ESTRO 2023 - Abstract Book

S426

Sunday 14 May 2023

ESTRO 2023

a wave of new tendences. After PACIFIC presentation, more physicians supported this strategy, however, after results from CheckMate 816 neoadjuvant strategy has reborn.

SP-0530 Biomarkers for patient selection and monitoring G. Walls 1 1 Queen's University Belfast, Patrick G Johnston Centre for Cancer Research, Belfast, United Kingdom

Abstract Text Individualisation of radiotherapy has been a goal for over 50 years and is beginning to be realised for patients with non small cell lung cancer (NSCLC) predominantly due to improvements in our biological understanding and imaging technology. In the advent of molecular stratification and targeted biological drug therapies in metastatic lung cancer, characterisation of tumour gene expression is now the gold standard for all new cases of NSCLC. In order to minimise turnaround times, testing is routinely performed in a reflex manner for all new NSCLC cases in many centres, and includes non-metastatic cases, providing an opportunity to study the impact of specific gene mutation on locoregional and distant control. Principles from the paradigm-shifting adjuvant targeted therapy studies such as ADAURA may be transferrable to the radiotherapy setting. The utility of ctDNA in patients with NSCLC for assessing the response to drug therapies and the detection of residual disease after surgery, designates it as an excellent candidate marker of disease relapse following radiotherapy. This non invasive assay has high specificity and sensitivity, and preliminary work by our group suggests that ctDNA levels change acutely during radiotherapy. Proof-of-concept papers by UK and Canadian groups demonstrated that acute fluctuations in ctDNA commonly occur during treatment of locally advanced NSCLC. Trials are ongoing with respect to the detection of minimal residual disease. Quantitative imaging analysis, ie the field of radiomics, has the potential to equip lung oncologists with numerical parameters to support clinical decisions. This type of analysis is of particular importance in lung cancer because of the difficulty in discerning tumour from radiation pneumonitis, fibrotic parenchyma and infective consolidation. Imaging derived indices hold promise in the search for lung cancer radiotherapy biomarkers in relation to treatment response, patterns of failure, normal tissue toxicity and survival. Before radiomic analysis can be implemented in practice however, evidence from randomised data is required and solutions to overcome artefactual findings that result from logistical variations. eg. scanner type and acquisition protocols. Following the recent advent of immunotherapy for advanced NSCLC, adjuvant immune checkpoint inhibitors have been integrated into the gold standard treatment of localised NSCLC being managed with definitive radiotherapy. Programmed death ligand 1 (PD-L1) expression levels are known to be predictive of response to immunotherapy agents, and may serve as a predictive parameter in relation to the radiation response. Curative-dose radiation may impact the PD-L1 expression of NSCLC cells, given the established immune effects of treatment on other elements of the tumour microenvironment. Personalised radiotherapy is increasingly achievable due to biological and technological advancements, and ultimately will inevitably result in better clinical and patient-reported outcomes though tailored dose, fractionation, systemic therapy sequencing and surveillance.

Debate: This house believes that C-arm linac is no longer the work-horse of radiotherapy

SP-0531 - 0535

Abstracts not available for this session

Symposium: AI for (online?) IGRT

SP-0537 From science to clinical practice: Activities in the ESTRO AI future group S. Korreman Denmark

Abstract not available

SP-0538 AI-based image synthesis M. Maspero 1 1 UMC Utrecht, Radiotherapy, Utrecht, The Netherlands

Abstract Text Medical imaging's impact on oncological patients' diagnosis and therapy has grown significantly over the last decades. Especially in RT, imaging plays a crucial role in the entire workflow, from treatment simulation to patient positioning and monitoring. In the last decade, synthesizing images inter or intra-modality has raised increasing interest, especially when based on artificial intelligence algorithms. Image synthesis generally aims at maintaining image quality or enabling treatment to reduce the number of images that need to be acquired. This talk will review the primary application for radiotherapy, covering, among others, MRI-only radiotherapy and CBCT based adaptive RT, providing an overview of the main approaches adopted. Finally, the SynthRAD2023 Grand Challenge (https://synthrad2023.grand-challenge.org) will be announced and briefly presented.

SP-0539 Towards AI-based motion modelling C. Paganelli 1 1 Politecnico di Milano, Dipartimento di Elettronica, Informazione e Bioignegneria, Milano, Italy