Spinal Ependymomas. Part 2: Filum Terminale Ependymomas
Spinal Ependymomas. Part 2: Filum Terminale Ependymomas
Myxopapillary ependymomas of the filum terminale are rare spinal tumors. The study presented here extended over 32 years to collect a series of 42 patien0ts representing 4.5% of all intradural tumors in that period. In a series of 1013 patients with intra- and extramedullary spinal cord tumors, Epstein's group identified a similar rate of 5.1% for myxopapillary ependymomas. With the exception of 3 children, all patients were adults. Males outnumbered females at a ratio of 1.5:1. The most common presenting symptom was back pain, mostly of dull intensity and restricted to the lower back. Some patients reported a pseudoradicular component as well. Neurological deficits were of mild to moderate degree in the majority of patients. Only 7% of patients were unable to walk, and 9% required urinary catheterization at the time of presentation (Table 2). This is in striking contrast to the often considerable size of these tumors and is an indicator of their usually slow growth. On average, it took about 3 years before patients presented and the diagnosis was made. Compared with the most common extramedullary tumors—neurino-mas (n = 170; 21 ± 31 months) and meningiomas (n = 189; 17 ± 26 months)—this time period was significantly longer (p = 0.01). More aggressive clinical presentations have been reported in pediatric series. Acute presentations with a sudden onset of severe pain are rare and related to tumor hemorrhages. This was encountered in only 1 patient in this series.
As expected, the preoperative clinical history tended to be longer for larger tumors. For those patients choosing to decline surgery, the clinical history tended to be considerably longer, reaching almost 10 years on average. This appeared to be the major reason to refuse an operation.
Interestingly, almost all ependymomas extending over a maximum of 3 spinal segments displayed a capsule, whereas the majority of those extending over 4 or more segments did not. This finding may suggest that ependy momas display a tumor capsule at the beginning, which is often quite fragile. Once they reach a certain size, further growth and tumor hemorrhages may lead to rupture of the capsule, infiltration of surrounding nerve roots, and even dissemination into the subarachnoid space. With the exception of 2 unoperated patients presenting with large ependymomas, all patients demonstrating subarachnoid dissemination had undergone incomplete tumor removal, suggesting that dissemination is an iatrogenic phenomenon in the majority of instances. Dissemination is unusual in previously unoperated patients and was restricted to large tumors without capsules in this series. Therefore, the entire s pinal axis should be examined on MRI whenever a large tumor extending over more than 4 spinal segments or a recurrent tumor is encountered.
The preoperative diagnosis requires MR images with and without contrast. Unlike intramedullary ependymomas, which may display various appearances on MRI, ependymomas of the filum terminale demonstrate a more uniform appearance, are isodense to cord tissue on T1-weighted images, and are hyperintense on T2-weighted images with strong contrast enhancement (Fig. 1). Heterogeneous contrast enhancement has been reported to indicate unencapsulated tumors (Fig. 2).
With large tumors filling the entire lumbar canal, the diagnosis of a myxopapillary ependymoma is straightforward. Especially in the sacral region, large tumors may penetrate the dura and lead to local bone destruction (Fig. 2). Smaller ependymomas, however, are impossible to distinguish from other extramedullary tumors such as schwannomas, for instance.
It is the general policy of the 2 institutions involved in the series to recommend surgery for intradural spinal tumors as soon as the diagnosis is made and symptoms are present. Myxopapillary ependymomas should be treated according to this policy as well, aiming for a GTR regardless of their size. The strategy for small tumors is to attempt in toto removal, with resection of the filum. With large tumors filling out the spinal canal completely and/or extending cranially beyond the conus, en bloc removals are impossible without significant neurological deficits and should not be attempted. In such instances, the tumor mass has to be debulked first. Debulking, however, raises the issue of subarachnoid spreading of tumor particles. Therefore, this must be done with great care. Placing cottonoids around the tumor appears to minimize this risk and also helps to protect nerve roots once they have been dissected away from the tumor. An assistant operating a second suction can be very helpful to avoid tumor dissemination during resection. Whenever a tumor displays a capsule, nerve roots will be displaced by the tumor mass rather than be encased in it. Once the mass has been reduced, the filum can be transected distally and then lifted up with the tumor remnant to continue dissection on the ventral side. To identify the exact position of the conus, the cranial pole of the tumor should then be dissected off the spinal cord surface and followed distally. By exposing and dissecting the ventral tumor portion from either end, the conus position can be identified safely and the filum finally cut immediately below it (Fig. 1).
With large tumors encasing nerve roots and causing atrophy of the inner dural layer or even both layers due to the effect of extradural extensions, dissection and removal are more difficult and time consuming and dura closure and reconstruction may also become challenging (Fig. 2). Fortunately, modern imaging techniques and MRI, in particular, seem to have contributed to the lower number of patients presenting with huge tumors in recent years. In this series, the last patient to present with a huge, previously unoperated tumor was in 1999. If tumors filling out the lumbar and sacral canal are encountered, they should be referred to neurosurgeons experienced in their management. Performing biopsies or attempting just a partial resection cannot be recommended as these procedures considerably raise the risk for subarachnoid dissemination. A realistic possibility for GTR in such cases is restricted to the first surgical attempt.
Despite the difficulty, complete resection should be attempted when dissecting a de novo ependymoma encasing nerve roots. Depending on the localization and number of nerve roots appearing to be infiltrated by tumor or carrying significant tumor vessels, the surgeon may have to decide whether to resect them with the tumor, increasing the radicality of the surgery, or to preserve them, risking a local recurrence. Intraoperative monitoring and stimulation may aid in such decisions and are highly recommended.
Once the tumor is resected, dural closure can be performed with a running suture whenever the consistency of the dura allows the surgeon to do so. With atrophy of the inner dural layer, it may be advisable to add a second layer with a fascia lata graft supporting the atrophied dura. Whenever parts of the dura require reconstruction, artificial rather than autologous materials should be employed. Autologous materials tend to cause profound adhesions between nerve roots, conus, and duraplasty and should be avoided. If the dural suture appears not to be reliable, a fascia graft may then be added, as a sandwich, on top of the artificial duraplasty. In such instances, a lumbar drain should be placed as well to lower the subarachnoid pressure for at least a week to prevent a CSF fistula.
With the use of this strategy in our patients, 77.7% of all tumors were completely resected. Incomplete resections were restricted to tumors without capsules (61.5%), whereas all encapsulated tumors underwent GTR. Similar results were reported by Xie et al., who were able to resect 87% of 38 tumors completely, and Kucia et al., who achieved 80% complete resections in 34 patients. In other reviews, considerably lower rates for GTRs were published. Oh et al. reported a rate of only 58.9% for 56 operations; Akyurek et al., 60% for 35 patients; Nakamura et al., 64% for 25 operations; and Celli et al., 71% for 28 patients. In their literature review, Figueiredo et al. calculated a rate of 59.79% GTRs for 199 patients. None of these studies provided separate figures for encapsulated and unencapsulated ependymomas.
Radiotherapy for spinal ependymomas is a controversial issue. It is not recommended for intramedullary ependymomas. These tumors are best treated surgically, can be resected completely, display no subarachnoid dissemination, and carry a very low risk of recurrence. For extramedullary ependymomas, however, issues such as infiltration of nerve roots, which need to be preserved, or potential for subarachnoid dissemination raise the question of postoperative radiotherapy even after GTR. In this decision process, results for radiotherapy of intracranial ependymomas are of no help as ependymomas behave differently according to their site of origin.
There appears to be no general consensus in the literature as to when radiotherapy should be applied or in what form. In a large literature review, Feldman et al., as well as Bagley et al. in their series of 52 patients or Celli et al. studying 28 patients, were unable to establish a benefit for postoperative radiotherapy of myxopapillary ependymomas. On the other hand, Pica et al. found postoperative radiotherapy to be the only independent factor in a multivariate analysis predicting a long progression-free survival in a series of 85 patients. Likewise, Nakamura et al. and Akyurek et al. recommended local radiotherapy as well as radiation for the entire spinal axis to prevent local recurrences as well as postoperative tumor dissemination, whereas Gilhuis et al. reserved radiotherapy for patients after incomplete resections. Some studies suggest that radiotherapy may have a particular role in pediatric patients. In their literature review, Lonjon et al. calculated a recurrence rate of 33% with radiotherapy and 55% without, suggesting that radiotherapy may have a role as adjuvant therapy. In another recent literature review, radiotherapy was recommended even after GTRs, as recurrence rates were reported to be lower than rates with surgery alone.
In this series, only 1 recurrence was encountered after GTR of encapsulated tumors. Surgery alone appears to be curative for this subgroup, similarly to their intramedullary counterparts (Fig. 3). For unencapsulated tumors, radiotherapy was employed in 5 patients. Four of these tumors had demonstrated subarachnoid dissemination prior to surgery. The fifth patient had undergone an incomplete resection of a large ependymoma early in the series. In the long term, all patients with unencapsulated ependymomas experienced a recurrence irrespective of the amount of resection or administration of radiotherapy, according to Kaplan-Meier analysis. However, there was a trend for longer recurrence-free survival for patients after radiotherapy in this subgroup: none of these patients experienced a neurological deterioration before 12 years, whereas without radiotherapy, 41.7% of patients had shown a clinical relapse within 5 years (Fig. 4). In a multiple regression analysis, postoperative radiotherapy was an independent predictor for a long progression-free interval. With the limitation of the small number of patients involved, this study suggests that radiotherapy may have a role in patients with subarachnoid dissemination and after incomplete resections of large unencapsulated tumors. However, attempting a GTR before these tumors have reached a size spreading over more than 3 spinal segments offers the best chance for a long progression-free survival.
The immediate postoperative outcome was very favorable for the great majority of patients of this series. Among 36 operations, 26 were followed by postoperative improvements. Only 7 patients experienced an unchanged status, and 3 patients underwent a permanent deterioration of bladder function after surgery. As one would expect, the results after 1 year somewhat depended on the degree of preoperative deficits and were better for encapsulated tumors, even though both subgroups benefitted from surgery, as shown in Table 3. The rate of 8.3% for permanent surgical morbidity is considerably lower than rates reported for other studies.
In the long term, tumor recurrence rates are low after GTRs for encapsulated tumors (Fig. 3). On the other hand, for unencapsulated tumors a GTR does not prevent a recurrence. In their literature reviews, Feldman et al. and Benesch et al. found significantly higher recurrence rates for pediatric than for adult patients. Agbahiwe et al. reported recurrence rates for children after GTRs of 38.5% and 70% within 5 and 10 years, respectively, and Bagley et al. reported a rate of 64%. In this study, only 3 pediatric patients were included. One of them experienced a recurrence related to tumor dissemination after undergoing a second attempt of GTR.
In this series, 9 tumors recurred. The overall recurrence rates according to Kaplan-Meier statistics were 6.6%, 19.0%, and 37.0% after 1, 10, and 20 years, respectively. Lonjon et al. reported a recurrence rate of 15% after complete resections according to a literature review, but without distinguishing tumors with and without capsules. In a more recent review, Figueiredo et al. reported higher overall recurrence rates of 21% and 34% after 5 and 10 years, respectively. Most studies, however, do not use statistical techniques such as Kaplan-Meier analysis to report recurrence rates and only provide the numbers of recurrent tumors: Xie et al. observed 6 recurrences among 38 patients; Kucia et al., 3 recurrences in 34 patients; Al-Habib et al., 3 recurrences among 18 patients; and Akyurek et al., 5 recurrences among 11 patients. Compared with intramedullary ependymomas, recurrence rates after GTRs are higher for ependymomas of the filum terminale.
In this series, recurrences were found almost exclusively after surgery on unencapsulated tumors, with all tumors having recurred within 24 years after surgery (Fig. 3). Likewise, Xie et al. found unencapsulated tumors, those extending over more than 2 spinal levels, and those invading the sacral canal to bear a worse long-term prognosis. Celli et al. considered a long preoperative clinical history and adherence to conus and nerve roots as unfavorable features associated with higher recurrence rates—characteristics more common in unencapsulated tumors. Whether a GTR of an unencapsulated tumor may prevent a recurrence in the long term cannot be evaluated in this series, as the maximum follow-up for this subset of patients was limited to 6 years.
Five patients died during follow-up in this series. Two tumor-related deaths occurred in the group of patients with unencapsulated tumors: death was due to widespread dissemination in 1 case (at 287 months after surgery) and urosepsis in the other (at 216 months after surgery). The corresponding 20-year survival rate was 88.9%. No comparable data have been published in the literature. The fact, however, that a benign tumor of WHO Grade I may be responsible for a patient's death unless it is diagnosed early and completely resected indicates that this pathology should be addressed surgically irrespective of the sometimes indolent course for long periods of time.
Discussion
Clinical Presentation and Diagnosis
Myxopapillary ependymomas of the filum terminale are rare spinal tumors. The study presented here extended over 32 years to collect a series of 42 patien0ts representing 4.5% of all intradural tumors in that period. In a series of 1013 patients with intra- and extramedullary spinal cord tumors, Epstein's group identified a similar rate of 5.1% for myxopapillary ependymomas. With the exception of 3 children, all patients were adults. Males outnumbered females at a ratio of 1.5:1. The most common presenting symptom was back pain, mostly of dull intensity and restricted to the lower back. Some patients reported a pseudoradicular component as well. Neurological deficits were of mild to moderate degree in the majority of patients. Only 7% of patients were unable to walk, and 9% required urinary catheterization at the time of presentation (Table 2). This is in striking contrast to the often considerable size of these tumors and is an indicator of their usually slow growth. On average, it took about 3 years before patients presented and the diagnosis was made. Compared with the most common extramedullary tumors—neurino-mas (n = 170; 21 ± 31 months) and meningiomas (n = 189; 17 ± 26 months)—this time period was significantly longer (p = 0.01). More aggressive clinical presentations have been reported in pediatric series. Acute presentations with a sudden onset of severe pain are rare and related to tumor hemorrhages. This was encountered in only 1 patient in this series.
As expected, the preoperative clinical history tended to be longer for larger tumors. For those patients choosing to decline surgery, the clinical history tended to be considerably longer, reaching almost 10 years on average. This appeared to be the major reason to refuse an operation.
Interestingly, almost all ependymomas extending over a maximum of 3 spinal segments displayed a capsule, whereas the majority of those extending over 4 or more segments did not. This finding may suggest that ependy momas display a tumor capsule at the beginning, which is often quite fragile. Once they reach a certain size, further growth and tumor hemorrhages may lead to rupture of the capsule, infiltration of surrounding nerve roots, and even dissemination into the subarachnoid space. With the exception of 2 unoperated patients presenting with large ependymomas, all patients demonstrating subarachnoid dissemination had undergone incomplete tumor removal, suggesting that dissemination is an iatrogenic phenomenon in the majority of instances. Dissemination is unusual in previously unoperated patients and was restricted to large tumors without capsules in this series. Therefore, the entire s pinal axis should be examined on MRI whenever a large tumor extending over more than 4 spinal segments or a recurrent tumor is encountered.
The preoperative diagnosis requires MR images with and without contrast. Unlike intramedullary ependymomas, which may display various appearances on MRI, ependymomas of the filum terminale demonstrate a more uniform appearance, are isodense to cord tissue on T1-weighted images, and are hyperintense on T2-weighted images with strong contrast enhancement (Fig. 1). Heterogeneous contrast enhancement has been reported to indicate unencapsulated tumors (Fig. 2).
With large tumors filling the entire lumbar canal, the diagnosis of a myxopapillary ependymoma is straightforward. Especially in the sacral region, large tumors may penetrate the dura and lead to local bone destruction (Fig. 2). Smaller ependymomas, however, are impossible to distinguish from other extramedullary tumors such as schwannomas, for instance.
Surgical Management
It is the general policy of the 2 institutions involved in the series to recommend surgery for intradural spinal tumors as soon as the diagnosis is made and symptoms are present. Myxopapillary ependymomas should be treated according to this policy as well, aiming for a GTR regardless of their size. The strategy for small tumors is to attempt in toto removal, with resection of the filum. With large tumors filling out the spinal canal completely and/or extending cranially beyond the conus, en bloc removals are impossible without significant neurological deficits and should not be attempted. In such instances, the tumor mass has to be debulked first. Debulking, however, raises the issue of subarachnoid spreading of tumor particles. Therefore, this must be done with great care. Placing cottonoids around the tumor appears to minimize this risk and also helps to protect nerve roots once they have been dissected away from the tumor. An assistant operating a second suction can be very helpful to avoid tumor dissemination during resection. Whenever a tumor displays a capsule, nerve roots will be displaced by the tumor mass rather than be encased in it. Once the mass has been reduced, the filum can be transected distally and then lifted up with the tumor remnant to continue dissection on the ventral side. To identify the exact position of the conus, the cranial pole of the tumor should then be dissected off the spinal cord surface and followed distally. By exposing and dissecting the ventral tumor portion from either end, the conus position can be identified safely and the filum finally cut immediately below it (Fig. 1).
With large tumors encasing nerve roots and causing atrophy of the inner dural layer or even both layers due to the effect of extradural extensions, dissection and removal are more difficult and time consuming and dura closure and reconstruction may also become challenging (Fig. 2). Fortunately, modern imaging techniques and MRI, in particular, seem to have contributed to the lower number of patients presenting with huge tumors in recent years. In this series, the last patient to present with a huge, previously unoperated tumor was in 1999. If tumors filling out the lumbar and sacral canal are encountered, they should be referred to neurosurgeons experienced in their management. Performing biopsies or attempting just a partial resection cannot be recommended as these procedures considerably raise the risk for subarachnoid dissemination. A realistic possibility for GTR in such cases is restricted to the first surgical attempt.
Despite the difficulty, complete resection should be attempted when dissecting a de novo ependymoma encasing nerve roots. Depending on the localization and number of nerve roots appearing to be infiltrated by tumor or carrying significant tumor vessels, the surgeon may have to decide whether to resect them with the tumor, increasing the radicality of the surgery, or to preserve them, risking a local recurrence. Intraoperative monitoring and stimulation may aid in such decisions and are highly recommended.
Once the tumor is resected, dural closure can be performed with a running suture whenever the consistency of the dura allows the surgeon to do so. With atrophy of the inner dural layer, it may be advisable to add a second layer with a fascia lata graft supporting the atrophied dura. Whenever parts of the dura require reconstruction, artificial rather than autologous materials should be employed. Autologous materials tend to cause profound adhesions between nerve roots, conus, and duraplasty and should be avoided. If the dural suture appears not to be reliable, a fascia graft may then be added, as a sandwich, on top of the artificial duraplasty. In such instances, a lumbar drain should be placed as well to lower the subarachnoid pressure for at least a week to prevent a CSF fistula.
With the use of this strategy in our patients, 77.7% of all tumors were completely resected. Incomplete resections were restricted to tumors without capsules (61.5%), whereas all encapsulated tumors underwent GTR. Similar results were reported by Xie et al., who were able to resect 87% of 38 tumors completely, and Kucia et al., who achieved 80% complete resections in 34 patients. In other reviews, considerably lower rates for GTRs were published. Oh et al. reported a rate of only 58.9% for 56 operations; Akyurek et al., 60% for 35 patients; Nakamura et al., 64% for 25 operations; and Celli et al., 71% for 28 patients. In their literature review, Figueiredo et al. calculated a rate of 59.79% GTRs for 199 patients. None of these studies provided separate figures for encapsulated and unencapsulated ependymomas.
Radiotherapy
Radiotherapy for spinal ependymomas is a controversial issue. It is not recommended for intramedullary ependymomas. These tumors are best treated surgically, can be resected completely, display no subarachnoid dissemination, and carry a very low risk of recurrence. For extramedullary ependymomas, however, issues such as infiltration of nerve roots, which need to be preserved, or potential for subarachnoid dissemination raise the question of postoperative radiotherapy even after GTR. In this decision process, results for radiotherapy of intracranial ependymomas are of no help as ependymomas behave differently according to their site of origin.
There appears to be no general consensus in the literature as to when radiotherapy should be applied or in what form. In a large literature review, Feldman et al., as well as Bagley et al. in their series of 52 patients or Celli et al. studying 28 patients, were unable to establish a benefit for postoperative radiotherapy of myxopapillary ependymomas. On the other hand, Pica et al. found postoperative radiotherapy to be the only independent factor in a multivariate analysis predicting a long progression-free survival in a series of 85 patients. Likewise, Nakamura et al. and Akyurek et al. recommended local radiotherapy as well as radiation for the entire spinal axis to prevent local recurrences as well as postoperative tumor dissemination, whereas Gilhuis et al. reserved radiotherapy for patients after incomplete resections. Some studies suggest that radiotherapy may have a particular role in pediatric patients. In their literature review, Lonjon et al. calculated a recurrence rate of 33% with radiotherapy and 55% without, suggesting that radiotherapy may have a role as adjuvant therapy. In another recent literature review, radiotherapy was recommended even after GTRs, as recurrence rates were reported to be lower than rates with surgery alone.
In this series, only 1 recurrence was encountered after GTR of encapsulated tumors. Surgery alone appears to be curative for this subgroup, similarly to their intramedullary counterparts (Fig. 3). For unencapsulated tumors, radiotherapy was employed in 5 patients. Four of these tumors had demonstrated subarachnoid dissemination prior to surgery. The fifth patient had undergone an incomplete resection of a large ependymoma early in the series. In the long term, all patients with unencapsulated ependymomas experienced a recurrence irrespective of the amount of resection or administration of radiotherapy, according to Kaplan-Meier analysis. However, there was a trend for longer recurrence-free survival for patients after radiotherapy in this subgroup: none of these patients experienced a neurological deterioration before 12 years, whereas without radiotherapy, 41.7% of patients had shown a clinical relapse within 5 years (Fig. 4). In a multiple regression analysis, postoperative radiotherapy was an independent predictor for a long progression-free interval. With the limitation of the small number of patients involved, this study suggests that radiotherapy may have a role in patients with subarachnoid dissemination and after incomplete resections of large unencapsulated tumors. However, attempting a GTR before these tumors have reached a size spreading over more than 3 spinal segments offers the best chance for a long progression-free survival.
Outcome
The immediate postoperative outcome was very favorable for the great majority of patients of this series. Among 36 operations, 26 were followed by postoperative improvements. Only 7 patients experienced an unchanged status, and 3 patients underwent a permanent deterioration of bladder function after surgery. As one would expect, the results after 1 year somewhat depended on the degree of preoperative deficits and were better for encapsulated tumors, even though both subgroups benefitted from surgery, as shown in Table 3. The rate of 8.3% for permanent surgical morbidity is considerably lower than rates reported for other studies.
In the long term, tumor recurrence rates are low after GTRs for encapsulated tumors (Fig. 3). On the other hand, for unencapsulated tumors a GTR does not prevent a recurrence. In their literature reviews, Feldman et al. and Benesch et al. found significantly higher recurrence rates for pediatric than for adult patients. Agbahiwe et al. reported recurrence rates for children after GTRs of 38.5% and 70% within 5 and 10 years, respectively, and Bagley et al. reported a rate of 64%. In this study, only 3 pediatric patients were included. One of them experienced a recurrence related to tumor dissemination after undergoing a second attempt of GTR.
In this series, 9 tumors recurred. The overall recurrence rates according to Kaplan-Meier statistics were 6.6%, 19.0%, and 37.0% after 1, 10, and 20 years, respectively. Lonjon et al. reported a recurrence rate of 15% after complete resections according to a literature review, but without distinguishing tumors with and without capsules. In a more recent review, Figueiredo et al. reported higher overall recurrence rates of 21% and 34% after 5 and 10 years, respectively. Most studies, however, do not use statistical techniques such as Kaplan-Meier analysis to report recurrence rates and only provide the numbers of recurrent tumors: Xie et al. observed 6 recurrences among 38 patients; Kucia et al., 3 recurrences in 34 patients; Al-Habib et al., 3 recurrences among 18 patients; and Akyurek et al., 5 recurrences among 11 patients. Compared with intramedullary ependymomas, recurrence rates after GTRs are higher for ependymomas of the filum terminale.
In this series, recurrences were found almost exclusively after surgery on unencapsulated tumors, with all tumors having recurred within 24 years after surgery (Fig. 3). Likewise, Xie et al. found unencapsulated tumors, those extending over more than 2 spinal levels, and those invading the sacral canal to bear a worse long-term prognosis. Celli et al. considered a long preoperative clinical history and adherence to conus and nerve roots as unfavorable features associated with higher recurrence rates—characteristics more common in unencapsulated tumors. Whether a GTR of an unencapsulated tumor may prevent a recurrence in the long term cannot be evaluated in this series, as the maximum follow-up for this subset of patients was limited to 6 years.
Five patients died during follow-up in this series. Two tumor-related deaths occurred in the group of patients with unencapsulated tumors: death was due to widespread dissemination in 1 case (at 287 months after surgery) and urosepsis in the other (at 216 months after surgery). The corresponding 20-year survival rate was 88.9%. No comparable data have been published in the literature. The fact, however, that a benign tumor of WHO Grade I may be responsible for a patient's death unless it is diagnosed early and completely resected indicates that this pathology should be addressed surgically irrespective of the sometimes indolent course for long periods of time.
Source...