Although brainstem tumors account for 10%–20% of cerebral neoplasms in children and 90% of them are gliomas [1, 2], exophytic pontine glioma located in the cerebello-pontine angle (CPA) is uncommon [3–6]. Because of their heterogeneous biological behaviors and neurological imaging manifestations, the classification of brainstem gliomas is not consistent. Diagnosis and therapeutic decision-making are often difficult for gliomas, especially in the case of rare neurological imaging features.

Here, we report a special case of multifocal glioblastoma derived from the pons, exhibiting an exophytic growth pattern, in a 5-year-old girl. In total, three lesions were revealed in the posterior fossa. Repeated magnetic resonance imaging (MRI) showed the progression of the lesions in the left side of the CPA and the upper left side of the pons as well as the appearance of a new lesion in the right side of the pons. Furthermore, repeated magnetic resonance spectroscopy (MRS) over 3 months revealed a gradually elevated ratio of choline (Cho) to N-acetyl aspartate (NAA).

The preoperative diagnosis of this case was difficult. Brain metastases, central nervous system (CNS) lymphomas, and inflammatory disease were considered in the differential diagnosis. It is notable that the patient presented with frequent fever and cough and had weak positivity for blood TB antibodies. Infectious diseases, such as tuberculosis, were considered. Although previous articles mentioned that ring-like enhancement and multiple lesions are indications of posterior fossa glioblastoma [7], these neurological imaging features are frequently observed in tuberculoma and tuberculous lesions that are occasionally secondary to tuberculous arachnoiditis [8]. Surgical resection or biopsy is necessary for a definitive diagnosis. The histological grading is based on the degree of anaplasia and dedifferentiation, number of mitoses, capillary endothelial proliferation, and presence of necrosis [9–11].

Immunohistochemical examination showed diffuse expression of GFAP, P53, and S-100 in our case. The expression of S-100 is not stable in glioma, whereas GFAP is a qualitative diagnostic marker of glioma [12]. The combination of all histopathologic evidence supported the diagnosis of glioblastoma in our case.

The progressive growth pattern of the CPA lesion is a confusing characteristic of this case, especially given the coexistence of multiple lesions. We searched PubMed and Google Scholar and found only two English-language articles reporting the coexistence of multiple lesions and a CPA lesion [13, 14]. According to the literature, CPA glioblastomas may have different origins [15]. Wu et al. [16] reported a CPA glioblastoma that originated from the cranial VIII nerve and reviewed seven other similar cases. Cranial nerve glioblastomas are presumably originated from CNS tissue locating at the root entry zone of the cranial nerve or from heterotopic neuroglia cells in the leptomeninges. Similarly, Mabray et al. [17] recently reported eight cases of glioblastoma originated from the cranial nerve. These glioblastomas were located at the CPA and involved the unilateral brainstem. The cerebellum is another unusual place for glioblastoma development. Matsuda et al. [18] reported a case of glioblastoma originated from the cerebellum and reviewed another three similar cases. In the reports by both Mabray et al. [17] and Matsuda et al. [18], the relatively clear boundary between the brainstem and tumor were demonstrated during the operation. In our case, neither preoperative MRI nor surgery alone supported an origin at the cranial nerve cerebellum. However, no clear boundary between the CPA lesion and the pons was discovered during our surgical procedure. This evidence, along with MRI findings on the CPA lesions and two pontine lesions, confirmed the diagnosis of primary exophytic pontine glioblastoma. The T2-weighted image showed that the CPA lesion was confined to the pons, with an outer layer that consisted of edematous tumor tissue infiltrating the pons. This finding may explain that the pathologic presentation of the outer layer of the tumor was not as typical for glioblastoma as that of the inner part of the tumor, emphasizing the diverse pathologic features of different parts of the tumor and the need for accurate and comprehensive specimen collection in such a case. As reported earlier, the biopsy samples should be collected from solid, non-necrotic, and gadolinium-enhanced areas, if possible [19].

The changes in the MRS results of our case merit attention. It was reported that the Cho:NAA ratio calculated according to MRS data can be used to monitor metabolism and glioma behavior [22]. In pediatric CNS tumors, changes in levels of metabolic markers as determined with MRS are associated with tumor grade [27]. Furthermore, an increasing Cho:NAA ratio predicts a short overall survival [22, 28]. PET measures the regional cerebral glucose metabolism, offering valuable information that can be used to differentiate between low-grade and high-grade gliomas [29–31]. For our case, the initial MRS results (obtained on October 8, 2015) showed a simultaneous decrease in NAA and Cho and an increased Cho:NAA ratio, leading to the diagnosis of low-grade glioma. The decrease in NAA, Cho, and Cho:NAA ratio of the first MRS may be technique variations such as the selection bias of regions of interests. Taking both MRS and PET results into consideration, a diagnosis of low-grade glioma seems reasonable. Earlier PET findings from another medical center were consistent with our initial MRS results. Repeated MRS performed 2 and 3 months later demonstrated an obvious decrease in NAA and an increase in Cho, indicating a high-grade glioma. In addition, we noticed peaks at approximately 1.3 ppm (Figs. 1c, 2c, f). Two situations were considered: (1) technical problems such as unstable baseline or lipid effection and (2) a lac-lip peak demonstrating necrosis or hypoxia. Although necrosis is usually found in high-grade glioma, PET–CT results and the increasing tendency of the Cho:NAA ratio both support the initial diagnosis of low-grade glioma. The MRS characteristics of this tumor seemed to be significantly altered within merely 2 months according to consecutive neurological imaging.

Glioblastomas are divided into two types of genesis patterns: primary glioblastoma, also termed de novo glioblastoma, which is defined as a fully developed tumor with no clinical, radiological, or histopathologic evidence of relative benign precursor lesion [32–34], and secondary glioblastoma progressing from low-grade astrocytoma, as supported by imaging or histological evidence [32, 33, 35]. According to the literature, the proportions of primary and secondary glioblastomas are approximately 90% and 10%, respectively [36]; the disease course of primary glioblastomas was less than 3 months, whereas the progression from relative low-grade astrocytoma to glioblastoma took approximately 2–5 years [37]. Nevertheless, rapidly progressive cases of secondary glioblastoma have been reported [34]. Furthermore, as for brainstem gliomas, it has occasionally been reported that low-grade gliomas behaved as aggressively as high-grade gliomas [38, 39]. Recent studies have revealed the histological heterogeneity of DIPG. Based on the autopsy results of the patients diagnosed with DIPG, It was reported that the primary and metastatic lesions displayed high and low histological grades, respectively [40, 41]. Accordingly, DIPG was classified as primary lesions, contiguous lesions, and metastatic lesions according to the locations [41]. The different pathologic manifestations of the two samples from our case may be related to the differences of “central” and “marginal” areas of the lesion. This may again emphasize the importance of accurate sampling.

Immunohistochemical examination of our case demonstrated a high positive rate of P53 (80%–90%), which is a frequently expressed marker of DIPG [42]. It was also reported that P53 mutation was even more frequently and increasingly expressed in secondary glioblastomas, which was in contrast with its rare expression in primary glioblastoma [34]. However, IDH1, a biomarker of secondary glioblastoma [43], showed low expression in our case. This may highlight the molecular differences between pediatric and adult secondary glioblastomas [44–46]. In children, low expression of IDH1 may be present in secondary glioblastomas. According to the literature, histological or imaging evidence of a progression from a less malignant astrocytoma can be used as a diagnostic criterion of secondary glioblastomas [34]. Based on this knowledge and results of repeated MRI (MRS), in combination with initial PET–CT findings, we were inclined to diagnose our case as a rapidly progressive secondary glioblastoma.

Multiple gadolinium-enhanced lesions are rarely seen in brainstem gliomas [47–49]. Multiple-lesion gliomas involve at least two enhanced lesions and account for 2%–9% of all gliomas [50–52]. This rare entity is classified into two distinct types, multifocal and multicentric gliomas. Multifocal gliomas possess homology or dissemination of definite pathways. Multicentric gliomas are defined as anatomically isolated lesions and are usually distributed at different lobes or hemispheres of the brain [51]. Multiple lesions can be present at the diagnosis or at a progressive stage for both multicentric and multifocal gliomas [53]. In our case, the right pontine lesion was developed at a later stage as compared with the other two lesions, and all lesions were located at the diffusely swelling pons, indicating a diagnosis of multifocal glioblastoma rather than multicentric glioblastoma.

Diffuse pontine glioma is the most common type of brainstem tumor and carries the poorest prognosis in children [23]. Early surgical intervention to determine the diagnosis is recommended [54–56]. The standard treatment of DIPG is external beam radiation therapy administered in fractions over approximately 6 weeks at a total dose of 60 Gy [54]. Diffuse intrinsic pontine glioma is almost invariably fatal with a mean patient overall survival of 9–12 months from the time of diagnosis [57]. Radiation therapy is the major treatment [58]. Adjuvant therapies, such as chemotherapy, targeted therapy, differentiation agents, and radiation sensitizers, have been studied. Although some of these therapies showed a beneficial impact on patient outcomes [25, 58], further research is required.

In summary, we presented a rare case of progressive, exophytic, multifocal pontine glioblastoma in a 5-year-old girl. A preoperative differential diagnosis was difficult because of the atypical neurological imaging features and frequent fever and cough symptoms. The clinical evidence supported a secondary glioblastoma. Regretfully, an early histopathologic diagnosis was not achieved.