Irradiating a moving tumor has been one of the ultimate goals of radiation oncology. Older forms of radiation oncology could not account for the motion associated with an organ containing a tumor. Radiation beams were targeted at a tumor with the assumption that most of the beam would hit the tumor and the rest of the beam would not have enough strength to damage surrounding healthy tissue. For example, irradiating a tumor within a man’s prostate needs to take into account that this organ will move during treatment.

This is a fascinating monograph that addresses the mathematics and algorithms associated with pulmonary image analysis. Since respiration results in motion within the lungs, there is an interest within pulmonary image analysis in predicting the motion of the lungs. There is also an interest in detecting certain structures of the lungs, such as fissures, which are difficult to detect using computed tomography (CT) and magnetic resonance imaging (MRI) scans because these structures have a low contrast on a medical image.

Registration uses multiple medical images of the same location within the body to bring more information into a composite image. Fissures, which may not appear on a single CT scan, may become discernible with the composite image that is created by using registration techniques to merge these images together.

Since the lungs typically have some movement associated with them, there is an interest in predicting the type of motion that can occur. The author points out that other organs in the body can have some motion associated with them. The approach for assessing motion within an organ was to use four-dimensional (4D) CT imaging. This consists of a series of three-dimensional views of an organ within the body over the course of time. With 4D analysis, assessing information about the patient-specific respiratory motion became possible. This let radiation oncologists optimize a treatment plan by dimensioning safety margins. In effect, the techniques that were developed gave radiation oncologists the ability to irradiate a moving target.

This monograph gives an overview of current methods for lung registration in chapter 2. Chapters 3 and 4 give the mathematical foundations of image registration and level-set-based segmentation. Chapters 5 and 6 present models for novel registration. Chapter 7 gives an extensive evaluation study based on clinical CT scans. Finally, chapter 8 gives a discussion of the developed methods. Strengths and limitations are summarized and an outlook on further developments is given.

The appendices in this monograph are also of interest. Appendix A provides a discussion of digital images, and Appendix B focuses on mathematical derivations. Finally, Appendix C gives supplementary results concerning the clinical data that were discussed in chapter 7.

The methods discussed in this monograph could be generalized to other organs. Although radiation oncology is suggested as benefiting from these methods, any form of computer-assisted surgery could also benefit.