Three-Dimensional Medical Image Registration with Applications in Proton Therapy

 Abstract: Three-dimensional Deformable image registration (DIR) is an important technique in medical imaging, facilitating the alignment of images acquired from different sources or at different times for clinical diagnosis and treatment planning. In this work, we develop a novel unsupervised deep-learning-based method to perform 3D DIR with better registration accuracy compared to state-of-the-art DIR methods with desirable diffeomorphic properties in comparable running time. In addition, a comprehensive deep-learning-based DIR performance evaluation method using target registration error (TRE) is developed to quantify the registration accuracy and further refine the initial registration result. We incorporated the DIR and its evaluation method into the joint statistical iterative dual-energy CT (DECT) reconstruction algorithm to reduce motion artifacts caused by inter-scan motion from the sequential DECT acquisition for the proton therapy treatment planning. 

Three-Dimensional Medical Image Registration with Applications in Proton Therapy

 Abstract: Three-dimensional Deformable image registration (DIR) is an important technique in medical imaging, facilitating the alignment of images acquired from different sources or at different times for clinical diagnosis and treatment planning. In this work, we develop a novel unsupervised deep-learning-based method to perform 3D DIR with better registration accuracy compared to state-of-the-art DIR methods with desirable diffeomorphic properties in comparable running time. In addition, a comprehensive deep-learning-based DIR performance evaluation method using target registration error (TRE) is developed to quantify the registration accuracy and further refine the initial registration result. We incorporated the DIR and its evaluation method into the joint statistical iterative dual-energy CT (DECT) reconstruction algorithm to reduce motion artifacts caused by inter-scan motion from the sequential DECT acquisition for the proton therapy treatment planning.