ESTRO 2024 - Abstract Book

S8

Invited Speaker

ESTRO 2024

3284

AI for automation of contouring in brachytherapy

Matteo Maspero

UMC Utrecht, Radiotherapy, Utrecht, Netherlands

Abstract:

Radiotherapy holds a central role in cancer treatment, with the objective of delivering precise radiation doses to tumor tissues while safeguarding adjacent healthy structures. Automatic segmentation of target and organs at risk structures is indispensable, enhancing efficiency, accuracy, and consistency in treatment planning. This paper offers a thorough overview of the integration of deep learning techniques into brachytherapy for automatic segmentation. The necessity for automatic segmentation arises from the intricate and variable nature of anatomical structures, necessitating meticulous contouring for treatment planning. Pre-artificail intelligence segmentation techniques, such as manual contouring and atlas-based algorithms, exhibit limitations in coping with the escalating complexity of contemporary treatment plans. Deep learning, a subset of artificial intelligence, has emerged as a potent tool for automatic segmentation, providing improved accuracy, efficiency, and speed. The implementation of these technologies in clinical practice is examined, emphasizing the necessity for rigorous validation and seamless integration into existing workflows. In the context of brachytherapy, this paper scrutinizes the principal clinical applications and the current status of autocontouring. Addressing challenges and opportunities in applying deep learning to brachytherapy, the paper underscores the importance of tailored solutions for this distinctive treatment modality. The conclusion encourages thinking beyond prevailing applications, advocating for the exploration of AI in novel areas within radiotherapy. As technology advances, contemplating unconventional applications of deep learning, such as treatment response prediction and adaptive planning, may unlock new possibilities for optimizing cancer care. Various approaches to autocontouring using deep learning are explored, encompassing a review of available commercial solutions. The capabilities and potential benefits of these solutions in clinical settings are underscored.

3285

AI in brachytherapy for dose optimisation

Shirin Abbasinejad Enger

McGill University, Medical Physics Unit, Department of Oncology, Montreal, Canada

Abstract:

Artificial intelligence (AI), particularly deep learning, promises to streamline and increase various aspects of the brachytherapy workflow. This includes tasks such as tumour and organs at risk segmentation, catheter digitization, dose prediction, treatment plan optimization, quality control and assurance of treatment plans. This will improve