ESTRO 2020 Abstract Book

S423 ESTRO 2020

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.

respiratory motion, especially in case of irregularly breathing patients. The recent integration of radiation- free Magnetic Resonance Imaging (MRI) systems with radiotherapy treatment units is put forward to overcome this issue. These systems provide time-resolved 2D imaging as state-of-the-art imaging for motion compensation, although limitation in the temporal-spatial trade-off prevents the acquisition of real-time volumetric (3D) data. Efforts have been also made to implement motion models able to predict motion states not depicted directly in 4D imaging, as a way to support on-line tumor tracking or off- line treatment verification and adaption. Local motion models correlating the internal anatomy with external respiratory surrogates are already available in the clinics for tumor tracking in X-ray radiotherapy (e.g. Vero and Cyberknife); global motion models designed to predict the entire 3D anatomo-pathological configuration at different respiratory phases are not yet clinically applied, though conventional and particle-based treatments could potentially drag great benefit from them. Motion models derived from 4D imaging and related motion prediction algorithms rely on deformable image registration (DIR), which has gained popularity to assess inter- and intra- fractional uncertainties due to patient anatomical changes and target motion. Although this raises promising perspectives for the improvement of treatment outcomes and quality of radiotherapy clinical practice, the variety of proposed DIR algorithms, combined with the lack of effective quantitative quality control metrics of the registration, is slowing the transfer of DIR into the clinical routine and still requires an appropriate awareness on related potentials and limitations in terms of geometric and dosimetric accuracy. Particularly, there are evidences that the quality of motion states prediction models is affected by DIR inaccuracies which may also depend upon artifacts affecting conventional respiratory correlated imaging modalities. Therefore, there is the need to define common guidelines and patient-specific quality assessment for DIR and related motion models, as a way to grant a rigorous geometric and dosimetric treatment evaluation and adaption. Recently, an AAPM task group report was published outlining the essential aspects of DIR for image guidance in radiotherapy and providing an overview of different methods and clinical guidelines which can be considered the first groundwork for a safe use of DIR. This applies for both the providers of commercial treatment planning systems featuring DIR capabilities and for clinical end-users, who are called to make a conscious use of DIR in the frame of common clinical guidelines. However, this effort should be complemented by the application of automatic and patient-specific DIR validation quantitative tools, to be also extended to the validation of motion modelling approaches. These latter also suffer from the lack of ground truth data to evaluate their predictive capability of geometric and dosimetric deviations as a function of motion. Digital phantoms turn out to be particularly useful to overcome this limitation, although their translation into clinical cases is not straightforward. This talk aims at providing insights on the current state of the art of DIR and motion states prediction model algorithms and the related applications for the treatment of mobile lesions with X-ray and particle radiotherapy, bringing about a strong focus on the problem of geometric and dosimetric validation.

SP-0772 For the motion (rebuttal) A. Dekker MAASTRO, Maastricht, The Netherlands SP-0773 Against the motion (rebuttal) A. McWilliam University of Manchester, UK

Abstract not available

Symposium: Tumour tracking – is it worth the effort?

SP-0774 Image registration algorithms and prediction models - overview and pitfalls G. Baroni 1 , C. Paganelli 2 1 Fondazione Cnao, Clinical Bioengineering Unit, Pavia, Italy ; 2 politecnico Di Milano, Department Of Electronics Information And Bioengineering, Milano, Italy Abstract text In external beam radiotherapy, organ motion detection and compensation is a crucial issue, in order to grant accuracy and effectiveness of the loco-regional treatment of mobile tumors. This is even more important in the perspective of a geometric and dosimetric treatment adaption, when tumor tracking techniques are adopted or when dealing with unconventional treatment modalities such as particle therapy. Over the last few decades, the use of volumetric imaging modalities to aid in target definition has become essential and respiratory-correlated (4D) imaging modalities are applied on a routine basis to quantify uncertainties due to organ motion, with 4D computed tomography (4DCT) being the current clinical standard. However, 4DCT cannot be considered representative of each breathing cycle (intra-fraction variability) at every therapy fraction (inter-fraction variability), thus limiting an accurate description of

SP-0775 How to QA tumor tracking in the clinic? Equipment and patient specific aspects