Retrospective analysis of 104 histologically proven adult brainstem gliomas: clinical symptoms, therapeutic approaches and prognostic factors

Adult brainstem gliomas are a very rare (<2% of gliomas) and poorly investigated disease. Recently several larger series on patients with brainstem gliomas have been reported, however, these series were only partly based on a histologically established diagnosis (Kesari [1], Landolfi [2], Guillamo [3], Salmaggi [4]) or involved also other pathologies than gliomas (Rachinger [5], Samadani [6]). Importantly, Rachinger and colleagues recently stated that intraaxial brainstem lesions with a radiological pattern of glioma represent a very heterogeneous tumour group with completely different outcomes and that metastasis, lymphoma, inflammation and cavernoma could be misinterpreted as a glioma by magnetic resonance imaging (MRI).

The goal of this study was therefore to analyze clinical, prognostic and therapeutic factors in the largest series of histological proven brainstem glioma reported so far.

Here we report one of the largest series of histopathologically proven gliomas of the brainstem. Patients presented with a wide variety of symptoms, and median survival was only 18.8 months for the whole patient population. The amount of neuroradiological differential diagnoses confirmed the necessity of histopathological evaluation. In multivariate analysis KPS ≤ 70, higher tumor grade and age ≥ 40 were negative prognostic factors, whereas radiation therapy or radio-chemotherapy improved prognosis.

Recently several large series about brainstem gliomas in adults with median overall survival rates between 54 and 85 months have been published (Table 4). However diagnosis of a brainstem glioma in these series was mainly based on neuroradiological imaging and confirmed by histopathological examinations only in 13% - 67% of cases. Rachinger found that in a series of 46 radiologically suspected brainstem gliomas histological examination confirmed a glioma in only 28 cases (61%) and revealed metastasis in 15% (n = 7), lymphoma in 11% (n = 5), inflammatory disease in 4% (n = 2), cavernoma in 2% (n = 1) and gliosis in 6% (n = 3). The authors pointed out, that intra-axial brainstem lesions with a radiological pattern of glioma represent a very heterogeneous tumor group with completely different clinical outcomes [5]. Samadani published a meta-analysis of 293 brainstem biopsies in children and adults. Stereotactic biopsy was in 96% diagnostic with a mortality rate of 0.3%, a transient morbidity rate of 4% and a permanent morbidity rate of 1%. Pathology showed that half of the adult brainstem intrinsic lesions were gliomas, 10% were metastases, and the remainders were hematomas, vascular malformations, lymphomas, demyelination, cysts, radiation necrosis, abscesses, vasculitis, infarcts, leukemia, cryptococcus, or granulomas. They pointed out that neuroradiological diagnosis of non-contrast-enhancing lesions of the brainstem as low grade glioma is insufficient and that histopathological diagnosis for non-enhancing brainstem lesions with a long duration of symptoms revealed hematomas, arteriovenous malformations, lymphomas, demyelination, radiation necrosis and infarction. Therefore the authors concluded that stereotactic biopsy is indicated for both enhancing and non-enhancing lesions of the brainstem [6].

Kickingereder et al. recently published a large meta analysis of 1480 stereotactic biopsies for brainstem tumors and found a diagnostic success rate of 96.2%, an overall morbidity rate of 7.8%, a permanent morbidity rate of 1.7% and a mortality rate of 0.9% [9].

The clinically relevant postoperative morbidity of 9.7% in our series is in line with the data of Kickengereder but higher in comparison with figures from the meta-analysis of Samadani and colleagues [6]. The main difference between these two series was the homogenous histopathology in our series which consisted only of gliomas with a proportion of high-grade tumors of nearly 50% as opposed to a broad variety of tumorous and non-tumorous diseases in the latter series. Because of the rich neovascularization of high-grade gliomas, the risk of of postoperative hemorrhage or malignant brain-edema is likely to be higher in malignant gliomas compared to other pathologies and may explain the differences in morbidity.

Several publication have discussed this issue with regard to stereotactic brain biopsy for supratentorial lesions. Bernstein suggested that biopsy of specific pathologies (e.g. glioblastoma, lymphoma) may be associated with an increased risk of either hemorrhage or severe edema, due to the abnormal neovasculature of these tumors [10]. Savin confirmed these data and identified malignant glioma pathology to be associated with a 4-fold increased risk of morbidity, especially from hemorrhage [11]. However other authors found no association between lesion pathology and complication rates [12]. A recent study about complication of frame-based stereotactic biopsy in 622 cases identified an association between mortality and glioblastoma pathology and suggested that abnormal tumor neovasculature of malignant glioma may be the reason therefore [13].

Overall morbidity of stereotactic biopsy in brainstem tumors ranges between 7.8% and 12% in larger series and is therefore not distinctly higher compared to a morbidity rate of 4.9% in general stereotactic brain biopsy, especially when considering the low rate of 1.7% of permanent morbidity in stereotactic brainstem biopsy [1, 4, 9, 13].

Dellaretti and colleagues investigated the correlation between magnetic resonance imaging findings and histological diagnosis of intrinsic brainstem lesions in adults in a series of 96 patients. Stereotactic biopsy established a precise histological diagnosis in 92 patients which consisted of 63 diffuse brainstem gliomas, 19 other neoplastic diseases (lymphomas, metastases, pilocytic astrocytomas, craniopharyngioma, ganglioma) and 10 non-neoplastic lesions (inflammatory disease, ischemic lesion, fungal abscess, gliosis). Overall morbidity rate was 9% and one patient died from exacerbated peritumoral edema. With regard to neuroradiological features the diagnostic effect of stereotactic biopsy was greater in patients with focal or enhancing lesions shown by MRI in whom the diagnosis of a diffuse gliomas was less frequent [14].

The value of additional imaging modalities to improve non-invasive diagnostic accuracy by MR spectroscopy or positron emission tomography is currently under investigation. However, Massager showed recently in a series of 30 brainstem gliomas that the integration of PET imaging can not replace histological analysis as MRI combined with PET data was only concordant with histological findings in 63% of cases [15].

The results of these studies are indicative that in adult patients with lesions of the brainstem therapeutic decisions should be based on a histopathological examination due to the wide spectrum of differential diagnoses.

Therefore we included in our study only patients with a histopathologically confirmed brainstem glioma to exclude a possible bias due to non-glioma lesions classified as gliomas by MRI, which may have a significant better prognosis. This and the high rate of malignant brain stem gliomas of 44.7% might explain the distinct difference in median overall survival of 18.8 months in our series in comparison to the actual literature of brainstem gliomas. We found that median overall survival of treated patients with HGG of the brainstem resembled the overall survival data of patients with supratentorial high grade gliomas. However median overall survival of patients with grade II gliomas of the brainstem was significant shorter than in brainstem glioma series of Kesari (26.4 months vs. 168 months) and in comparison to supratentorial low grade gliomas (26.4 months vs. 7–8 years). Reasons for this difference might be the location within a highly eloquent area, faster malignant transformation than in supratentorial gliomas for unknown reasons or possible histopathological undergrading. It is also notable that in Kesari’s series grade I brainstem gliomas had a significant shorter median overall survival of 83 months in comparison to grade II brainstem gliomas with a median overall survival of 168 months. Inaccurate neuroradiological diagnosis and subsequent undergrading may also explain this as up to 60% of high grade brainstem gliomas show no contrast enhancement after gadolinium application [6].

The location within a highly eloquent area may also explain the finding that the percentage of glioblastoma is significant lower in the brainstem compared to its supratentorial counterpart (12.5% vs 60-75%) [16] as brainstem gliomas may become clinically symptomatic very early in the course of the disease. We also found that pilocytic astrocytomas of the brainstem are surprisingly not such a benign disease as supported by many authors [17‐19] with rapid progression especially in the first 20 months after diagnosis and stabilization in survival thereafter (see Figure 1a). These data are also supported by Stuer [20] who observed after a median follow-up of 55 months 30% tumor recurrence and 18% deaths. Another surprising result was the lack of pure oligodendroglial brain stem gliomas and the low proportion of oligoastrocytic tumors. Interestingly in the series of Guillamo 25% of biopsied brainstem gliomas were oligodendrocytic or mixed gliomas.

Established prognostic factors in the literature are age, duration of symptoms, KPS, contrast enhancement, MRI “necrosis”, histology and location in the pons and medulla, mainly based on univariate analysis [2, 3, 21, 22]. We confirmed the strong prognostic impact of KPS and age in multivariate analysis and the positive effect of radiation therapy as the cornerstone of therapeutic measures on overall survival (therapy reduces the risk of death by a factor of 0.4). The value of chemotherapy in low and high grade brainstem gliomas is still undefined. Efficacy of different chemotherapeutic agents (temozolomide, nitrosureas or platinum based chemotherapeutic protocols) is currently unproven. At relapse a wide variety of chemotherapeutic agents are used which included BCNU [1,2-bis(2-chloroethyl)-1-nitrosourea], BCNU - procarbazine, CCNU [1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea] -procarbazine-vincristine (PCV), carboplatin, carboplatin-VP16, carboplatin-VP16-ifosfamid, ifosfamid, procarbazine-VP16, temozolomide, CCNU, vincristine, irinotecan, cisplatin and temozolomide, ACNU and procarbazine. Effectiveness of these protocols were limited: Guillamo [3] reported a radiological response rate of 7% three months after onset of chemotherapy and clinical improvement lasting longer than 6 months in 15% of patients and Samlaggi [4] described a temporary clinical and radiological stabilization in 22% of patients after chemotherapy.

Alternative therapeutic strategies like interstitial radiosurgery with implantation of iodine-125 seeds and application of anti-angiogenic drugs like bevacizumab have to be also considered in the therapeutic concept. Mundinger treated in a series of 89 low grade brainstem gliomas 55 patients with stereotactic brachytherapy. 29 patients received iodine-125 seeds with a 5 year survival rate of 54.8% and 26 patients received iridium 192 seed with a 5 year survival rate of 26.9% in comparison to 5 year survival rate of 14.7% in patients who underwent only biopsy [23]. Ruge recently published a series of 47 patients with inoperable focal brainstem gliomas WHO grades I and II treated by stereotactic implantation of iodine-125 seeds with a 5-year overall survival rate of 97.4 ±2.6% [24].

Survival rates of interstitial radiosurgery are therefore at least comparable to external beam radiotherapy with reported 5-year survival rates between 45%-58% [1, 2] and both methods should be evaluation against each other in prospective randomized trials.

Reports about the use of antiangiogenic substances in the literature are rare. Besides two case reports [25, 26] a small series of 3 patients [27] showed the effectivness of bevacizumab as a salvage therapy for progressive brainstem gliomas with improvement of clinical condition, reduction of daily dexamethasone dosage and radiological response.

Adult brain stem gliomas present with a wide variety of neurological symptoms and neuroradiological differential diagnoses. Stereotactic biopsy is the procedure of choice to obtain a histopathological diagnosis. Prognosis of high-grade gliomas resembles its supratentorial counterparts whereas low-grade gliomas of the brainstem have a worse prognosis compared to the actual literature. Cornerstone of therapy remains radiation and alternative strategies like interstitial radiosurgery, chemotherapy or antiangiogenic drugs need to be further explored, ideally in the context of molecular profiling for common alterations such as 1p/19q codeletion, MGMT promoter methylation and IDH mutation.