Applications of generative adversarial networks (GANs) in radiotherapy: narrative review

Background and Objective: Radiation therapy (RT) is the dominant method for clinical cancer treatment, which aims to ensure that planning target volume (PTV) receives a sufficient dose while organs-at-risk (OARs) are exposed to little or no radiation. However, obtaining a clinically acceptable radiotherapy plan often requires a long time, tedious work, and a high level of physician experience. The general steps to perform RT include planning [computed tomography (CT)/magnetic resonance imaging (MRI)/positron emission tomography (PET)] image acquisition, contouring the treatment area (gross tumor volume, OARs, etc.), and developing a treatment plan and treatment implementation. But there are still some challenges that need to be overcome. Fortunately, with the development of the computer science, Generative Adversarial Network (GAN) which is composed of a generator and discriminator with opposing optimized goals has been widely used by an increasing number of applications in various fields, especially in CT, MRI, and other images and plays a great role in RT. This review aims to provide an up-to-date snapshot of GAN applications in radiotherapy. Methods: We searched for studies published from January 2018 to March 2022, with English language restrictions on PubMed and IEEE Xplore databases. Key Content and Findings: (I) GAN is an active field of research to support the automation of the radiotherapy workflow; (II) clinical validation is still needed to guarantee the usability of GANs in radiotherapy. Conclusions: GAN model has already been widely used in RT. Thanks to their ability to automatically learn the anatomical features from different modalities images, improve quality images, generate synthetic images and make less time consumption automatic dose and plan calculation. Even though the GAN model cannot replace the radiotherapy doctors’ work, it still has great potential to enhance the radiologists’ workflow. There are lots of opportunities to improve the diagnostic ability and decrease potential risks during radiotherapy and time cost for plan calculation.