Abstract: |
Artificial intelligence (AI) has emerged as a powerful technology with widespread applications in diverse areas ranging from autonomous cars to medical decision support. In this chapter, we review the role of AI in radiation oncology. Rapid developments in radiation oncology and other fields of medicine using AI have been due to the deep learning revolution that has demonstrated remarkable versatility and capacity to learn from several different types of modalities for several different tasks. Radiation oncology arguably uses the highest level of automation not only in cancer care but also across medicine generally. The need for automation to ensure accurate targeting of cancer while ensuring patient safety, and the increased reliance on image guidance for image-guided radiation treatments, has made AI and deep learning particularly relevant for radiation oncology. In this chapter, we will review the current state of the art in the advances of AI techniques with respect to radiation oncology and the anticipated, though uncertain, impact of AI. Emphasis will be placed on applications that rely on the analysis of images, as the clinical impact is imminent and likely transformative, in particular, automated tissue segmentation for normal tissues but also for tumors. The ability to extend auto-segmentation to “on-treatment” images, cone beam in particular, will allow for more accurate setup, dose-accumulation estimates, and possibly smaller margins (“AI-guided RT”). © Springer Nature Switzerland AG 2022. |