Brainstem Glioblastoma Multiforme in a Patient with NF1

Discussion

Brainstem gliomas that occur in patients with NF1 are typically low-grade, indolent, and generally have a better prognosis compared to patients with sporadic brainstem gliomas (13-15). Further, most of these lesions are discovered in childhood (1, 4, 16). High-grade astrocytomas, while rare, have been reported in NF1, but are almost always supratentorial (5, 7-9, 12). Two cases of infratentorial GBMs in patients with NF1 have been reported, both occurring in the cerebellum (10, 11).

Molecular analysis of NF1-related low- and high-grade astrocytomas showed NF1 inactivation in all NF1-associated astrocytomas (12). Molecular analysis of NF1-associated GBMs has shown both an NF1 deletion and either CDKN2A loss or homozygous chromosome 10q losses, which are commonly seen mutations in sporadic GBMs. Since both low-grade and high-grade astrocytomas in NF1 have NF1 inactivation, but only the latter has additional tumor markers that are commonly seen in GBMs, it is suspected that their respective tumorigenesis occurs via distinct pathways. Further, it has been proposed that large germ-line deletions of NF1 may be the driving factor that places patients with NF1 at an increased risk of malignant astrocytomas as opposed to low grade gliomas (12).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Brain MRI before and after treatment. A: Three months pre-treatment axial T1 post-contrast. B: Three months pre-treatment axial Flair. C: 5 months after treatment initiation axial T1 post-contrast. D: Five months after treatment initiation axial Flair.

Our case has analogous findings of both NF1 inactivation and loss of the CDKN2A tumor suppressor gene on chromosome 9. Our case had a nonsense NF1 mutation rather than a large NF1 deletion previously noted in other cases with high-grade astrocytoma. Thus, we propose protein-truncating mutations in NF1 may yield similar behaviors to large NF1 deletions when occurring in conjunction with loss of CDKN2A.

Biopsy of a cerebellar GBM in 2009 revealed two distinct lesions: (1) a grade 4 astrocytoma (GBM) (+GFAP, −S-100, 30% immunopositivity for p53) and a diffuse neurofibroma (−GFAP, +S-100). First, this supports previous evidence that mutations in both NF1 and p53 cooperate in the development of astrocytomas (12, 17, 18). Second, the authors hypothesized that the pre-existing neurofibroma may have created a favorable environment for tumor-initiating cells. Thus they recommended close surveillance of existing lesions in NF1 for the development of high grade astrocytomas (10). Our patient's GBM had no p53 mutation. Though he refrained from MRI surveillance for the 4 years prior to his diagnosis, he had no previous MRI evidence of an infratentorial lesion; nor did his pathology or molecular reports indicate any separate cell type. Therefore, it is likely that his GBM developed de novo. Regardless, the utility of MRI surveillance for patients with NF1 to monitor for both existing and de novo lesions is further reinforced.

While the molecular profiles of these tumors are of significant interest for the development of targeted treatment, more research is required. To date, treatment for NF1 associated GBMs is identical to treatment for sporadic GBMs (5, 6). Targeted therapies, such as immune checkpoints inhibitors, are currently being assessed in clinical trials and could be beneficial for patients with NF1 associated GBMs (19). Of the cases discussed above, three received surgery followed by concurrent radiation and TMZ followed by monthly adjuvant TMZ. One patient lived 41 months after surgery (frontal GBM), (5) another 6 months after surgery (cerebellar GBM), (10) and the third was still alive at the time of publication (frontal GBM, no progression 9 months after surgery) (7). A separate case of an NF1 associated cerebellar glioblastoma showed substantial improvement on MR imaging after 3 weeks of targeted immunotherapy with trametinib, a MEK inhibitor (11). Genomic testing for this tumor revealed no IDH-1 mutation, with NF1 splice site 3870 + 1G>T, CDKN2A/B. Trametinib was used after failure of the standard radiation and TMZ, carboplatin, and also progression on everolimus, an mTOR inhibitor. It should be noted that trametinib was chosen not based on a specific MEK amplification, but because of potential over-activity of the MEK pathway in NF1 (11). More research is required to determine if additional NF1-associated GBM patients would benefit from similar immunotherapy.