Clinical Characteristics

Brainstem gliomas account for up to 300 cases of pediatric brain tumors in the United States per year, and represent a heterogeneous group of lesions. These lesions are typically classified based on anatomic location and radiographic appearance. They may arise in the midbrain, pons, medulla, tectum, or cervicomedullary junction. They can be categorized as diffuse or focal, and intrinsic or exophytic. Up to 75% of pediatric brainstem gliomas are diffuse intrinsic pontine gliomas (DIPGs).1,2 Tumor extension into the cerebellar peduncles, cerebellum, cerebrum, spinal cord, and leptomeninges has been described.3

The clinical features of pediatric brainstem gliomas are dependent on tumor morphology and anatomic location. Focal tumors typically arise in the tectum, midbrain, dorsal brainstem, or cervicomedullary junction, and often exhibit insidious onset and localizing symptoms. Exophytic lesions may cause symptoms of mass effect including cranial nerve deficit or contralateral hemiparesis, as well as hydrocephalus leading to headache, nausea/emesis, and lethargy. Extremity weakness and myelopathy may also be detected in patients with cervicomedullary lesions, whereas tectal lesions can cause ocular disturbance and aqueductal obstruction, resulting in hydrocephalus. Intrinsic infiltrative tumors typically present with a shorter clinical course (usually 2 to 6 months) of symptoms including long tract signs, cerebellar ataxia, dysmetria, and cranial nerve deficits. The most common cranial nerves affected are VI and VII, and may be unilateral or bilateral.4,5

The mean age at diagnosis of brainstem gliomas is 6 to 9 years, with no gender predilection. Tectal gliomas typically have a more indolent course, whereas DIPGs are frequently aggressive, carrying a median survival of 10 months with up to 90% mortality within 18 months. DIPG lesions are often large at the time of clinical presentation, with most children exhibiting tumors over 2 cm in size at initial radiographic diagnosis.1,6,7 Patients with neurofibromatosis type 1 (NF-1) have an increased frequency of brainstem gliomas, which are typically low grade.8 Brainstem gliomas are rarely seen in adults,5 and rare cases of infantile diffuse intrinsic brainstem glioma have been described in the literature9 ( Table 33.1 ).

Radiographic Findings

Brainstem gliomas are typically classified on the basis of radiographic appearance on magnetic resonance imaging (MRI). Several studies have been published on the radiographic appearance of brainstem gliomas, culminating in a classification scheme for brainstem gliomas based on computed tomography (CT) and MRI appearance ( Table 33.2 ).6,10–13 The differential diagnosis of a child with a brainstem lesion includes glioma, atypical teratoid/rhabdoid tumor, primitive neuroectodermal tumor (PNET), hemangioblastoma, demyelinating disease, and focal brainstem encephalitis.

In general, brainstem gliomas may be classified as diffuse, focal (intrinsic or exophytic), or cervicomedullary.6 DIPG is characterized by expansile hypertrophy and diffuse infiltration of the ventral pons (basis pontis), with most lesions measuring over 2 cm at the time of presentation. DIPG appears hypointense to brain on T1-weighted MRI, and hyperintense on T2-weighted MRI, with indistinct margins and often minimal to no enhancement with gadolinium administration ( Fig. 33.1 ), although some can display variable contrast enhancement. Perilesional edema, although rare, may be visualized as increased signal on fluid-attenuated inversion recovery (FLAIR) sequence. Regions of cystic necrosis and extension along white matter tracts into the cerebellum, midbrain, cerebrum, medulla, or cervical spinal cord may also occur. Cases of tumor dissemination with leptomeningeal enhancement, particularly at late stages of disease, have been reported.14,15 Diagnosis of DIPG can often be made based on radiographic appearance.

In contrast, focal brainstem gliomas are typically well-defined lesions of the dorsal pons (pontine tegmentum), midbrain, or medulla, which may be exophytic. They can have homogeneous to heterogeneous contrast enhancement ( Fig. 33.2 ). They are most often hypointense on T1-weighted imaging and hyperintense on T2-weighted imaging. Fluorodeoxyglucose positron emission tomography (FDG-PET) may be useful for differentiating low- and high-grade gliomas,16–18 whereas diffusion tensor imaging (DTI) can reveal degeneration of motor and sensory white matter tracts and play a role in determining tumor extension19–21 ( Table 33.3 ).

Pathological Features

Knowledge of the histology and molecular characteristics of DIPG has been limited due to the lack of access to tissue, with current data arising mainly from small case series and autopsy data. On histopathological examination, DIPGs are typically high-grade anaplastic astrocytomas (World Health Organization [WHO] grade III) or glioblastoma multiforme (WHO grade IV). Characteristic histopathological findings of high-grade glioma may be seen in DIPG, including cellular atypia, mitotic figures, pseudopalisading necrosis, and microvascular proliferation ( Fig. 33.3 ). Rare diagnoses of well-differentiated diffuse pontine astrocytomas (WHO grade II) have been reported. Subgroup analysis in one case series of 134 DIPG patients showed that prognosis is independent of tumor grade. Dissemination within the neuraxis with positive cerebrospinal fluid (CSF) cytology also may be observed in the setting of advanced progressive disease.14 In contrast, focal brainstem gliomas are typically low-grade, pilocytic astrocytomas (WHO grade I). Ganglioglioma (WHO grade I) and diffuse astrocytoma (WHO grade II) of the brainstem have also been reported22,23 ( Table 33.4 ).

Diagnosis

Due to the anatomic location and frequently aggressive nature of pediatric brainstem gliomas, diagnosis and treatment is often a significant clinical challenge. Barkovich et al11 first defined radiographic criteria for diagnosis and established the high sensitivity and positive predictive value of MRI for brainstem gliomas. In 1993, Albright′s group24,25 reviewed the Children′s Cancer Group experience with diagnosis and management of patients with DIPG, and questioned the need for diagnostic biopsy given its characteristic appearance on radiographic imaging. In a multicenter review, Schumacher et al26 compared MRI with biopsy results in children with brainstem tumors. This study demonstrated that with specific imaging criteria, coupled with clinical history and laboratory data, a diagnosis of brainstem tumor could be made with 94% sensitivity and 43% specificity, with a positive predictive value of 96%. In cases of brainstem glioma, radiographic determination correlated with WHO grade in up to 74% of cases. As a result, it is argued that diagnosis of pediatric brainstem glioma can be made on the basis of imaging alone, with diagnostic biopsies only performed for atypical lesions.

However, recent studies have suggested that MRI cannot predict prognosis in children with DIPG without correlation to histological findings or treatment response.27 Further, with the implementation of stereotactic techniques, the utility of diagnostic biopsy for pediatric brainstem glioma has once again come into question. Roujeau et al28 reported a series of 24 children undergoing stereotactic biopsy of brainstem lesions, with no deaths and minimal morbidity, consisting of transient focal neurologic deficit, including extremity weakness and cranial nerve palsy. They argue that histological data afforded by biopsy can, in some cases, direct therapy. Pincus and colleagues29 report a single-center series of 10 children with atypical brainstem lesions in which stereotactic biopsy yielded a 100% diagnostic rate. They compared these findings to a literature review consisting of 157 children undergoing stereotactic brainstem biopsy, which demonstrated a diagnostic rate of 93.6%, with 5.1% morbidity and 0.5% mortality. A separate review of the literature consisting of 378 children undergoing stereotactic biopsy of brainstem lesions found a diagnostic rate of 94.8%, with 0.5% mortality and 6.6% morbidity, 1.5% of which was permanent.30

Although controversy persists over whether to perform diagnostic biopsies for children with brainstem gliomas, diagnosis based on radiographic evidence with implementation of biopsy only in the case of atypical lesions is currently the established paradigm. Hankinson and colleagues7 demonstrate a lack of consistency about which brainstem lesions pediatric neurosurgeons would define as “typical” or “atypical” based on MRI appearance, or which lesions they would consider for biopsy. With the advent of emerging molecular data suggesting the presence of possible tumor subtypes, the utility of biopsy in children with brainstem gliomas continues to be a topic of debate ( Table 33.5 ).