Generation of Synthetic Computed Tomography from Magnetic Resonance Images: A Literature Review

Abstract

Magnetic Resonance Imaging (MRI) is used in radiotherapy to designate target volumes and organs that are at risk because it offers greater contrast in soft tissues than Computed Tomography (CT) imaging does. The tissue electron density needed for dosage calculation, however, is not provided in MRI data. The most frequently studied anatomical localizations are the brain and pelvis, followed by the neck and head, abdomen, breast, and liver. A variety of methods for producing synthetic CT (sCT) using MRI scans have been designed to estimate radiation exposure. This study reviews different deep-learning methods used to generate synthetic CT images from MRI images.