(Tuesday, 4/7/2020) 8:00 AM - 10:00 AM [Mountain Time (GMT-6)]
Artificial intelligence (AI) has recently raised tremendous interests in a spectrum of domains, including medicine. There have been significant research and development efforts on the applications of AI to solve problems in radiation oncology. As one of the indispensable steps in radiotherapy workflow, treatment planning using modern treatment planning systems (TPSs) embraces AI-led opportunities. Potentially, with high throughput computation power in information analysis and automation execution, AI implementation in modern TPS could improve treatment plan quality, plan quality consistency, and enhance plan efficiency in the current human operator-centered practice. As such, a few research groups have pioneered in the development of in-house AI-based software working in modern TPSs, ranging from organ contouring to inverse optimization, from external beam radiotherapy to brachytherapy. Several vendors have also released AI-related tools to improve the treatment planning process. Hence, it becomes critical for clinical physicists to understand the development of these tools, as well as the process to test, accept, validate them, and to evaluate their values in clinical practice. This session will bring together experts with experience in this area to discuss advances of several AI tools in modern TPS and their experience and lessons learned during the clinical translation at different institutions. We will briefly introduce the evolution of AI-related works of treatment planning and discuss the frontier works of current stage AI developments. We will also share our experience of implementing AI in clinical TPS with clinical examples. Lessons, concerns and potential caveats of AI developments and implementation in TPS will be discussed. We hope the session will trigger discussions and foster collaborations on the safe and effective use of AI tools to continuously improve our practice.
Learning objectives:
1. To give an overview of AI-related developments in treatment planning: including early efforts and current frontier AI developments
2. To learn how to implement AI in modern TPS through clinical examples: including clinical rationale judgment, execution workflow, quality assurance, and effective analysis
3. To understand the benefits/caveats of AI implementation in modern TPS
Funding Support, Disclosures, and Conflict of Interest: Chunhao Wang was partially supported by NIH RO1 CA201212.
Contact Email
