Pediatric Brain Tumors: Target Delineation

Pediatric brain tumor target delineation starts with staging, immobilization, and anatomy, not with beam arrangement. In the chapter devoted to this topic, the authors build the workflow around three diseases that genuinely change dose, field design, and boost strategy: medulloblastoma, ependymoma, and pure germinoma.

That is why this chapter is useful at the console. It keeps pulling the reader back to the same operational questions: what is the true disease extent, which structures must stay inside the volume, and where should the margin stop. For the broader series, return to the Target Volume Delineation and Field Setup – Complete Clinical Guide; this article stays with the pediatric CNS deep dive.

Medulloblastoma in pediatric brain tumors

For medulloblastoma, the decision that drives both dose and volume is risk grouping. The chapter classifies patients as high risk when CSF cytology is positive, gross metastases are present, or postoperative MRI shows at least 1.5 cm² of residual disease in the tumor bed. Standard-risk disease means no spread, M0 staging, and less than 1.5 cm² of residual tumor bed disease.

The staging package is deliberately comprehensive. Every patient needs a full history and physical examination, preoperative and postoperative contrast-enhanced brain MRI with 1 to 3 mm slices, contrast-enhanced spine MRI, and CSF sampling to rule out dissemination. Without that, the later choice between standard-risk CSI, high-risk CSI, focal boost, and posterior fossa coverage becomes guesswork.

Simulation is handled with the same level of seriousness. The chapter recommends a reproducible setup, often with full-body Vac-Lok or alpha cradle immobilization plus a standard brain mask carrying multiple marks for triangulation and longitudinal spinal alignment. Variable CT slice thickness can be useful, with thinner cuts through the brain for detailed contouring and thicker cuts through the spine to keep the contouring burden realistic. The scan should include all immobilization devices, the vertex superiorly, and the gonads inferiorly.

One of the most practical sections deals with anesthesia. If the child starts treatment intubated but is expected to transition later to a laryngeal mask airway or even nasal cannula, the authors suggest incorporating an oral airway when the mask is made so chin position remains stable and a second simulation can be avoided. They are equally firm about contouring every target and organ at risk on every CT slice. If your department still spends time reconciling inconsistent OAR names between physicians and planners, our article on structure standardization in radiotherapy (TG-263) is a useful companion before building these pediatric sets.

Craniospinal irradiation and risk-adapted dosing

The neuraxis volume is defined from the dura and CSF space outward, not from field borders inward. In a growing child, the CTV spans the entire vertebral body and canal; in a fully grown patient, the canal alone is sufficient. That distinction matters because the chapter treats skeletal growth as a planning issue, not a footnote.

Table 34.1: target volumes for the craniospinal phase

The chapter separates craniospinal irradiation from the posterior fossa boost. This table condenses the volume definitions used for the full neuraxis treatment phase.

Target volume	Definition and description
GTV	Tumor bed with all gross residual disease, the walls of the resection cavity seen on MRI, and areas of concern identified by the neurosurgeon. Surgical defects that only reflect the operative path are not included. Any gross spinal disease should also be contoured if a boost is being considered.
CTVCSI	The entire volume enclosed by the dura and in contact with cerebrospinal fluid, including any postoperative pseudomeningocele. In a growing child, this CTV includes the vertebral body and canal; in a fully grown patient, it includes the entire canal.
PTVCSI	CTVCSI plus 3 to 10 mm, depending on daily setup confidence and institutional experience.

Source: Target Volume Delineation and Field Setup, 2nd Edition (Table 34.1)

The proton discussion is more nuanced than many quick summaries imply. Some groups have recommended treating the full vertebral body to 30 Gy when CSI reaches 36 Gy in a growing child, but the chapter points out that evidence remains limited and cites the ongoing trial NCT03281889 as the key effort testing whether deliberate whole vertebral body coverage is truly necessary. Many clinicians instead include bone within the PTV without expanding further, mainly to avoid intentional dose to lung and esophagus.

The inferior limit of the thecal sac is another planning detail with real consequences. It is often, but not always, located at S2. If a posterior photon field extends lower than necessary, gonadal exit dose rises. That is not a proton issue, but it absolutely is a photon issue.

For standard-risk disease, the chapter is explicit: the 18 Gy CSI arm on COG ACNS0331 had more failures, and 23.4 Gy remains the standard of care. The same trial showed that involved-field boost was equivalent to whole posterior fossa irradiation in standard-risk patients. The authors therefore recommend 23.4 Gy to the entire craniospinal axis followed by an involved-field boost to 54 Gy.

Tumor bed boost versus whole posterior fossa boost

Once the CSI phase is complete, the planning logic becomes more anatomical. The recommended margin for the postoperative tumor bed is 1 to 1.5 cm, but only inside plausible pathways of spread. The skull and tentorium act as natural barriers. Brainstem invasion can occur, so the authors advise including 2 to 3 mm of brainstem in the CTV when the tumor contacts the brainstem. If there was no contact on preoperative imaging or at surgery, the brainstem can be excluded. For the PTV, they recommend 3 to 5 mm, while noting that their own practice uses 3 mm with daily image guidance.

Table 34.2: tumor bed boost within the posterior fossa

After CSI, the book shifts to the focused tumor bed boost. The main theme is simple: keep the margin generous enough for microscopic spread but stop where anatomy makes further extension biologically implausible.

Target volume	Definition and description
GTV	Tumor bed with all gross residual disease, the walls of the resection cavity defined on MRI, and areas of concern outlined by the neurosurgeon. The surgical route itself is not part of the cavity if it was not initially involved by tumor. Gross spinal disease should be outlined separately if a boost is under consideration.
CTVtbboost	GTV plus a 1 to 1.5 cm anatomically confined margin. The expansion should stop at barriers to spread, such as the tentorium, and should limit brainstem inclusion to 2 to 3 mm where the tumor contacts the brainstem.
PTVtbboost	CTVtbboost plus 3 to 5 mm, based on daily imaging and institutional experience.

Source: Target Volume Delineation and Field Setup, 2nd Edition (Table 34.2)

The metastatic scenarios are just as clearly separated. Patients with M2 disease, meaning intracranial subarachnoid spread, may receive boosts to 54 Gy for supratentorial or posterior fossa metastatic sites. M3 disease is split into focal spinal deposits and diffuse spinal dissemination. Diffuse disease means radiographically visible lesions in at least three of the four spinal regions: cervical, thoracic, lumbar, or sacral. On COG ACNS0332, diffuse spinal disease was prescribed 39.6 Gy, focal disease above the cord 45 Gy, and focal disease below the cord 50.4 Gy.

For high-risk disease, or for patients who will not receive chemotherapy, the chapter recommends 36 Gy to the craniospinal axis with a boost to 54 Gy. The authors acknowledge that retrospective series support tumor bed-only boost even in high-risk patients without excess posterior fossa failures outside the bed, but they also state plainly that randomized confirmation is still missing.

If the whole posterior fossa is selected for boost, the target becomes broader and less forgiving. Everything below the tentorium is included, the anterior border extends to the posterior clinoids, and the entire brainstem belongs inside the posterior fossa CTV.

Table 34.3: entire posterior fossa boost volumes

Some patients still need a whole posterior fossa boost rather than a tumor bed-only boost. The table below shows the exact boundaries used for that larger volume.

Target volume	Definition and description
GTV	Tumor bed with all gross residual disease, the walls of the resection cavity seen on MRI, and neurosurgeon-identified areas of concern. The operative corridor is not part of the volume when it was not involved by tumor. Any gross spinal disease may be contoured separately for a boost.
CTVpf	The entire posterior fossa. The whole brainstem is included. The borders are the skull base anteriorly, the tentorium superiorly, and the foramen magnum inferiorly; skull bone constrains the volume posteriorly and laterally.
PTVpf	CTVpf plus 3 to 5 mm, depending on daily imaging and institutional experience.

Source: Target Volume Delineation and Field Setup, 2nd Edition (Table 34.3)

Ependymoma in pediatric brain tumors

Intracranial ependymoma begins with surgery because extent of resection remains the dominant prognostic factor. The chapter is direct on that point and goes further: if postoperative MRI shows residual disease, repeat resection should be considered whenever the expected morbidity is acceptable.

The restaging package resembles medulloblastoma but with a different probability profile. Brain MRI before and after surgery, total spine MRI, and detailed clinical evaluation are expected. Unless medically contraindicated, spine MRI and CSF cytology should still be obtained to exclude dissemination, even though intracranial ependymoma spreads at diagnosis in fewer than 10% of patients.

Simulation is CT-based, without contrast, and should use 1 to 3 mm slices to support image fusion, target definition, and organ-at-risk contouring. The scan must include every immobilization device and the entire cervical cord. Many of these children require daily anesthesia, so the authors again stress early coordination with anesthesiology so the mask will reproducibly accommodate the airway support that treatment actually needs.

For the GTV, the chapter warns the reader not to stop at the central cavity. The foramina of Luschka and Magendie deserve special scrutiny because they are often involved. This is exactly where a quick conversation with the surgeon can correct what MRI alone may understate.

Margin reduction, conedown, and cord protection

The modern trend is toward smaller margins. On COG ACNS 0831, patients were treated with CTV equal to GTV plus 0.5 cm to a total dose of 54 Gy in 30 fractions. To limit brainstem toxicity, brainstem expansion was capped at 3 mm. The same trial used a conedown for children older than 18 months to 59.4 Gy, explicitly excluding the entire brainstem, optic chiasm, and cervical cord from the final boost volume.

Table 34.4: recommended volumes for infratentorial ependymoma

For ependymoma, the chapter separates the initial 54 Gy volume from the 59.4 Gy conedown. That distinction is what allows late-phase sparing of the brainstem, optic chiasm, and cervical cord.

Target volume	Definition and description
GTV	Tumor bed with all gross residual disease, the walls of the resection cavity defined on MRI, and areas of concern outlined by the neurosurgeon. Special attention is required for the foramina of Magendie and Luschka.
CTV54	GTV plus 5 to 10 mm, constrained by bone and tentorium. The expansion should not enter the brainstem by more than 3 mm.
CTV59.4	On ACNS 0831, this volume was defined as GTV plus 5 mm while excluding the entire brainstem, optic chiasm, and cervical cord.
PTV54/59.4	CTV plus 3 to 5 mm, depending on daily imaging and institutional experience. The chapter notes that parts of the PTV may be intentionally underdosed when needed to respect cervical cord and optic chiasm tolerance.

Source: Target Volume Delineation and Field Setup, 2nd Edition (Table 34.4)

The authors also acknowledge that many pediatric radiation oncologists still favor larger margins, using CTV equal to GTV plus 1 cm and a total dose of 54 Gy for patients not enrolled on ACNS 0831. They describe that approach as acceptable, which is useful context when institutional practice has not fully converged.

When 59.4 Gy is prescribed, the suggested workflow is two-phase treatment with conedown at 54 Gy to spare additional dose to the brainstem, optic chiasm, and cervical spinal cord. Operationally, the crucial rule is that the 59.4 Gy PTV should not extend below the foramen magnum, regardless of how far inferior the original tumor reaches.

The chapter also gives a hard spinal cord number that is easy to miss. On ACNS 0831, the target cervical spinal cord constraint was D10% of 57 Gy or less. To meet that objective, the spinal cord should receive no more than 70% of dose, or 126 cGy per fraction, during each of the final three fractions of the 59.4 Gy phase. That is the kind of planning detail that usually lives in protocol notes, yet it changes how a conedown is drawn.

Pure germinoma in pediatric brain tumors

For pure germinoma, biology determines field size before geometry does. The chapter again asks for preoperative and postoperative brain MRI, total spine MRI, and full clinical assessment, but it adds mandatory serum and CSF beta-hCG and AFP testing to exclude a nongerminomatous component.

That distinction is not cosmetic. Any AFP elevation leads to treatment as NGGCT. On ACNS 1123, only patients with serum or CSF beta-hCG of 100 mIU/mL or less were treated as germinoma; values above 100 were managed as NGGCT. The chapter also notes that, off protocol in North America, NGGCT is still commonly treated with CSI while more limited fields remain under study.

Localized disease uses standard supine facial mask immobilization with conventional brain triangulation marks. If disseminated disease requires CSI, immobilization returns to the medulloblastoma-style setup. Bifocal germinoma confined to the suprasellar and pineal regions is still treated as localized disease, with whole ventricular irradiation followed by a boost to the original gross disease.

The target definition is one of the clearest practical lessons in the chapter. The prechemotherapy tumor volume must be contoured at the start of planning because the boost often extends outside the normal ventricular anatomy. The target includes the prechemotherapy tumor, any residual disease, and the ventricles. For boost planning, the CTV is the prechemotherapy GTV plus 1 to 1.5 cm. Coverage of the prepontine cistern is optional, but the authors specifically advise considering it after third ventriculostomy and in patients with bulky suprasellar tumors.

If radiation is the only treatment modality, the whole ventricular volume receives 21 to 24 Gy and the boost takes the prechemotherapy volume to 45 Gy. Because prognosis is favorable and long-term neurocognitive toxicity matters, 1.5 Gy fractions are often used, though 1.8 Gy per fraction remains reasonable.

After neoadjuvant chemotherapy with complete response of the primary tumor, the whole ventricle receives 21 Gy and the boost adds 9 to 15 Gy for a total primary dose of 30 to 36 Gy. ACNS 1123 also tested reducing the whole ventricular dose to 18 Gy. There were no ventricular failures among the 74 evaluable patients, but the study did not demonstrate noninferiority against its planned threshold of 95% 3-year progression-free survival. Patients with partial response or progression still require a boost bringing the primary tumor dose to 36 to 45 Gy.

If there is one consistent message across the entire chapter, it is this: 3D conformal therapy, IMRT, VMAT, and proton therapy can all be successful, but none of them rescues a poorly reasoned target. In medulloblastoma, ependymoma, and germinoma alike, the chapter keeps returning to thin imaging, reproducible setup, honest anatomy, and dose selection tied to real risk. For the wider book roadmap, revisit the complete guide.