The clinical target volume (CTV) includes the GTV with an added margin that is meant to treat subclinical microscopic disease and is usually “anatomically confined.” The term anatomically confined means that it does not extend beyond the bony calvarium, base of the skull, falx, or tentorium. The CTV extends to but not beyond neuroanatomic structures through which tumor extension or invasion is unlikely. When the GTV approaches the boundary of an anatomic compartment, the CTV extends to and includes the boundary (Fig. 7.2).

Planning target volume (PTV) includes a margin which is added to the CTV in three dimensions to create the PTV. It is geometric and not anatomically confined. The PTV has two components: the internal margin (IM) and the setup margin (SM). The IM is meant to compensate for all movements and variations in size and shape of the tissues contained within the CTV. The SM is meant to account for setup and mechanical and dosimetry uncertainties related to daily patient positioning and treatment equipment and software. For most brain tumor studies in children, the PTV margin is 3 or 5 mm (Fig. 7.2). The use of a PTV margin of 3 mm requires high precision immobilization and treatment verification methods known as image-guided radiation therapy. Given that the CTV is generally confined to the intracranial space, the PTV may extend into or beyond cranium but is unlikely to extend beyond the surface of the patient. Considering the CTV and PTV margins used for most brain tumor protocols in children, the imaging studies used to define the target volume should have a resolution less than the specified margins. Recent studies have shown that for children with brain tumors, a PTV margin of 3 mm is adequate provided that daily imaging is performed using cone beam CT [4]. Patients treated in the supine position have a smaller setup error than those treated in the prone position. The latter is often used for children with posterior fossa tumors. Similarly, those treated under anesthesia have the least amount of intra-fractional motion compared to children who are not anesthetized (Table 7.1). Children with posterior fossa tumors tend to be younger than those with supratentorial tumors.

Medulloblastoma is the most common malignant tumor in children and is named for its posterior fossa location. It is the signature tumor when demonstrating the importance of pediatric neuroradiotherapy because of the curative potential of irradiation, the complexities of requisite craniospinal irradiation, and the systematic studies that have been performed to test craniospinal dose reduction and to prove the role of chemotherapy. It is the only pediatric brain tumor system to use a staging system to guide the use of radiation therapy – a staging system that has served investigators for nearly four decades with few modifications. Irradiation of medulloblastoma is notable for the evolution in target volumes, for the treatment of the primary site, and soon for future changes that will be driven by tumor biology. The specter of side effects observed after the irradiation of this disease is legendary in the field of pediatric oncology.

In the United States, the recent negative (no difference between treatment arms) A9961 trial [5] for standard risk (nonmetastatic) medulloblastoma in children ages 3–21 years, which compared two chemotherapy regimens, vincristine (VCR)/cisplatin (CDDP)/cyclophosphamide (CYC) versus VCR/CDDP/lomustine (CCNU), was paused after nearly two decades of progress where standard dose craniospinal irradiation (36 Gy) and the posterior fossa boost (≥54 Gy) for all patients was replaced by reduced dose craniospinal irradiation (23.4 Gy) and the posterior fossa boost (55.8 Gy) followed by postirradiation chemotherapy to achieve an equivalent outcome in average-risk patients. Current and future progress includes further reducing the role of radiation therapy or at the very least the craniospinal dose for biologically favorable tumors, reducing the volume for the primary site boost (omitting the posterior fossa boost after craniospinal irradiation in favor of focal treatment of the primary site), increasing treatment intensity for biologically unfavorable or high-risk patients, and understanding better the effects of radiation dose and volume, clinical and treatment factors and host factors, and their effect on side effects after radiation therapy (Table 7.2).

Most clinical trials for medulloblastoma accrue a large number of patients and involve many institutions. Large trials are required to have the statistical power to detect small improvements in treatment. Large trials may lead to poor quality control in radiation therapy and diagnostic imaging [5, 6]. This is most evident in trials where craniospinal dose reductions have been the primary question and stems from inadequate staging or poor radiotherapy technique. This tragedy could be avoided if procedures were implemented to review all imaging and treatment plans prior to the initiation of treatment – a lofty goal which will not be achieved soon.

As noted earlier, during the past two decades there has been considerable evolution in radiotherapy target volumes for CNS embryonal tumors including medulloblastoma (Table 7.3). When there was a preliminary indication that full posterior fossa irradiation was not required for children with medulloblastoma [7], investigators focused on testing the feasibility and safety of reducing the target volume for the treatment of the primary site which customarily follows the craniospinal component of therapy. The first prospective report was from the SJMB96 study which was a multi-institution collaborative study that included all types of CNS embryonal tumors [8]. The SJMB96 study [9] was carried out from 1996 to 2003 and included prospective treatment with postoperative, risk-adapted, craniospinal irradiation (CSI) and postirradiation chemotherapy that included dose-intense cyclophosphamide, vincristine, and cisplatin. Average-risk (nonmetastatic) cases (n = 148) were treated using 23.4 Gy CSI, 36 Gy posterior fossa irradiation, and 55.8 Gy primary site irradiation using a 2 cm CTV margin and daily fractionation of 1.8 Gy. High-risk cases were treated using 36–39.6 Gy CSI and 55.8 Gy primary site irradiation using a 2 cm (pre-2003). There was no difference in craniospinal dose, target volume margins, or cumulative total dose based on tumor type. The results of this study showed that the 5-year cumulative incidence of posterior fossa failure was only 4.9 ± 2.4 % for patients with average-risk tumors [8] with a median follow-up of more than 5 years. The targeting guidelines used in this study resulted in a mean reduction of 13 % in the volume of the posterior fossa receiving doses in excess of 55 Gy compared with conventionally planned posterior fossa boost. The prospective trial has demonstrated that irradiation of less than the entire posterior fossa after 23.4 Gy craniospinal irradiation for average-risk medulloblastoma results in disease control comparable to that after treatment of the entire posterior fossa. These guidelines were the first prospective series that showed the ability to reduce the targeted volume safely in children with CNS embryonal tumors. With the exception of the A9934 study, the other studies are works in progress. The A9934 study [10] was a trial of systemic chemotherapy, second-look surgery, and conformal radiation therapy limited to the posterior fossa and primary site for children between 8 months and 3 years with nonmetastatic medulloblastoma. Following resection, these very young children received four 4-week courses of induction chemotherapy (cyclophosphamide, vincristine, cisplatin, and etoposide) followed by a second surgery when necessary and age- and response-adjusted irradiation of the tumor bed and posterior fossa. Patients then received maintenance chemotherapy consisting of four alternating cycles of cyclophosphamide and vincristine followed by oral VP-16. For the 78 eligible patients, the 3-year event-free survival (EFS) was approximately 50 % and by extent of resection 58 % (gross-total resection) and 36 % (<gross-total resection). There was a striking pattern of recurrence related to the timing of radiation therapy. Among the eight patients that progressed prior to irradiation, seven progressed with a component of failure in the posterior fossa or primary site. Among the 18 patients who progressed after irradiation, the initial sites of recurrence in 15 were either in the frontal lobes or the spinal compartment below C2/3. There was only one recurrence in the posterior fossa and there were no recurrences in the volume that received a dose in excess of 12–15 Gy (cumulative) during the 6 week course of the treatment. It has been concluded that, despite the neuraxis failures, the addition of postoperative conformal radiation therapy limited to the posterior fossa and primary site appeared to be a promising treatment for nonmetastatic medulloblastoma in very young children (Fig. 7.3).

Critical to the success of this approach were the follow-up psychology data that suggested no significant or consistent decline in either cognitive or motor functioning as measured by either the phone-based interview technique or formal neuropsychologic assessments that could be performed after the delivery of chemotherapy and local conformal RT [11]. The approach of irradiating only the tumor bed in very young children has been carried forward by a number of groups and will likely be continued in the next series of cooperative group studies for intermediate- or higher-risk patients (unfavorable biology or residual or metastatic disease). Future studies will limit irradiation to the primary site and not include the posterior fossa as the initial volume (Fig. 7.4).

The past results for the treatment of ependymoma have been dismal with very little changed until 20 years ago with the advent of conformal radiation therapy (Table 7.4). Conformal radiation therapy was first introduced for very young children with this disease in the St. Jude RT1 protocol [12]. For the first time, children under the age of 3 years were offered immediate postoperative radiation therapy. The disease control rates increased significantly, especially for children who underwent gross-total resection prior to irradiation.

Radiation therapy for ependymoma can be very complex because of the configuration of the original tumor, postoperative changes, and the intimate association with vital structures. For example, these tumors can extend through the foramen magnum onto the surface of the upper cervical cord; through the foramen of Luschka to involve cranial nerves, the brainstem, and the basilar and vertebral arteries; or through the foramen of Magendie posteriorly into the subarachnoid space of the cisterna magna.

The standard treatment for localized ependymoma includes maximum safe resection followed by conformal or intensity-modulated radiotherapy. The extent of surgical resection is consistently the most important prognostic factor for these tumors. A 5-year survival for patients who have undergone a GTR is reported at 75–93 % compared to 22–52.4 % for those who received less than a GTR [13–15]. Radiation therapy has also demonstrated a survival advantage for these patients, with adjuvant radiotherapy resulting in a 5-year survival of 63 % compared to 23 % without radiotherapy in one series [15, 16]. Very often, disease control reports for ependymoma include patients with supratentorial tumors; however, in the recent publication by Merchant et al. [13], patients with infratentorial ependymoma appeared to have a poorer outcome than those with supratentorial ependymoma. This difference was not statistically significant.

The impact of histologic grade has been reported as controversial in the past; however, a report of 50 patients with localized ependymoma from St. Jude demonstrates a dramatic difference in a group of patients who received adjuvant radiotherapy [17]. The estimated 3-year progression-free survival rate was 28 % for patients with anaplastic ependymoma compared to 84 % for patients with differentiated ependymoma. Merchant et al. [17] provided further evidence from a prospective trial of conformal radiotherapy for resected ependymoma in which tumor grade had a significant impact on both EFS and overall survival (OS). A literature review of 1,444 pediatric patients with posterior fossa ependymoma suggests grade (WHO grade II vs. III) is an independent prognostic indicator for EFS but may not impact OS [18].

Craniospinal irradiation was once used prophylactically for ependymoma because of possible CNS dissemination; however, reviews demonstrated no benefit, even for anaplastic tumors. In fact, the most common site of recurrence is local. The best results in the literature come from the St. Jude prospective trial of conformal radiotherapy using a 10 mm margin which included patients 12 months of age or greater. The 7-year local control, EFS, and OS were 87.3, 69, and 81 %, respectively. The rate of gross-total resection in the study was 82 %, and the majority of patients had 59.4 Gy prescribed to their operative bed (children less than 18 months of age received 54 Gy).

The conformity of irradiation is key to decreasing long-term neurocognitive side effects and making radiation a safe option for young children. The percent volume of the supratentorial brain receiving radiation doses between 0 and 20 Gy, 20 and 40 Gy, and 40 and 65 Gy had a significant relationship with postirradiation IQ [19]. A prospective trial from St. Jude demonstrated that neurocognitive outcomes remained stable and within normal limits post radiotherapy for children receiving conformal adjuvant radiotherapy [12]. It is important to note that children less than 3 years of age at the time of conformal radiotherapy had significantly lower neurocognitive scores at baseline but demonstrated improvement after radiotherapy. The ongoing COG trial now utilizes a 5 mm margin for the CTV, thus reducing normal tissue exposure to radiotherapy.

The role of adjuvant chemotherapy for ependymoma is an ongoing study question. Two cooperative group trials have utilized chemotherapy after craniospinal irradiation and found no clear benefit from the chemotherapy [20, 21]. A prospective cooperative group study randomized patients after postoperative CSI to receive lomustine, vincristine, and prednisone versus observation and found no benefit from the addition of chemotherapy for failure-free or OS [20]. A second prospective trial randomized patients to two adjuvant chemotherapy regiments: lomustine, vincristine, and prednisone versus the eight drugs in a 1-day chemotherapy regimen and found no difference between the regimens [21]. There was no difference between the two chemotherapy regimens and no benefit from radiotherapy alone when compared to other series. Seventy-one percent of the relapses on this trial were isolated local relapses, and only 47 % of patients received a gross-total resection. The impact of chemotherapy after more aggressive surgery and radiotherapy is not fully known.

Omitting radiotherapy from the treatment regimen of children with ependymoma has been an objective of two recent trials based on clinical factors that predict the low risk of failure even in the absence of radiation therapy. The recently completed COG ACNS0121 trial [22] and the current COG ACNS0831 trial [23] both include observation of children with supratentorial differentiated (WHO grade II) ependymoma after microscopically complete resection. These trials require central review of the neuropathology and written confirmation from the operating neurosurgeon that a microscope was used and that no visible tumor cells were present at the completion of surgery. That patients with supratentorial anaplastic ependymoma were not eligible for observation was based on data suggesting a higher rate of failure after radiation therapy of children with supratentorial anaplastic ependymoma regardless of the extent of resection [17]. Studies purely evaluating observations about the absence of chemotherapy have not been performed in children. Koshy et al. [24] reported using the SEER registry that among 804 patients diagnosed with intracranial ependymoma between 1988 and 2005, postoperative radiation was administered only to 35 % of patients younger than 3 years and that among children younger than 3 years, the 3-year OS was significantly greater among those who received postoperative radiation compared to those who did not (81 % vs. 56 %, respectively, p = 0.005). Indeed, RT is required for adult patients according to the work by Rogers et al. [25].

Brainstem glioma is a diffuse intrinsic tumor involving the pons. It has a characteristic appearance on MRI and does not require pathological confirmation. It is uniformly fatal despite excellent initial responses to irradiation; therefore, radiotherapy is considered palliative. Radiation dose escalation increases toxicity but does not improve outcome. Radiosensitizers have not improved the therapeutic ratio, and now biologics are combined with radiotherapy in clinical trials. Brainstem glioma (BSG) or diffuse infiltrating pontine glioma (DIPG) as it is also known is a tumor that extends along neural tracts to involve adjacent regions of the brain. The full extent of the tumor at diagnosis may not be fully appreciated but is readily apparent when it progresses after irradiation. When radiation therapy planning is performed, CT is used for dose calculation and verification. T2-weighted MRI is registered to the CT dataset for the purpose of targeting. A margin of security surrounding the obvious tumor is necessary to account for subclinical microscopic disease or extension along neural tracts, and an additional margin is subsequently added to account for variability in patient setup for daily (fractionated) radiation therapy. The MRI-visible tumor is called the gross tumor volume (GTV), the margin surrounding the gross tumor volume to account for subclinical microscopic disease and unappreciated tumor extension is called the clinical target volume (CTV), and the margin added to the CTV is called the planning target volume (PTV). The CTV margin used in clinical trials has been arbitrarily chosen because pattern-of-failure data are not available to support a specific margin (Table 7.5).

To make progress and compare results among treatment regimens require uniform and consistent guidelines for irradiation. Because there is no agreement on target volume definitions for conformal treatment planning and some suggest that the first site of progression is at or beyond the targeted margin, the patterns of failure with respect to the targeted volume are currently under evaluation. T2-weighted MR imaging can be used to demonstrate treatment in almost all cases. Brainstem glioma extends along neuronal tracts in all dimensions including superiorly into the cerebral peduncles and thalamus, axially with posterior extension into the cerebellar peduncles, and in a limited manner into the cervical spinal cord. In one study [26], the average distances from the contoured GTV were superior, 1.1 cm with a range of 0.3–2.75 cm; inferior, 0.4 cm with a range of 0.1–0.6 cm; and lateral, 1.1 cm with a range of 0.3–2.0 cm. It was then recommended that the GTV should include identifiable tumor on post-contrast T1- and T2-weighted MRI. The CTV should include a 2 cm anatomical expansion in the transverse and inferior dimensions and a 3 cm anatomical expansion in the superior dimension. Tumors confined to the pons should have the CTV expanded superiorly to cover the midbrain. PTV margins should be individualized (0.3–0.5 cm) but not less than the MR image section thickness used for treatment planning. Treatment planning MRI should include the upper cervical spinal cord. In summary, the “required” CTV margin used in the treatment of BSG will vary according to the location and volume of the GTV and access to specific neural pathways. Properly encompassing the tumor and subclinical microscopic disease will help patients to benefit most from radiation therapy while considering normal tissue effects which are often dose related.