From the Viewpoint of Therapeutic Radiology
The biological effects of radiation were investigated within a year after Wilhelm Conrad Rontgen discovered X-rays in 1895, and then it was recognized as beneficial effect for curing malignant tumor , which may be considered the beginning of radiation therapy. What are required in radiation therapy from medical physics point of view are (1) high conformity of dose distributions to tumor regions and (2) accurate tumor localization and patient positioning. To achieve these requirements, radiation therapy researchers have dedicated their efforts since around 1960 to the development of novel technologies such as conformal radiotherapy , intensity- modulated radiation therapy (IMRT) , real-time tracking radiotherapy (RTRT) , and image-guided radiation therapy (IGRT) [55, 56].
Since radiation therapy can preserve organ function and is useful in patients, particularly elderly patients, unsuited to surgery, it has attracted greater attention. Consequently, this treatment modality is considerably important for developed countries such as Japan and the United States of America, whose populations are rapidly aging. In Japan, the percentage of older people (65 years and over) was estimated to be around 23% in 2011 . Therefore, great benefits from radiation therapy can be provided for many patients, including elderly patients, whose quality of life could increase.
The primary aim of radiation therapy is to deliver as high a dose as possible to the tumor while causing as little damage as possible to normal tissues and organs at risk (OARs) [55, 56]. The OARs are normal tissues whose radiation sensitivity may significantly affect radiation treatment planning (RTP) and/or the prescribed dose . To achieve these goals, high-precision radiation therapy methods have been developed, such as stereotactic body radiation therapy (SBRT), intensity-modulated radiation therapy (IMRT), adaptive radiotherapy (ART), realtime tracking radiotherapy (RTRT), and image-guided radiation therapy (IGRT). These advanced techniques have recently led to impressive progress regarding the
Fig. 1.6 Five steps of radiation therapy and examples of image engineering techniques, including computational anatomy, in each step
precision of radiation delivery. As a result, high-precision radiation therapy has been reported to provide outcomes comparable to surgery for some cancers . In these high-precision radiation therapies, novel methods of image analysis, including computer graphics, image processing, and pattern recognition, play considerable roles in improving the accuracy of radiation therapy and assisting in treatment planning.
Radiation therapy procedure includes five steps: diagnosis, treatment planning, patient setup, radiation administration, and follow-up. Computational image engineering techniques are employed to assist the radiation oncology staff in decision-making at each step. Figure 1.6 describes the steps and provides examples of image processing techniques including computational anatomy in each step.
The first step is diagnosing the cancer. Computer-aided diagnostic technologies may be useful if the oncologist employs multiple imaging modalities. Then the radiation oncologist and clinicians determine the radiation treatment goal, i.e., curative treatment or palliative treatment.
The second step involves developing the treatment plan, in which the gross tumor volume (GTV), clinical target volume (CTV, the volume including gross tumor and areas of likely microscopic involvement), and planning target volume (PTV, the actual volume to be irradiated to ensure that the entire CTV receives a therapeutic dose, assuming some target motion and other inaccuracies) are defined. A plan is created by arranging beam paths to maximize the tumor dose and minimize the OAR dose. Figure 1.7 depicts illustrations of a radiation treatment plan: (a) a beam’s eye view with a GTV region and organs at risk (light blue, bladder; pink, rectum) and (b) dose distribution images produced using CT. Image engineering techniques are used to help define these volumes and OARs. In current clinical practice, GTV regions have been manually delineated by radiation oncologists using treatment planning
Fig. 1.7 Illustrations of a prostate radiation treatment plan: (a) a beam’s eye view with a gross tumor volume (GTV) and OARs outlined (light blue, bladder; pink, rectum) and (b) dose distribution using IMRT treatment planning
computed tomography (CT). However, the subjective manual contouring of a tumor region is tedious and time-consuming, and its reproducibility is relatively low, which could cause inter- and intra-observer variability [60-62]. A number of automated contouring methods for determining the GTV have been proposed to reduce this variability and planning time and increase the segmentation accuracy of the GTV [63-65].
The third step is patient setup. In this step, the radiation technologist positions the patient manually on the treatment couch as accurately as possible, often using immobilization devices. After that, fine-tuning of target position is performed by using image registration techniques, which register a moving image with a reference image with respect to the target. In general, in the patient setup phase, digitally reconstructed radiographic (DRR) images and planning CT images are used as the reference images. An electronic portal imaging device (EPID) and cone-beam CT (CBCT) images produced using kilovoltage or megavoltage X-rays during treatment are employed as the moving images. Previous studies have revealed that these techniques are effective for reducing setup error [66, 67].
The fourth step is the actual radiation treatment. A photon or particle beam is delivered to the planning target volume (PTV) in a patient according to the treatment plan. For tumors subject to respiratory motion such as lung and some liver cancers, respiratory gating or fiducial markers are used to follow the target. For example, an RTRT system has been developed, which employs pattern recognition techniques to follow gold markers within the tumor to track it and switch the treatment beam on and off .
Computational anatomy-based technologies could make it possible to develop statistical models of events happening in tumors and human bodies as well as mathematical prediction models of tumor or normal tissue responses during radiation treatment. If these novel technologies were developed, they could make large impacts on automated segmentation of GTV, CTV, and OAR and optimization of radiation treatment planning.