S3043
Invited Speaker
ESTRO 2026
neoadjuvant chemotherapy, outcomes appear comparable to EMBRACE cohorts treated with optimized radiotherapy alone, despite less favourable baseline characteristics of EMBRACE patients. Similar observations apply to recent immunotherapy trials such as KEYNOTE-A18. These findings suggest that the full therapeutic potential of high-quality, image-guided radiotherapy may not be fully reflected in trials incorporating suboptimal radiotherapy techniques. EMBRACE-based radiotherapy should therefore not only serve as the standard in future clinical trials but also be implemented broadly in modern radiotherapy centres worldwide. In this lecture, practical aspects of such implementation will be presented, based on experience from introducing these techniques across multiple institutions. Key challenges, common pitfalls, and the critical role of interdisciplinary collaboration in achieving optimal outcomes will be highlighted. In addition, real-world clinical outcomes of this complex treatment approach will be discussed, including emerging data and the recently initiated EMBRACE III REWIND study, which aims to evaluate its performance in routine clinical practice. 5450 Machine learning (and AI) opportunities for radiobiology research Sarah Catharina Brüningk Department of Radiation Oncology, Inselspital and University of Bern, Bern, Switzerland. Department of Digital Medicine, University of Bern, Bern, Switzerland Radiobiology seeks to understand and quantify radiation response across multiple levels of biological organization, from molecular damage signaling and cellular repair processes to tissue response and clinically observable outcomes. This multiscale nature makes the field both scientifically rich and methodologically challenging, particularly when attempting to integrate heterogeneous experimental and clinical data into predictive frameworks. In this context, machine learning and artificial intelligence (AI) offer a set of computational approaches that can complement mechanistic radiobiological modelling by identifying complex patterns, integrating diverse data modalities, and improving prediction of biological and clinical endpoints. This lecture will first position radiobiology as a discipline concerned with mechanisms and outcome definitions across different biological scales, emphasizing the implications of this complexity for data interpretation and model development. It will then introduce the fundamental components of AI- based modelling, including data representation, model architecture, training strategies, and evaluation
enabling RTTs to contribute beyond implementation and conduct, into study conception, design, and dissemination. The session will also address the role of RTTs as contributors to, and leaders within inclusive research programmes. This includes recognising how RTT-led research can improve relevance, feasibility, and impact, particularly in populations affected by cancer inequality. Attendees will be encouraged to consider actionable steps that can strengthen research capacity within their own settings, including mentorship, multidisciplinary networks, and smaller scale projects as stepping stones toward larger studies. By the end of the lecture, participants will be able to describe the key features of a developing radiotherapy research culture, identify barriers and enablers to RTT- led research, and consider practical approaches to advancing research activity within their own practice. 5447 Chemoradiation for locally advanced cervical cancer; Integrating EMBRACE II into clinical practice: Clinician perspectives Primoz Petric Cantonal Hospital, Cantonal Hospital, Lucerne, Switzerland. Department of Oncology, Aarhus University Hospital:, Aarhus, Denmark Standard treatment of locally advanced cervical cancer (LACC) consists of external beam radiotherapy (EBRT), concurrent chemotherapy, and brachytherapy. Over the past decades, the introduction of intensity- modulated radiotherapy (IMRT) and image-guided adaptive brachytherapy (IGABT), supported by GEC- ESTRO and ICRU ecommendations, has enabled individualized dose delivery adapted to tumour and organ-at-risk dynamics. Prospective and retrospective evidence, including the multicentre EMBRACE-I study, has demonstrated that MRI-guided IGABT achieves excellent tumour control with limited toxicity, establishing chemoradiation with IGABT as the standard of care. Five-year outcomes from EMBRACE-I showed local control of 92% and overall survival of 74%, with low rates of severe morbidity, and long-term follow-up suggests sustained disease control beyond five years. Further improvements in outcomes have been pursued through implementation of IGABT planning aims and dose constraints, and risk-adapted strategies, including risk-adapted para-aortic irradiation and integration of systemic therapies. The EMBRACE-II protocol demonstrated high rates of locoregional and distant control with minimal severe toxicity (ESTRO 2025). In contrast, randomized trials investigating additional chemotherapy have generally not shown benefit in unselected populations. Although the INTERLACE trial reported improved survival with
Made with FlippingBook - Share PDF online