Immobilization devices can reduce body and organ movement to a certain degree, whilst image-guided localization and tracking, with or without immobilization systems, may reduce the targeting spatial uncertainty to a minimum.

We first discuss the imaging techniques used for locating the tumor in 3D. A variety of technologies are used—some of which involve imaging before the treatment, while others allow monitoring of target during treatment, either continuously or intermittently. These techniques include:

With the target’s 3D position determined, either the patient position or the beam position/direction can be adjusted as desired if the target is well-immobilized. For moving targets, such as targets with respiratory motion (lung and upper-abdominal tumors), the next step, image-guided tracking, becomes necessary. Significant effort has been devoted to solving the problem [14]. Various techniques have been developed in the past, including:

The above two techniques, however, may not be good enough for SBRT in certain locations. Real-time dynamic tumor-tracking methods or respiratory-synchronized techniques offer a much improved solution. Currently this technique has only been realized by Synchrony^TM Respiratory Tracking system (Accuray, Sunnyvale, CA). In this system, the treatment plan is based on a static CT scan, usually in the end-expiration phase. The accelerator is robotically controlled, which allows for non-isocentric beam delivery within a solid angle of over 2π. In order to compensate for the time delay between data acquisition and accelerator repositioning, the tumor’s real-time position is predicted by the correlation model of an optical monitor system (with three optical light-emitting markers fitted on a tight vest worn by the patient) and a pair of orthogonal X-ray images which locate the internal gold fiducials implanted inside the tumor. The external marker position is continuously measured using the optical system and the X-ray images are taken every 3–5 beams (or in newer systems according to how much time has elapsed since the last image, a flexible user-defined variable). The patient can breathe freely; the beam is delivered by a linear accelerator mounted on a robotic arm which follows the corresponding model. This model is checked and updated by acquisition of new pairs of X-ray images. With this tracking system, the tumor’s translational movement can be largely compensated, but not for the tumor rotation and body deformation. As a result, a margin of up to 5 mm may be required in selected cases [23]. An improved tracking mode, X-sight Lung by Accuray, allows the dynamic tracking to be achieved based on the lung tumor radiograph images without using fiducials for certain lung cases (about 30 % patients in some centers).

Besides using the robotic controlled linac, the dynamic tumor-tracking can be achieved by a technique currently being developed by Vero, in which a linac is mounted on gimbals that allows tilting of the linac following a moving target. Conceptually, this tracking may also be achieved by dynamic moving MLCs in a conventional linac [24–31] based on the MV EPID or KV imaging. This approach, however, is still in a development stage.

CT is a primary imaging modality for SBRT. Since CT provides geometric information with minimal spatial distortion (less than 1 mm in 20 cm as the general QA requirement) and provides the tissue density needed for dose calculation, it is the basis for treatment planning. CT images are helpful in identifying pulmonary nodules, parenchymal diseases, and chest wall involvement for superior sulcus tumors and lung diseases [32, 33]. Contrast-enhanced CT is the most sensitive study for the hepatic system [34]. A typical scan length should extend at least 5–10 cm superior and inferior beyond the treatment area; for non-coplanar systems this length should be extended to 15 cm. Tomographic slice thicknesses of 1–3 mm are required.

MR and ¹⁸F-fluorodeocyglucose (¹⁸FDG) PET are frequently used. As the gold standard for visualization of brain neoplasms, MR also found increasing applications in prostate, spinal tumors, and chest and solid abdominal tumors. PET (Positron Emission Tomography) is valuable in staging, planning and evaluating treatment response and might predict long-term outcome [35]. PET-CT is now popular for reduced image registration uncertainties and widely used for lung cancer, head-and-neck tumors, colon cancer, etc. Whenever possible, the initial patient setup (along with any immobilization device) during CT simulation should be reproduced during other imaging modalities, which helps to significantly reduce the uncertainty when it is fused with the CT imaging.

Among the photon beam characteristics, the beam penetration power and the beam penumbra are important factors to achieve the following dosimetry goals:

A new development is the photon beam generated without usage of the flattening filter (Varian TrueBeam^TM, Varian Medical System, Palo Alto, CA) [38, 39]. When flattening filter-free (FFF) beams are used, out-of-field doses can be reduced significantly. This is mainly due to reduced head scatter and residual electron contamination. FFF beams should therefore lead to reduced peripheral doses and patients may benefit from decreased exposure of normal tissue to scattered doses outside the field. Removal of the flattening filter implies also the possibility to deliver treatments with higher dose rates, up to a factor of 4 at 10 MV, and with a much higher dose per pulse. Besides further improving time efficiency for delivery, this might have subsequent potential radiobiology implications. Whilst interest for FFF beams is increasing in the medical physics field, there are few instances where FFF beams are applied in clinical practice, particularly in SBRT treatments. Initial clinical outcomes for local control are promising. However, aspects related to standardization, dosimetry, treatment planning, and optimization need to be addressed in more detail in order to facilitate the clinical implementation [38, 39].

Tomotherapy is a new concept for beam delivery [8]. A small megavoltage X-ray source was mounted on a ring gantry in a similar fashion to a CT x-ray source, the patient moves through the bore of the gantry simultaneously with gantry rotation. The beam intensity is modulated by temporally modulating multiple independent leaves that open and close across the slit opening. The geometry provided the opportunity to provide CT images of the body in the treatment setup position. The incident electron beam with 6 and 3.5 MeV in the treatment and imaging modes, respectively.

Cobalt-60 with MLC is also a suitable beam source. It has been shown that the common assumption that the cobalt penumbra is inferior to linac penumbra for MLC-based IMRT is not supported by comparative analysis [40]. Nearly identical plans can be achieved using cobalt source when compared to 6 MV IMRT.

Proton beams are being investigated for SBRT in certain sites. Dosimeric studies for using proton beam in SBRT have been performed by comparison between photon and proton-based SBRT for liver and lung treatment [41, 42]. It indicates that protons resulted in lower doses to critical organs at risk and a smaller volume of non-targeted normal lung exposed to radiation. However, the clinical significance and relevance of these dosimetric improvements remain unknown at the moment.

The terminology defined for conventional radiation therapy by ICRU 50 and ICRU 62 [43, 44], such as GTV, CTV, PTV, and ITV, is still used in SBRT. In many cases, especially for the metastatic lung, liver, and paraspinal cases, the GTV and CTV are often identical [45–47].

Due to the inevitable movement in most of the extracranial targets, SBRT is inherently imprecise compared to SRS and SRT, unless adequate immobilization or compensation for the target motions, both inter- and intra-fractional, is achieved. Thus dosimetric benefits of targeting essentially rely on the appropriate utilization of immobilization devices and/or on the techniques for precise organ motion compensation [23, 48–50]. The uncertainty in the CTV is generally accounted for by an internal margin added to the CTV, resulting in the internal target volume (ITV) [44]. The PTV then addresses all other possible geometrical variations, including setup uncertainties, machine tolerances, and intra-treatment variations, by adding all these margins to the CTV. This margin should be determined based on the particular SBRT system, the target site, and other clinical conditions.

The main objectives for high-dose-per-fraction SBRT are the precise and conformal targeting and rapid dose fall off to normal surrounding tissues. These objectives can be spelled out in the following dosimetric considerations during treatment planning: