CBCT/CT-Based Image Synthesis
Hao Zhang · 2023
Before radiation therapy, the patient usually undergoes computed tomography (CT) simulation to acquire images of the area of the body to be treated with radiation. The planning CT images are used to delineate the tumors and surrounding organs-at-risk, and then to design an optimal treatment plan and dose calculation for the patient. Cone-beam CT (CBCT) is equipped on most linear accelerators for image-guided radiation therapy. It can provide up-to-date volumetric information on patient anatomy, and therefore is widely used for accurate patient positioning and anatomy change monitoring (e.g., patient weight loss, tumor shrinkage) in radiation therapy. Inspired by the successes of deep learning in many fields, researchers have also investigated synthesizing images from CT/CBCT using deep learning techniques for various clinical applications in radiation therapy. This chapter reviews some of the major efforts including synthetic CT from CBCT, synthetic magnetic resonance imaging (MRI) from CT/CBCT, and synthetic dual-energy CT from single-energy CT.