomic features to identify tumour phenotypes on medical images, to describe tumour heterogeneity and possibly use these features as predictors of clinical outcomes. Many studies are trying to identify “image biomarkers” with predictive/prognostic significance. Various studies obtained some initial promising results on the associations between radiomic features and oncological outcomes. Despite the clinical potential of radiomics, most times enthusiasm was not coupled to clever use of radiomics. Heterogeneity of acquisition protocols and image reconstruction methods impairs the robustness of radiomic studies. Furthermore, the biological/physical meaning of most radiomic features is largely unknown and statistical associations were not able to gain an understanding of the relationship between imaging and biology. All these pitfalls emerge in an important way when multicenter studies are considered, and multicenter studies are for sure a need in radiomic studies, due to the need of large population sizes for both development and validation of models exploring the possible contribution of a large number of features. The debate will first focus on the exploration of the statement “This house believes that radiomics will improve predictive models in RT”: (a) predictive vs prognostic models, (b) discussing the meaning of improvement in models, AUC vs clinical utility vs decision making. Next current unmet needs and limits of inclusion of radiomics in predictive models will be discussed: (1) population size, event prevalence, overfitting; (2) sensitivity of the radiomic features to acquisition modes, patient positioning, reconstruction parameters; (3) biological/imaging meaning of radiomic features; (3) internal validation vs external validation and generalizability; (4) integration with clinical features, quantification of the potential complementary value of the different predictors. In summary, clinical experience is underlining that radiomics could be only one of the elements to be included in models, “integration” with all relevant (-omics ?) feature is the way to move forward.

SP-0772 For the motion (rebuttal) A. Dekker MAASTRO, Maastricht, The Netherlands

Abstract not received

SP-0773 Against the motion (rebuttal) T. Rancati (Italy) Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy

Abstract not received