Image Processing in Radiation Therapy PDF

Preface

The demand for image processing tools to provide integration of data into the radiotherapy process has led to exciting collaborations among medical physicists, computer scientists, and mathematicians. Image processing addresses the reconstruction, adaptation, integration, and evaluation of imaging data into the radiation therapy process to ensure an accurate, efficient, and effective use of the information for radiation planning, treatment delivery, and outcomes assessment. Image processing includes the mathematical manipulation of images for integration at the reconstruction and evaluation stage as well as the incorporation of this information into the increasingly complex radiation therapy workflow. This book provides a description of the techniques used as well as their applications in radiation therapy.

Over the past two decades, the influx of images into the radiotherapy process has dramatically increased. The field quickly moved from the acquisition of a single computed tomography at the start of treatment to the inclusion of multiphase computed tomography scans, magnetic resonance, positron emission tomography, and single-photon emission computed tomography for planning, volumetric imaging at the time of delivery, and the additional acquisition of diagnostic-quality images throughout the treatment course for reassessment. The addition of these images enables clinicians to identify the tumor and critical normal structures as well as the impact of physiological motion on the position of these structures, with increasing precision. In addition to the ability to identify these structures, advances in treatment planning (e.g., intensity-modulated radiation therapy) have enabled sophisticated treatment plans to be developed to precisely avoid these critical structures and target the tumor. The escalation of the number of images, the number of structures to be contoured, and the precision of treatment planning placed an increasing demand on the clinical process to integrate and segment these images. The time spent contouring increased from minutes to hours, introducing the demand for automated segmentation and the need for highly accurate registration of multimodality images. These demands further increased with the introduction of in-room imaging, which provided additional information of the patient, including the complexities that are exhibited over the course of treatment including motion and changes in motion patterns and response to treatment. This wealth of new imaging devices demanded novel reconstruction techniques and methods to handle uncertainties in the data acquisition, such as breathing motion. The acquisition of in-room volumetric images placed additional demands on image registration, requiring high efficiency to allow a streamlined integration of these technologies into a busy clinical workflow. With the ability to view the 3D anatomy of the patient on a daily basis, the detection of tumor and normal tissue response over the course of treatment created the demand for additional diagnostic imaging along with its associated registration and segmentation. This overall patient treatment can easily result in more than 40 3D image volumes and more than 20 structures to be segmented per image. With this trajectory, the demand for image processing, including reconstruction, segmentation, and registration, became clear.

This work spans the topics of deformable registration, segmentation, image reconstruction, and integration of these practices into the radiation therapy environment. Section I of this book motivates the need for advances in image processing in radiotherapy applications, including adaptive radiotherapy, online monitoring and tracking, dose accumulation, and accuracy assessment. Section II describes the mathematical approach to deformable registration. The description of similarity metrics used for registration techniques is described, including their effectiveness and applicability in the radiation therapy environment. In addition, Section II includes a detailed description and evaluation of parametric and nonparametric image registration techniques and their applications in radiation therapy processes. Section III addresses the techniques for image segmentation. This section addresses atlas-based, level set, and registration-based techniques, assessing their efficiency, robustness, and breadth of application. Section IV of this book focuses on advanced imaging techniques for radiotherapy. Advancements in 3D image reconstruction, respiration sorting, in-room imaging techniques, and advanced applications of image registration using a graphics processor unit are described.

The goal of this book is to provide a comprehensive description of techniques and algorithms for image segmentation and registration, in-room imaging techniques and advanced reconstruction techniques, and clinical rationale and implementation. The target audience includes medical physicists, clinicians, dosimetrists, radiation therapists, and trainees in medical physics and related fields. This book benefits from the contributions of nationally and internationally recognized authors who have developed state-of-the-art image processing techniques and pursued their integration into the clinical environment. Their extensive knowledge is evident from their comprehensive descriptions of the algorithms and techniques, practical examples, and resulting clinical translation. The editor is sincerely thankful for their excellent contributions.