ESTRO 2024 - Abstract Book

S59 ESTRO 2024 this session, we delve into different MRI sequences in radiation oncology, their clinical applications, and the implications for training and future advancements. Invited Speaker

3376

Lessons to learn from adaptive photon therapy: What traps to avoid?

Lorenzo Placidi

Fondazione Policlinico Universitario Agostino Gemelli, IRCCS, Department of Diagnostic Imaging, Oncological Radiotherapy and Hematology, Roma, Italy

Abstract:

Adaptive radiation therapy (ART) is an evolving paradigm able to account for ongoing changes in the patient's anatomy during the treatment course, supporting an increasingly more accurate targeting of disease. The idea is to use imaging information to update the plan daily (or on a more frequent basis in an intra-fractional manner) instead of keeping one static plan throughout the whole course of therapy. Therefore, ART offers potential improvements in delivery accuracy and adaptation to anatomic and/or functional changes. ART is essentially based on the ability to quantify and evaluate the patient's anatomical situation before and/or during treatment, significantly shifting the paradigm from treatment reproducibility to treatment re-optimisation. Image guided radiotherapy (IGRT) plays a unique and important role in this new approach. To date, the main IGRT techniques used in clinical practice for ART are: CBCT and MRI guided radiotherapy (MRIgRT). CBCT provides high quality 3D imaging, allowing precise tumour localisation and patient positioning, although image contrast and degradation are strongly influenced by anatomical location. MRIgRT, especially when integrated with a radiotherapy system (MRI LINAC), offers ideal soft tissue contrast visualization and, above all, the ability to track and gate the tumour in real time and to adapt the treatment online based on the daily anatomy. Not only IGRT plays a key role in the online ART workflow: deformable image registration, automatic segmentation, accelerated replanning, patient-specific quality assurance (QA) and online tracking are also relevant components of a complex workflow in which AI will play an increasingly important role. Registration uncertainties: simulation CT registration introduces uncertainties that can compromise treatment accuracy. Strategies to minimise these uncertainties are essential for effective treatment delivery. Accuracy of synthetic CT (sCT): the reliability of sCT generation is critical for accurate dose prediction and optimisation. Overcoming the challenges of sCT accuracy is essential to ensure treatment efficacy. Fast and accurate auto-contouring: auto-segmentation is a key component of treatment planning automation and presents challenges in accurately delineating tumour volumes and critical structures. Ensuring the reliability and consistency of auto-segmentation algorithms is critical to treatment precision, speed up the process and standardize such process. Planning challenges: maintaining robust treatment plans amidst anatomical variations, especially in extracranial sites such as the upper and lower GI and pelvis, presents significant planning challenges. Optimisation of beam selection and intra-fraction re-optimisation strategies are essential to adapt to these variations. Real-time imaging: the integration of real-time imaging modalities is paramount for on-the-fly adjustments. However, challenges such as tracking latency and interplay effects need to be addressed to reduce treatment uncertainties. Based on real-patient data and clinical experiences with the photon therapy system, several pitfalls and considerations can be identified when applying (online) ART to proton therapy.