After fabrication of the facial mask, thin slice CT is performed for planning. The data from the CT images are also used for absorption correction during the planning of dose delivery. For target delineation, MRIs are also obtained and merged into the planning CT. Target volumes are contoured on these images in line with the International Committee on Radiation Units and Measurements (ICRU) reports 50 and 62 [35, 36]. Gross tumor volume (GTV) is usually defined as the enhanced lesion on T1-weighted MR images and the surgical cavity. In the treatment of patients with GBM, a simultaneous integrated boost technique is usually employed, and layered clinical target volumes (CTVs) are settled to cover both the tumor bulk and microscopic spread of the tumor. Because the dominant pattern of failure is local after standard GBM treatment, higher dose delivery to the regional tumor is required. Therefore, the CTV of the tumor bulk (CTVb) is defined as being equal to the GTV. GBM is a highly infiltrative tumor, and migration of tumor cells into the surrounding brain should also be included in the CTV. Two different approaches exist regarding the definition of the CTV for the infiltrating area (CTVi). The first approach is based on the experiences of treatment failure after conventional RT. The majority of the recurrences have been reported to arise within 15–20 mm of the tissue surrounding the enhanced lesion, and the CTVi may be defined as the expansion of the GTV to encompass the surrounding 15–20 mm zone. This expansion is automatically performed in the planning software without consideration of the anatomical spread of tumor cells. Therefore, excessive volume should be manually removed from the CTVi after automatic expansion (Fig. 7.3). The other approach regarding the definition of the CTVi is based upon the estimated microscopic spread of tumors cells. As described above, the volume of the edema surrounding the enhancing lesion includes microscopically detected migrating tumor cells and is also defined as the CTVi. MRIs of both the T2-weighted images and fluid-attenuated inversion recovery (FLAIR) images are used to visualize edema. Although discordance between CTVs based upon T2-weighted and FLAIR images has been reported [37], FLAIR images are more frequently employed because the CSF is also visualized brightly on T2-weighted images and may impair the visualization of edema. After contouring the CTVs, the planning target volume (PTV) is defined by expanding the CTV with a 3.0–5.0 mm margin, which includes the immobilization and treatment errors.

Although MRIs provide precise geometrical information, they do not directly visualize the location of tumor cells. The contrast-enhanced area is usually diagnosed as tumor bulk, but enhancement is only the result of leakage of contrast medium through the impaired brain-blood barrier (BBB) of tumor vessels. The high-intensity area on T2-weighted or FLAIR images may include infiltrating tumor cells, but this does not mean that migrating tumor cells exist anywhere in these areas. Furthermore, induced neovascularization as a physiological reaction after surgery also causes contrast enhancement, and changes in the blood circulation after surgery also cause an increase in the intensity of normal brain surrounding the surgical cavity on T2-weighted or FLAIR images. Therefore, CTVs contoured on these images may include the area without the need for dose delivery. To decrease the excessive radiation delivery to the critical structures adjacent to the target, biological imaging is required for the delineation of real targets.

Amino acid (AA) positron emission tomography (PET) has been tried to improve the delineation of the distribution of glioma cells. l-[methyl-¹¹C]methionine (MET) and O-(2-[¹⁸F]fluoroethyl)-l-tyrosine (FET) are the most widely used tracers. These tracers are transported across the BBB by means of the membrane transport system, and the accumulation of these tracers reflects the protein synthesis taking place in tumor cells. The uptake of these tracers is not dependent on the disruption of the BBB, and the significantly higher accuracy of AA-PET relative to CT and MRI in defining the distribution of tumor cells has been reported [38–43]. MET-PET and FET-PET provide comparable diagnostic information for the delineation of tumors [44], even though the short half-life of ¹¹C (20 min) limits the use of MET-PET to institutions with an onsite cyclotron. In the majority of the primary cases, the AA tracer accumulates beyond the enhanced lesion and within the high-intensity area on T2-weighted or FLAIR MRIs (Fig. 7.4). Although target delineation based upon these accurate images will contribute to the sparing of neurological functions in patients after IMRT, the major remaining problem regarding the use of PET information is the threshold of uptake. The uptake of AA is semiquantitatively expressed by the ratio of uptake in the lesion to that in the contralateral normal brain (T/N ratio). However, limited data is available regarding the optimal value of the T/N ratio required for drawing the target volume and the optimal radiation dose for the control of these areas.

On the other hand, PET is also used to define the radioresistant area in the tumor. Hypoxia is one of the major causes of radioresistance in GBMs, but heterogeneous distribution of hypoxic cells has also been reported [45, 46]. Although some limitations remain, the significance of [¹⁸F]FMISO-PET in the visualization of the hypoxic area in gliomas has been reported [47, 48]. In the treatment of head and neck cancers, [⁶⁰Cu]ATSM-PET- and [¹⁸F]FMISO-PET-guided radiotherapy planning have been reported [49, 50], and these images may also contribute to the treatment of GBM in the future.

As the prognosis of patients with GBM is poor, some oncologists may feel little need for the preservation of critical structures from late toxicities [51]. However, temozolomide prolongs the survival of GBM patients, and the 5-year survival rates have improved to 13.8 % for patients with the methylated O-6-methylguanine-DNA methyltransferase (MGMT) gene [4]. Therefore, the issue of late toxicity has become more significant.

In treatment planning for brain tumors, OARs include the eye lens, retina, optic nerve, chiasm, cochlea, and brain stem. A study by Emami et al. in 1991 was the first to summarize the clinical experience regarding partial organ tolerance doses for OARs [52]. More recently, QUANTEC (quantitative analysis of normal tissue effects in the clinic) articles have updated/refined these data (Table 7.2) [53–56]. As these data are based upon clinical outcomes after conventional fractionated radiotherapy or radiosurgery involving single fraction treatment, the tolerance doses for OARs have been estimated from these data using the LQ model. However, the reliability of the LQ model and the value of the α/β ratio are still debated in hypofractionated RT (see Chap. 4), and further investigations are required to define the tolerance doses for OARs.

After contouring the PTVs and OARs, the prescribed doses and the tolerance doses are inputted into the inverse planning system, and the dose distribution calculated automatically. In cases where the tumor exists adjacent to the critical organs, it may be difficult to deliver the prescribed doses to the PTV homogeneously while keeping the doses to the OARs below the limit of the tolerance doses. In such cases, the preferred dose delivery/avoidance should be decided individually based upon the estimated prognosis and functions of the patient.

Although IMRT has improved the conformity and homogeneity of dose delivery to targets, especially concave tumors, conventional coplanar beam IMRT still has limitations with regard to the sparing of critical organs. In the treatment of tumors located in the posterior or temporal fossa, PTVs and OARs (optic nerve, chiasm, brain stem, and cochlea) may be included in the same axial plane, and the coplanar beam cannot completely avoid critical organs (Fig. 7.5a) [57]. In such cases, noncoplanar beam arrangement may decrease the doses to the OARs (Fig. 7.5b) [58].

IMRT is usually performed using multiple static intensity-modulated beams. However, using static fields, the isodose curves in the range of lower doses extend along the gantry angle [58]. This “tail” of the low-dose field extends more sharply and over a greater distance when the angle or number of fields is limited to decrease the doses to the OARs adjacent to the PTV (Fig. 7.5a). Recently, VMAT, a rotating IMRT, has also become available for the treatment of patients with GBMs. With this method, a more homogeneous distribution can be achieved, not only in the low-dose field but also in the PTV, while maintaining a decreased dose to the OARs (Fig. 7.5c). The range of rotation can be changed and a noncoplanar arc can also be made available to further decrease the doses to the critical organs.