(Wednesday, 7/15/2020) 11:30 AM - 12:30 PM [Eastern Time (GMT-4)]
Room: Track 5
Artificial Intelligence is rapidly evolving from an academic research tool into a method that has direct clinical applications for many activities in the radiotherapy process. Currently automatic segmentation based on deep-learning techniques, followed by automatic planning based on AI where a large number of previously irradiated patients are used to come up with a treatment plan for the next patient, have been introduced into a few commercial products. However future perspectives are not yet clear. New applications will arise in areas such as automated machine QA, creation of pseudo CT from MRI and deep learning based deformable image registration and predictive modeling from real world data. The actual clinical benefit of most of these still needs to be proven as, from the implementation side, the proper route of how to train, commission and validate these algorithms is still under discussion. This opens opportunities for many groups, both in research and from industry, to discuss the future needs of radiotherapy and how to implement them clinically.
In this symposia experts on diferent AI application fields will share their expirience. The first talk will review the State of the art in deep learning contouring. In particular, it will focus on how it has transitioned from academic research to real clinical implementation over the past 5 years. The steps for the validation of this Technology will be reviewed. The main focus of this talk will be on the barriers to adoption within clinical practice, and how medical physicists can help to overcome these.
The second talk will discuss how AI has been successfully applied to radiotherapy planning, including VMAT and 3DCRT approaches. AI is also being applied to the quality assurance of treatment plans, including evaluation of field apertures (e.g. for cervix 4-field box), and complex dose distributions. Several of these approaches will be reviewed. Finally, the key components of safe deployment of AI-based planning and quality assurance will be discussed.
The last talk will review how to build an information infrastructure for the continuous large-scale capturing of multimodal health information, describe characteristics of self-cognizant hospitals, and illustrate its utility in generating novel prognostic models using AI applications.
Learning Objectives:
1. Appreciate the possibilities of Artificial Intelligence in radiotherapy
2. Review the various steps for safe implementation of AI techniques in clinical practice
3. Discuss pitfalls and drawbacks of using AI techniques in radiotherapy. (e.g. lack of accessible big data, multi-centre collaborations)
Handouts
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email
