Electronic portal imaging devices (EPIDs) generate 2D images for portal verification of the radiation field as well as verification of the patient positioning [6]. There are a variety of EPID types, including liquid-filled (Li-Fi) EPID, camera-based EPID, and amorphous silicon (a-Si) EPID. Among these, a-Si EPID is the most commonly used in clinical practice [7, 8]. The EPID is equipped with the treatment machine where the therapy beam is along with (Fig. 5.6). In regard to the advantages and disadvantages of using EPIDs for image guidance, the contrast resolution of an EPID is lower than that of kV imaging [9]; however, the accuracy in matching the EPID images is usually excellent, reported to be on the order of 1 mm or 1° from a phantom study performed using (semi-) automated matching software [5, 10, 11]. On the other hand, large intra- and interobserver variations exist in the interpretation of images in image registration using an EPID [12–15]. Usage of an EPID can provide positional information regarding whether the target volume is properly covered or not even during treatment with respiratory or physiological movements of target tumors or organs.

MV-CBCT imaging can generate 3D images using the EPID. During rotational irradiation of the MV beam using relatively low exposure, projection images can be acquired automatically and the CBCT images subsequently reconstructed. The main disadvantages of using MV beam energy for the acquisition of volumetric images are the degraded quality of the images and increased radiation exposure of the patients due to increased Compton scatters. An important benefit of MV-CBCT is the high resolution of the acquired images despite the presence of dense metal objects such as hip prostheses or high-Z dental enamels [16]. When MV-CBCT was applied clinically, the absorbed dose in the irradiated area was about 5–9 cGy according to the anatomical site, thickness of the patient’s body or the irradiated volume, and the maximum dose was between 9 and 17 cGy [17]. In MV-CBCT images, the bony anatomy and 3D boundaries between the soft tissues and bony structures can be easily identified, and image registration can also be carried out easily with low-dose exposure. However, the dose required for identifying soft tissue boundaries, such as those of the prostate gland, urinary bladder, rectum, etc., would increase [17].

2D kV imaging using an on-board imager installed in the treatment machine is an effective method to evaluate setup errors precisely and adjust the patient’s position properly. The kV imaging system consists of a conventional X-ray tube and amorphous silicon X-ray detectors mounted orthogonally to the treatment beam axis (Fig. 5.7). The X-ray tube and detectors are retractable during treatment. The advantages of kV imaging over MV imaging for the correction of setup errors include not only lower dose exposure but also smaller interobserver variability, which allows improved reproducibility of the patient’s setup [18].

kV-CBCT imaging is performed during rotational irradiation of a kV beam and simultaneous acquisition of projection images. The kV-CBCT images are reconstructed from the projection images produced by the retractable X-ray tube and the detectors; the image contrast in kV-CBCT is generally superior to that in MV-CBCT [19, 20]. CBCT has the advantage of having the ability to detect soft tissues, which enables improved accuracy of setup of the patients, accompanying rapid 3D image registration, as compared to MV-CBCT. However, streak artifacts caused by high-Z materials are found in kV-CBCT images because of significant photon attenuation as compared to the case in MV-CBCT images.

CT-on-rails is defined as a diagnostic CT scanner installed in the treatment room for the purpose of verification of patient positioning (Fig. 5.8). The gantry of the CT-on-rails scanner can move across the patient, instead of the couch moving. The CT-on-rails scanner moves in line with the couch orientation and yields CT images, while the patient remains in the treatment position before or after the treatment. Diagnostic-quality CT images can be obtained by CT-on-rails, because the CT scanner installed in the CT-on-rails was originally developed for diagnostic radiology. Therefore, usage of CT-on-rails as a tool for image guidance may be expected to improve the accuracy of detection of soft tissue lesions, including the target tumors or organs, and reduce the interobserver variations in image registrations [21, 22]. However, several disadvantages of CT-on-rails have also been pointed out. From the structural basis of CT-on-rails, the CT scanner and the treatment machine need to be operated independently, because the CT-on-rails does not share the same gantry. These systematic limitations may cause the uncertainties in the following: (1) the patient couch position after a rotation, (2) the precision of the CT coordinates, and (3) alignment of contours with structures in the CT images. In addition, there is sometimes a vital time lag between the actual treatment and the image acquisitions for image guidance, resulting in unexpected movement of the patients during the time lag and impairment of the setup accuracy [23].

2D kV stereoscopic imaging allows measurement of 3D geometric variations with 2 kV X-ray tubes and the corresponding X-ray detectors in which the two X-ray beam axes are crossed at a certain point (Fig. 5.9). Two kV 2D images are acquired from a patient, and subsequently, two sets of 2D coordinates of geometric variations from the two images can be converted to the 3D variation.

Here, we introduce three commercially available systems. The CyberKnife system consists of a compact linear accelerator mounted on a robotic arm. The manipulator arm is controlled to direct the radiation beams to the region of the beam intersection of the two orthogonal X-ray imaging systems integrated to provide image guidance during the treatment. The ExacTrac system uses the optical positioning system and two sets of kV imaging systems to enable patient positioning during the treatment. The optical positioning system is used for the initial patient setup as well as the couch motion control. In addition, the system can manage the patient’s respiration and gated beam irradiation. The X-ray imaging system is used for verification of the patient’s position and readjustment using the internal anatomy or implanted fiducial markers on the images. The accuracy of positioning using the optical guidance is on the order of submillimeters [24]. The real-time tumor-tracking system was designed to track the targets using fiducial markers implanted in or close to the tumors in real time. The system consists of two X-ray cameras mounted in the ceiling and two X-ray tubes on the floor. The system has been used clinically for lung and liver tumors [25–29] and reduced tumor motion to less than 5 mm [26]. The real-time tumor-tracking system has been applied for the treatment of prostate cancer, and it has been shown that hypofractionated IMRT (2.5 Gy × 28 fr = 70 Gy, equivalent to 80 Gy at a daily dose of 2.0 Gy) can be administered safely and provides a reasonable biochemical control rate [30].

During treatment, the video-based 3D optical surface imaging system can generate 3D models of the patient surface using photogrammetry [31–34], and then the system evaluates and quantitates the difference in the positional gap between the planned 3D models of the patient as the reference and the observed surface models [35] (Fig. 5.10). The major advantages of the video-based 3D optical surface imaging system, as demonstrated in a phantom study, are that it provides accurate and reproducible image guidance on the order of submillimeters [36]. In addition, the optical surface imaging system allows for continual monitoring of the patient surface without any radiation exposure during the actual treatment. In regard to its clinical applicability, the video-based 3D optical surface imaging system is especially useful as an image guidance system for radiation therapy of breast cancer, because the breast surface is nonrigid, and alignment of the 3D breast surface achieved by using this system yields greater breast correspondence than that obtained with laser or portal imaging systems [31].

US images of the target tumors or organs are acquired before the start of treatment and the integrated computer software program overlays the planned contours of the target tumors or organs such as the prostate gland, which are delineated at the treatment planning, over the acquired US images (Fig. 5.11). Subsequently, adjustments of the patient’s position are performed to align the contours to the targets while monitoring the actual target position on the US image. The advantages of US imaging as an image guidance system are as follows: (1) no radiation exposure, (2) noninvasive image guidance, (3) excellent visualization of soft tissues, (4) relatively rapid acquisition of images, and (5) low cost, including the entire system. The first commercially available US imaging system was the BAT system (North American Scientific, USA), which can provide 3D information of the target based on two US images. Subsequently, a 3D US imaging system was developed (SonArray, Zmed Inc., USA) so as to overcome the limitations of the former system. Currently, the US imaging system is mostly used for image guidance in radiotherapeutic approaches for localized prostate cancer. The disadvantage of US image guidance is that it is partly dependent on the user’s experience and expertise, resulting in high inter-user variability and reduced accuracy relative to CBCT guidance with or without fiducial markers. The US imaging systems that are currently in use clinically calculate the required shift, but do not allow for automated correction of the treatment couch. This results in a compromise of accuracy and prolongation of the treatment time. Other limitations of the US imaging system are forced movement or deformation of the target tumors or organs by the external pressure of the US probe during the acquisition of images and also non-visualizability of the target tumors or organs located behind bony structures [37].

The Calypso system (Varian Medical Systems, USA) can track the locations of target tumors or organs in real time by using tiny electromagnetic transponders implanted within the tumor before the treatment (Fig. 5.12). Upon excitement of the system, the transponders yield radiofrequency waves, and the detection of radiofrequency waves by the receiver activates analysis of the location of the transponders. The transponders are implanted by an invasive method; however, the system provides real-time tracking of the target during the treatment with no radiation dose delivered for imaging. The system has been demonstrated in some studies to show interfraction localization within 2 mm of the X-ray-based positioning [38, 39].

MRI-based radiation therapy systems include MRI imaging and treatment beam delivery in the same room, and both radiation therapy and monitoring of the target location in 2D/3D can be performed in real time using MRI simultaneously (Fig. 5.13). The main advantage of MRI for image guidance is that it provides high contrast of soft tissues, yielding clear detection of the boundary between the target tumor and the surrounding normal tissues, and the modality does not entail radiation exposure. However, the clinical usefulness of a linear accelerator equipped with an MRI imaging system is still under investigation [40–49]. The cobalt irradiation system with MRI (Viewray, USA) is already available commercially.