Brain Pathology Case of the Month - November 2015


Glioblastoma with primitive neuroectodermal (PNET) features and metastasis of the PNET component


Glioblastoma (GBM) is the most common malignant primary central nervous system tumor and carries an overall poor prognosis with a median survival ranging from 12-15 months (Stupp et al., 2009). GBM is a primarily astrocytic tumor, however, there are emerging variants with unique cellular and molecular features (Tanaka et al., 2013). GBM with primitive neuroectodermal features (GB-PNET) is one of the less well-characterized variants with unusual morphology, behavior and unknown cell of origin. Within a GB-PNET tumor, GBM-like areas express glial fibrillary acid protein (GFAP) in the absence of synaptophysin, while PNET-like areas express the neuronal markers synaptophysin and NeuN, but while lacking GFAP expression (Karsy et al., 2012). The variant GB-PNET tumor has distinct architectural and molecular findings when compared with either PNET or GBM alone.

This case illustrates several features of this tumor entity: 1) The uniquely poor prognosis despite multimodal therapies tailored both to the treatment of GBM and PNET, 2) the highly invasive and metastatic tumor behavior driven by the PNET component, 3) the unusual treatment resistance and poor outcome despite IDH-1 positive mutation status of the glial component, and 4) and the unique molecular features of the metastatic tumor with presence of N-MYC amplification in tumor cells with neuronal features, but not in glial tumor cells.

Notably, it is unknown whether positive prognostic factors known to be relevant in GBM-MGMT promoter DNA methylation (Hegi et al., 2005) or isocitrate dehydrogenase (IDH) 1/2 mutation (Parsons et al., 2008) are also applicable to the behavior and overall outcome in variant GB-PNET. Glioblastoma patients with positive IDH-1 mutation and MGMT promoter methylation status are considered to carry a more favorable prognosis, which appears not to hold true in this case. It has been reported that positive IDH1 mutation confers a prognostic advantage in GB-PNET, although the case series in which this was described consisted of only 10 patients (only 2 had positive IDH1 mutations) (Song et al., 2011). One series of 40 patients reported a median survival of patients with GB-PNET of 24 months (Varlet et al., 2004). The prognostic significance of MYC gene amplification in GBM was not assessed. Unfortunately, our patient survived approximately 5 months despite aggressive treatment, which was attributed to rapid tumor cell proliferation, metastatic spread, and recurrence of the PNET-like component.

In one of the largest series of GB-PNET to date, Perry et al. (Perry et al., 2009) presented 53 cases of GB-PNET with a median survival of 9.1 months, somewhat shorter than the survival of patients with classical GBM. Interestingly, Perry et al. had one case with rapid tumor recurrence that showed only the PNET-like histopathology on re-resection, much like the case presented here. We hypothesize that the recurrence in both cases may be related to the presence of N-myc amplification, possibly resulting in high cell proliferation and extensive subarachnoid spread. In malignant gliomas with a PNET component, the primitive areas contain C-myc or N-myc amplification in 43% of cases (Perry et al., 2009). Additionally, this patient did not have gross-total resection, and gross-total resection has been reported as an important prognostic factor predicting survival (Varlet et al., 2004).

This case suggests a propensity of GB-PNET to ultimately behave more like PNET despite a previously proposed hypothesis that the PNET component secondarily develops from a primary glial tumor (Joseph et al., 2013; Perry et al., 2009). Tumor cell infiltration into adjacent tissues poses a major therapeutic challenge for the management of these tumors.

The optimal treatment of this rare variant is still unknown. Current GB-PNET therapies are based on treatment of GBM and pediatric medulloblastomas. However, medulloblastoma is known to be a distinct molecular entity. Our patient underwent surgical debulking followed by standard therapy for glioblastoma. At first recurrence, however, and based on the predominant recurrence of the PNET component, he was treated with a regimen of cyclophosphamide, cisplatin and vincristine, following guidelines for the management of childhood PNETs. The GBM component of his tumor appears to have been treated sufficiently based on the lack of solid GBM at the time of first recurrence and at autopsy. Temozolomide-based treatment is administered to the majority of patients with newly diagnosed GB-PNET, however, it has been suggested that given the frequency of cerebrospinal fluid (CSF) spread of the PNET-like component, platinum-based chemotherapy might be a reasonable option to consider up-front (Perry et al., 2009). Gross total surgical resection appears to be one of the most important interventions to improve prognosis (Varlet et al., 2004). Further research towards achieving a standard protocol for treatment of these unusually aggressive tumor variants, albeit rare, is urgently needed.


Contributed by Jamie Nicole LaBuzetta, Daniel Mordes, Isabel Arrillaga, Andrew Chi, Jorg Dietrich

International Society of Neuropathology