Surgical Management of Adult Intrinsic Brainstem Tumors

The literature on adult patients with brainstem gliomas is sparse. This is mostly a reflection of the rarity of these tumors, which account for < 2% of all primary brain tumors in adults. This is contrary to the incidence in children of 10% and the subsequent prolific number of series concentrating on children alone.1-6 Furthermore, wide variations in prognosis in the adult literature are a consequence of variations in the definitions of brainstem tumors, with some series including malignant with low-grade and tectal lesions and many series determining histopathological “grading” on radiological characteristics alone.7-14 Much of the nihilism associated with primary brainstem tumors in children has been reduced in the last decade through the promising results of a few pediatric series.3,6 The same cannot be said for the attitudes of adults faced with a primary, intrinsic brainstem tumor. Certainly, recommendations for surgery of these treacherous lesions have been poorly defined. Although there are some absolute indications for surgery such as atypical tumor characteristics on magnetic resonance imaging (MRI), a significant exophytic component, or a clearly focal lesion, the same cannot be said for a patient who presents with a diffuse low-grade tumor, an expanding tectal glioma, or a clinically and radiologically malignant tumor. This surgical series of adult patients with brainstem gliomas attempts to address some of these unanswered questions. PATIENTS AND METHODS We retrospectively reviewed all patients with intrinsic brainstem lesions who underwent surgical resection by the senior author (C.T.) at our institution over a 10-year period (January 2000 through December 2010). Using our database, we identified all patients who underwent surgery for presumed intrinsic brainstem tumors. From this cohort, we selected patients > 18 years of age with a histopathologically proven diagnosis of glioma for further analysis. Patients < 18 years of age, those with extra-axial lesions, and patients with histopathologically confirmed nonglial lesions were excluded. Patient demographics, presenting symptoms, tumor location, extent of resection, postoperative complications, histopathology, use of adjuvant therapies, and clinical and radiological follow-up were obtained from medical records. Telephone interviews were conducted by a neurosurgeon to obtain missing follow-up data when necessary. Preoperative imaging was reviewed to determine the location of the lesion and whether it was focal or diffuse. A lesion was considered focal if there was definitive peripheral contrast enhancement or if the T1 image corresponded exactly with the T2 image on MRI and showed well-demarcated edges. Lesions that did not have these properties were defined as diffuse and were considered for surgery if they exhibited either an exophytic component or partial relatively well-defined T2/fluid-attenuated inversion-recovery signal characteristics. Patients were subsequently divided into those with low- or high-grade lesions on the basis of the clinical history and MRI features. The preoperative presumptive grade for each case was then compared with the final pathological diagnosis after resection to ascertain the reliability of clinicoradiological features in predicting the histopathological grade. In each case, the extent of resection was analyzed on the basis of both the surgeon’s intraoperative impression and the postoperative MRI and defined as gross total resection if there was no visible evidence of residual tumor, near-total resection if there was minimal contrast enhancement or high T2/fluid-attenuated inversion-recovery signal change, and subtotal resection (> 90%, 70%-90%, 50%-70%, and < 50%), depending on the amount of tumor remaining. Volumetric analyses were not performed. RESULTS Patient Demographics Over a 10-year period, 138 patients underwent surgery for presumed intrinsic brainstem tumors. Of these, 104 procedures were performed in patients < 18 years of age. Two patients were found to have nonglial tumors. One patient had a secondary neoplasm, and the other had an intra-axial tuberculoma (Figure 1). The remaining 31 patients met our inclusion criteria and were analyzed further. The demographic characteristics of these patients are given in Table 1. Eighteen patients (58%) were male and 13 (42%) were female. The mean age at surgery was 33 years (range, 19-56 years). Presenting symptoms/signs included imbalance/incoordination in 10 (32.3%), motor weakness in 8 (25.8%), visual disturbance in 5 (16.1%), headache in 4 (12.9%), sensory loss in 4 (12.9%), facial weakness in 3 (9.7%), vertigo/dizziness in 2 (6.5%), tinnitus in 2 (6.5%), and dysphagia in 1 (3.2%). The mean duration of presenting symptoms/signs varied, depending on the histopathology of the lesion, from 25.2 months (range, 2-96 months) for low-grade lesions (juvenile pilocytic astrocytoma [JPA], ganglioglioma, and low-grade astrocytoma [LGA]) to 4.7 months (range, 0.5-24 months) for high-grade lesions (anaplastic astrocytoma [AA] and glioblastoma multiforme [GBM]). Similarly, malignant lesions were more likely to present with multiple neurological deficits. Tumor location and characteristics are summarized in Table 2. There were 7 tumors located in the midbrain (tegmentum/peduncle), 3 in the tectal plate, 9 in the pons, 3 at the pontomedullary junction, and 9 in the medulla. Indications for operating on tectal lesions included symptomatic enhancing lesions or radiological progression of nonenhancing lesions. The majority of tumors (n = 28, 90.3%) were focal in nature, and only 3 lesions were considered diffuse.FIGURE 1: T1-weighted magnetic resonance image of the brain with contrast (axial view) demonstrating a large focally enhancing right-sided pontine lesion in a 32-year-old man presenting with left-sided hemiparesis. The patient underwent surgical resection of the lesion, which was found to be a tuberculoma. This case was excluded from the present study; however, it shows that other pathological entities can mimic primary brainstem tumors and emphasizes the importance of tissue diagnosis.TABLE 1: Demographic Data in 31 Adult Patients With Intrinsic Brainstem TumorsTABLE 2: Tumor Location and Characteristics in 31 Adult Patients With Intrinsic Brainstem TumorsSurgical Results The extent of tumor resection is summarized in Table 3. A gross or near-total resection was achieved in 64% of patients. Final histopathology demonstrated 14 low-grade lesions (4 JPAs, 4 gangliogliomas, 6 LGAs) and 17 high-grade lesions (7 AAs, 10 GBMs; Table 4). Of the 3 patients with tectal lesions, histopathology showed 1 tumor to be a LGA and the other 2 tumors to be malignant gliomas (1 AA and 1 GBM). In the majority of cases (n = 27, 87%), the preoperative clinicoradiographic features were reliable in predicting the histopathological grade. Of the 4 cases (13%) in which the preoperative diagnosis was incorrect, the lesion was undergraded in 3 patients and overgraded in 1 patient, underscoring the importance of tissue diagnosis in patients with brainstem tumors. None of the patients required an endoscopic third ventriculostomy or permanent cerebrospinal fluid shunting. There were no surgery-related deaths. Sixteen patients (51.6%) were worse neurologically after surgery; however, 12 patients (38.7%) made a complete recovery to baseline or better at the last follow-up (Table 5). Permanent neurological morbidity occurred in 4 patients (12.9%) and was seen only in cases with gross or near-total resections (Table 6).TABLE 3: Extent of Resection in 31 Adult Patients With Intrinsic Brainstem TumorsTABLE 4: Histopathology in 31 Adult Patients With Intrinsic Brainstem TumorsTABLE 5: Surgical Complications in 31 Adult Patients With Intrinsic Brainstem TumorsTABLE 6: Influence of Extent of Resection on Surgical ComplicationsClinical and Radiological Follow-up The overall mean clinical and radiologic follow-up period was 43.3 and 36.6 months, respectively (range, 3-132 months). All patients with malignant gliomas (AAs and GBMs) underwent postoperative radiation with or without temozolomide therapy except for 2 patients who had undergone preoperative radiation therapy at other institutions. None of the patients with low-grade lesions (JPA, ganglioglioma, LGA) received adjuvant therapies. The mean time to radiologic tumor progression for nontectal JPA, ganglioglioma, LGA, AA, and GBM was 69, 57, 50.4, 31.5, and 6 months, respectively. The mean time to tumor progression for the 3 patients with tectal lesions was 120 months for the LGA, 12 months for the AA, and 8 months for the GBM (Table 7). At last clinical follow-up, a total of 16 patients had died of their disease at a mean time of 28.3 months from surgery (Table 8). Overall, 13 of the 17 patients (76.4%) with malignant gliomas and 3 of the 14 patients (21.4%) with low-grade lesions died at a mean time of 16.8 and 78 months from surgery, respectively. Both patients with malignant tectal lesions died, whereas the 1 patient with a tectal LGA remains alive and well at 10 years. Two of the 6 patients with nontectal AA died at a mean time of 42 months from surgery. The remaining 4 patients are alive at a mean time of 54 months. On the other hand, all 9 patients with nontectal GBM died at a mean time of 12.3 months after resection. Three of the 13 patients with nontectal low-grade lesions died. Two patients with LGA died of tumor progression and suspected malignant transformation 96 and 60 months from surgery. One patient with a medullary ganglioglioma, who suffered a devastating neurological complication after surgery, died of pneumonia 78 months later. The remaining 10 patients with low-grade lesions are alive at a mean of 55.2 months after surgery.TABLE 7: Mean Time to Radiologic Tumor Progression Based on Location and PathologyTABLE 8: Survival Outcomes of 31 Adult Patients With Intrinsic Brainstem TumorsILLUSTRATIVE CASES Case Illustration 1 A 39-year-old woman presenting with a 6-month history of tinnitus was found to have a diffuse but relatively well-defined hyperintense lesion within the right pons and middle cerebellar peduncle (Figure 2A). The patient underwent surgical resection via a presigmoid/petrosal approach. Postoperative MRI demonstrated a gross total resection (Figure 2B). The patient developed a transient facial paresis and cerebellar dysfunction. Histopathology proved the tumor to be an AA, and she subsequently underwent radiation alone. She remains alive without evidence of tumor recurrence 7 years after surgery.FIGURE 2: A, preoperative T2-weighted magnetic resonance image (MRI) of the brain (axial view) demonstrating a diffuse but well-defined hyperintense lesion within the pons and middle cerebellar peduncle. B, postoperative MRI showing a gross total resection. Histopathology revealed an anaplastic astrocytoma.Case Illustration 2 A 19-year-old woman presenting with a 3-year history of facial paresis was found to have a focally enhancing lesion in the left pons (Figure 3A). A near-total resection was performed (Figure 3B). The tumor proved to be a JPA. The patient is alive and well 11 years after surgery.FIGURE 3: A, preoperative T1-weighted magnetic resonance image (MRI) of the brain with contrast (axial view) showing a left-sided focal pontine lesion. B, postoperative MRI demonstrating a near-total resection. Histopathology revealed a juvenile pilocytic astrocytoma.Case Illustration 3 A 20-year-old man presenting with a 6-month history of motor weakness was found to have a focally enhancing pontotegmental lesion (Figure 4A). A near-total resection was performed (Figure 4B). Histopathology revealed a JPA. The patient was neurologically worse after surgery. He is alive at 2 years after surgery.FIGURE 4: A, preoperative T1-weighted magnetic resonance image (MRI) of the brain with contrast (axial view) demonstrating a focally enhancing pontotegmental lesion. B, postoperative MRI showing a near-total resection. Histopathology revealed a juvenile pilocytic astrocytoma.Case Illustration 4 A 36-year-old female patient presenting with a 3-month history of left hemiparesis was found to have a right-sided focally enhancing midbrain lesion extending rostrally into the thalamus (Figure 5A). The patient underwent gross total resection (Figure 5B). Postoperatively, the patient suffered a transient worsening of her hemiparesis and a mild third nerve paresis. Histopathology proved the tumor to be an AA, and the patient received postoperative radiation and chemotherapy. The tumor recurred after 6 months, and she ultimately died 18 months after surgery.FIGURE 5: A, preoperative T1-weighted magnetic resonance image (MRI) of the brain with contrast (axial view) demonstrating a right-sided focally enhancing lesion involving the midbrain with rostral extension into the thalamus. B, postoperative MRI showing a gross total resection of the tumor, which proved to be an anaplastic astrocytoma.Case Illustration 5 A 40-year-old man presenting with a 1-year history of visual disturbances was found to have a large focal tectal lesion (Figure 6A). The patient underwent surgery and was found to have an AA. Postoperative MRI demonstrated a near-total resection (Figure 6B). The patient subsequently received postoperative radiation and chemotherapy. The tumor recurred 1 year later, and he ultimately died 24 months after surgery.FIGURE 6: A, preoperative T2-weighted magnetic resonance image (MRI) of the brain (sagittal view) showing a large focal tectal lesion. B, postoperative T1-weighted MRI (sagittal view) demonstrating near-total resection. Histopathology revealed an anaplastic astrocytoma.DISCUSSION The management of adult patients with primary brainstem tumors is challenging. The current literature is sparse and confusing. In some series, the classification of brainstem gliomas includes the diffuse intrinsic nonenhancing, the focal enhancing, and the tectal varieties but fails to mention the focal nonenhancing type.7,12 In other series, pediatric and adult patients are discussed together despite the fact they are very distinct entities with different tumor biology and prognosis.10,14 Accurate data on prognosis are also lacking because some series have combined malignant with low-grade gliomas,13,14 yet others have separated the 2 groups.7,12 Furthermore, most of these series were based on a radiological classification without tissue confirmation. Ueoka et al14 attempted to address this potential confounder by using the patient’s clinical course to determine malignancy, resolving that the tumors that progressed within 12 months were likely to be malignant and those after 12 months to be low grade. Finally, the literature fails to define clear surgical indications. The prevailing trend is either partial resection or biopsy followed by adjuvant therapy or management based on the clinicoradiological features of the lesion without tissue histology. In a large surgical series comprising 40 pediatric and adult patients, Mehta et al10 offered surgery to patients if they had “expanding focal lesions.” However, the question of surgical intervention for those patients who initially present with a symptomatic focal glioma or the benefit of aggressive resection is not clarified by a review of the current literature. Although the numbers are small, to the best of our knowledge, this is the largest surgical series of adult brainstem gliomas in the current literature. Furthermore, we attempted to define clear surgical indications. For tectal gliomas, surgery was recommended only for those patients with symptomatic enhancing lesions or radiological progression of nonenhancing lesions. Most tectal gliomas appear as stable, nonenhancing lesions and present with hydrocephalus secondary to aqueductal compression. The management of these lesions involves an endoscopic third ventriculostomy and serial follow-up MRI scans. The indolent nature of this glioma subtype is well recognized and usually evokes no controversy in the literature.7 Consequently, there were only 3 patients who had surgery for a tectal lesion in this series. Two patients had enhancing lesions that were proven to be malignant, and the third patient had radiological progression of a low-grade tumor. One patient with an AA survived 24 months; the other patient with a GBM died 18 months after surgery. The patient with an LGA remains alive and well 10 years after a > 90% resection. For nontectal gliomas, we offered surgery to all patients with symptomatic focal lesions regardless of the enhancement pattern and whether or not there was documented radiological progression. Surgery was also recommended for symptomatic diffuse lesions if they were relatively well circumscribed (Figure 2) or if they had a significant exophytic component. On the other hand, patients with diffuse ill-defined lesions not amenable to radical resection were not offered surgery (Figure 7).FIGURE 7: T2- (A) and T1- (B) weighted magnetic resonance image of the brain (axial views) demonstrating a right-sided ill-defined pontine lesion in a 45-year-old male patient. The patient was not offered surgery because it was thought that lesion was too diffuse to achieve a good resection.Of the 28 patients with nontectal gliomas, those with low-grade lesions (n = 13) had a longer clinical history and generally a better overall prognosis. A total of 10 patients with nontectal low-grade gliomas are alive at a mean of 55.2 months after surgery. Two patients were significantly worse postoperatively and still recovering. Both patients had total or near-total resections. The remaining 8 patients (61.5%), 4 of whom had a total or near-total resection, are either neurologically better or stable postoperatively. A total of 3 patients with low-grade nontectal tumors died. Two patients with LGA died of tumor progression and suspected malignant transformation 96 and 60 months from surgery. One patient with a medullary ganglioglioma, who suffered a devastating neurological complication after surgery, died of pneumonia 78 months later. Historically, adult patients with malignant brainstem gliomas do very poorly, with a median survival time of 11.2 months.7 In this surgical series, the median survival time for nontectal malignant lesions was 25.5 months. Only 1 of these patients suffered a permanent neurological deficit. Two of the 6 patients with nontectal AA died at a mean time of 42 months from surgery. The remaining 4 patients are alive at a mean time of 54 months. On the other hand, all 9 patients with nontectal GBM died at a mean time of 12.3 months after resection. The benefit from surgical resection of adult intrinsic brainstem tumors remains controversial. In the present series of 31 patients, a > 90% resection was achieved in 24 cases (77.4%) and resulted in a permanent neurological deficit in 4 cases (12.9%). Similarly, in the large surgical series by Mehta and colleagues,8 the authors reported good outcomes in the majority of cases with radical excision. However, it should be noted that this series included a significant number of pediatric patients. On the contrary, Mursch et al9 questioned the benefit of surgery for adult intrinsic brainstem tumors. In their series of 16 adult patients, a > 80% resection was achieved in only 2 cases, and the permanent neurological morbidity rate was 63.6% (n = 7). The authors found that patients with low-grade tumors did as badly as those with malignant lesions. In our opinion, surgical resection of intrinsic brainstem lesions, when possible, has several potential advantages. Although the clinical course and radiological characteristics help in differentiating low-grade from high-grade lesions, this is not without error.8 In the majority of cases in our series, we were able to reliably predict, on the basis of the clinicoradiological features, whether a lesion was going to be pathologically low or high grade. However, in 4 cases (13%), our preoperative diagnosis was incorrect. The lesion was undergraded in 3 patients and overgraded in 1 patient, underscoring the importance of tissue diagnosis. Similarly, we excluded 3 patients from the present study after they were found to have histologically proven nonglial tumors (secondary neoplasm and tuberculoma) or an extra-axial lesion. Another potential advantage of surgery is the relief of mass effect. In several cases, we have observed improvement of preoperative deficits after tumor resection, which may be related to relief of tumor compression on adjacent neural structures. It is also important to note that despite several patients having preoperative hydrocephalus, none of our patients required an endoscopic third ventriculostomy or shunt placement. Finally, the most controversial aspect of this aggressive surgical approach resides in the philosophical question of the price one must pay to enjoy longer survival. It is sobering to note that all 4 patients who were significantly and permanently worse after surgery had a near-complete or complete resection. Would they have survived as long and had a better quality of life had they received a less radical resection? Indeed, 2 long-term survivors in this series had > 10% residual tumor. Unfortunately, the answer to this question is beyond the scope of this study. CONCLUSION The surgical management of adult intrinsic brainstem tumors remains challenging. Aggressive surgical resection can be accomplished with acceptable morbidity and may result in improved outcomes compared with historical controls. Although complications may be incapacitating, they are relatively uncommon. The indications for surgery are different for tectal and nontectal tumors. The importance of differentiating focal, well-defined diffuse, and ill-defined diffuse lesions cannot be overemphasized. Disclosure The authors do not have any personal or institutional financial interest in drugs, materials, or devices described in this submission.