Department of Neurosurgery, Karolinska Hospital, Stockholm, Sweden
OBJECTIVE: Long-term data on the natural history of traditionally treated cranial base meningiomas are necessary to judge the benefit of modern cranial base techniques for individual patients and to understand when nonradical surgery of a meningioma is in the interest of the patient. The only available means of obtaining such data is investigation of patients treated before the present surgical era.
METHODS: The records of 315 patients who were operated on at Karolinska Hospital between January 1, 1947, and December 31, 1982, were reviewed. Of the patients, 10.8% died perioperatively and 9.7% died within 10 years. The remaining patients were followed for 10 to 36 years (mean, 18 yr).
RESULTS: The 5-year recurrence rate was 4% for patients undergoing radical surgery (Grades 1 and 2) and 25 to 45% for patients undergoing Grade 3 or 4 operations. Follow-up periods longer than 5 years revealed that 16% of Grade 1 and 20% of Grade 2 patients had symptomatic recurrences, whereas a majority of Grade 4 and 5 patients showed symptomatic progression. Forty-two of 69 patients who underwent Grade 4 or 5 operations died as a result of their tumors, usually within 10 years after the first operation. No patients who underwent Grade 4 or 5 operations were free from symptomatic progression after 20 years. The tumor progression or recurrence was usually detected within the 1st 10 years, but late recurrences were seen ¾25 years after the operation. The worst outcome was found in medial sphenoid wing/clinoidal meningiomas and in tumors invading the cavernous sinus. Subfrontal tumors showed unexpectedly high recurrence rates, with a mortality rate ¾14% in the late phase.
CONCLUSION: The findings emphasized the necessity to plan the management of patients with cranial base meningiomas according to a 10- to 20-year perspective. Patients must be followed to evaluate the treatment results and to detect recurrences. Nonradical surgery must be viewed as a temporizing or palliative measure; a continued search for means of radical tumor treatment is warranted in these often surgically difficult tumors.
(Neurosurgery 39:29, 1996)
Key words: Brain tumor, Cavernous sinus, Cranial base, Meningioma, Radiation therapy, Recurrence
Meningiomas are considered benign because of their slow growth rates and the feasibility of surgical cure. The initial surgical challenge in cranial base meningiomas is to evacuate tumor without causing mortality or morbidity. With improved surgical knowledge and technical progress, an uneventful clinical course after tumor removal is expected. It is nonetheless worthwhile to analyze and optimize the long-term results of surgery of benign lesions.
Extensive cranial base surgery involves a risk of morbidity and mortality. Some surgeons advocate a radical approach with total tumor removal (2, 26), whereas others favor a conservative attitude, supported by the thought that the natural history of these lesions is also favorable if tumor remains (4, 10, 23). Extensive extradural dissection and bone work paired with improved microsurgical skills have allowed the removal of tumors that were previously regarded as unoperatable. After the postoperative phase, eventual results depend on whether the recurrence of totally removed tumors or the continued growth from subtotally removed tumors causes neurological deficits or death in the long term. Knowledge of the natural history of traditionally treated meningiomas must therefore be considered when an aggressive surgical approach is in the interest of the individual patient. The only available sources of long-term data on these lesions are records of patients treated in the past.
This analysis was undertaken to fulfill the need of data regarding extended follow-up after traditional cranial base surgery. The report comprises a 10- to 36-year follow-up of cranial base meningiomas treated at one institution between 1947 and 1982. The recurrence rates and long-term outcomes are reported for different degrees of radicalness and for different tumor locations.
The site of the original tumor was recorded as subfrontal (anterior fossa with attachment anterior to optic foramina), tuberculum sellae, medial sphenoid wing, middle-lateral sphenoid wing, or posterior fossa (clivus, cerebellopontine angle, petrosal). Its pathological characteristics and invasion of the cavernous sinus were recorded. Further, the clinical course, radiation therapy, outcome at 1 year, recurrence, and cause of death were noted. All patients were followed until recurrence or death or for a minimum of 10 years. The mean follow-up was 18 years. Only clinical recurrences were recorded to allow use of data from before the computed tomography era, to allow comparison with previous series for which clinical recurrences were recorded, and because clinical recurrences are of much higher clinical significance than radiological recurrences.
The actual recurrence rates were recorded at different time intervals, without a need of life table analysis. Such data presentation avoids extrapolation. However, some patients died because of unrelated causes or were lost to follow-up. Figures are therefore shown for both the detected recurrences and an estimated maximum of possible recurrences. The recurrence rates at n year's follow-up are given as intervals between the cumulative number of observed recurrences divided by the total number of patients studied (lower percentage) and the cumulative number of observed recurrences divided by the number of patients alive and available for follow-up at the nth year (higher percentage). The second figure will be lower than a hypothetical "true" figure only if patients dying from unrelated causes or not followed longer would have had a much higher recurrence rate than those followed. Patients who died as a result of their tumors or had known recurrences were entered as "recurrences." Patients dying from other causes were entered in the "lower percentage group" described above. Analyses were made with McIntosh computers (Apple Computer, Inc.) and Cricket Graph (Abacus Concepts, Inc.) or Statview (Computer Associates International, Inc.) software.
Overall long-term results
Figure 1 shows recurrence rates for cranial base meningioma removal at 5 to 25 years. The 5-year recurrence rates for patients who underwent radical surgery (Grades 1 and 2) were 3.5 to 4%, whereas 25 to 45% of patients who underwent less than radical operations experienced symptomatic progression in this interval. Longer follow-up periods revealed that ¾16% of Grade 1 and 20% of Grade 2 patients had symptomatic recurrences, whereas a majority of Grade 4 and 5 patients showed symptomatic progression (Table 1).
FIGURE 1. Recurrence rates for cranial base meningiomas at 5 to 25 years. White boxes, Simpson Grade 1; black circles, Grade 2; black triangles, Grade 3; white triangles, Grade 4. For each grade, the best case situation is shown by a black line and the worst by a dotted line.
| Grade | |||||
|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | |
| No. of patients | 53 | 137 | 56 | 55 | 14 |
| Years after surgery | |||||
| 5 | 3.5% | 4% | 25% | 45% | 36% |
| 15 | 710% | 1115% | 3743% | 81100% | 63100% |
| 25 | 1316% | 1520% | 3976% | | |
The results with subtotal removal are shown in Table 2. A minority of patients with Grade 4 or 5 operations remained stable, whereas the rest showed progressive symptoms. Forty-two of 69 patients died as a result of their tumors, usually within 10 years after the first operation. No patients with Grade 4 or 5 operations who had been followed for >20 years were free from symptomatic progression; the longest follow-up period without a progression was 17 years.
| Survival (yr) | No. of Patients |
|---|---|
| <3 | 17 |
| 310 | 13 |
| 1020 | 10 |
| >20 | 2 |
The tumor progression or recurrence was usually detected within the 1st 10 years, but late recurrences were also seen. Of the 57 patients observed in the interval of 20 to 25 years, 7 showed recurrences during that interval (12%). These recurrences were truly late in at least two patients who had been investigated with computed tomographic scans at 15 and 17 years, respectively, without any signs of an expanding tumor. The 46 patients followed for 25 to 36 years did not show any recurrences.
Histology
Seven of the 315 patients (2%) had tumors with histological signs of malignancy or atypia (i.e., increased cellularity, mitoses, necroses and/or brain invasion). Of the patients with totally benign tumors that recurred, 26 underwent subsequent operations. At subsequent operations, seven patients (27%) showed changes indicative of increased malignancy or atypia as compared with the original specimens.
Recurrence by location
Table 3 shows the recurrence rates with different tumor locations. An extremely high rate of recurrence was found for tumors of the central cranial base, but 16 to 27% of subfrontal tumors also recurred in 25 years. As many as 14% of the patients with subfrontal tumors died as a result of tumor recurrence. Tumors of the middle/lateral sphenoid wing had the lowest rates of recurrence.
| Location | No. of Tumors | % at 10 years | % at 25 years |
|---|---|---|---|
| Subfrontal | 75 | 910 | 1627 |
| Tuberculum sellae | 49 | 1217 | 2741 |
| Temporal fossa floor | 23 | 1619 | 2628 |
| Cliniodal/medial sphenoid wing | 49 | 60100 | 72100 |
| Sphenoid wing en plaque | 9 | 2533 | 50 |
| Middle/lateral sphenoid wing | 61 | 1317 | 1540 |
| Posterior fossa | 44 | 1421 | 1821 |
The very late recurrences (2025 yr) were always located in the central cranial base (7 of 44 patients followed) and not laterally (0 of 13 patients followed). Of Simpson Grade 1 and 2 tumors in the subfrontal, clinoidal, or cerebellopontine angle regions, 2% recurred in the interval of 20 to 25 years.
Medial sphenoid wing/clinoidal meningioma
The worst results regarding recurrence were found in medial sphenoid wing/clinoidal meningiomas. Of 49 patients, 6 died during surgery. Of the remaining 43 patients, 22 died as a result of their tumors, 7 died because of other causes, 10 were alive with progressing symptoms, and 4 had been followed without recurrences for 11 to 14 years after Grade 1, 2, or 3 operations.
Cavernous sinus invasion
Table 4 shows the results for 48 patients with cavernous sinus invasion, as described by the operative notes or shown radiologically. Of these, 10 died because of unrelated causes or from the original surgery and 38 could be followed until death from tumor or for >10 years. Fourteen patients died within 5 years from tumor progression, and 13 died between 5 and 15 years after surgery. Two died later. Only three were alive and without progressive symptoms 10 to 15 years after surgery.
| Survival (yr) | No. of Patients | No. of Patients Remaining Asymptomatic |
|---|---|---|
| <2 | 4 | 12 |
| 25 | 10 | 9 |
| 510 | 7 | 7 |
| 1015 | 6 | 1 |
| 1520 | 1 | 0 |
| 2025 | 1 | 0 |
Radiation therapy
Radiation therapy for patients with central cranial base invasion in the central cranial bone or for invasion of the cavernous sinus was not administered according to a uniform protocol during the study period. Table 5 shows the 10- and 20-year results for patients with inoperative tumor growth in the central cranial base. The 10-year mortality rate was statistically lower in the radiated groups. This difference is not present in the 20-year follow-up. No patients seemed to be cured by radiation therapy. Symptomatic progression or death was noted in a majority of the patients in both groups with sufficient follow-up, which could, however, be 15 to 20 years.
| Radiation (No. of Patients) | No Radiation (No. of Patients) | |
|---|---|---|
| Alive at 10 yr | 10 of 13 | 14 of 33 |
| Alive at 20 yr | 5 of 13 | 6 of 33 |
This long-term follow-up study revealed very high recurrence rates in central cranial base tumors. Furthermore, the recurrences were of high clinical relevance, causing progressive neurological deficit and/or death. Remaining tumor regrew in a large majority of patients with a sufficiently long follow-up. Obviously, the difficult locations must have precluded results that can be achieved with the surgical techniques available today. Any long-term follow-up studies include patients who were operated on long ago (in our study, by H. Olivecrona) and cannot analyze patients who have been treated with the most recent methods. The immediate surgical results are vastly improved (2, 5, 25, 26) as compared to the old surgical results (7, 14, 24). The rationale of performing this analysis was, however, to understand the natural course in the long term after treatment. Such data are important for planning therapy and follow-up for patients expected to live 30 to 50 years after treatment.
Asymptomatic meningiomas are common incidental autopsy findings (31). Meningiomas account for only 7% of death from brain tumors (9). These tumors are therefore considered relatively benign. However, these benign tumors were hazardous to our patients in a long-term perspective, even if the initial surgery was successful. The natural history of these tumors must be considered and respected in tailoring a treatment that aims further than the first operation.
Literature review
High recurrence rates of meningiomas have been shown in previous reports (1, 8, 15, 17, 19, 22, 32). The long-term results with reportedly radical operations were better in this survey than in the previous reports. Our results also indicate fewer recurrences of surgically accessible tumors, i.e., tumors of the lateral cranial base, than reported previously (1, 19, 22, 27). In contrast, the recurrence rates and final outcomes for patients with inaccessible tumors (i.e., tumors of the central cranial base) and those who underwent subtotal operations were not good. Recurrent tumors were detected >20 years after the initial operation. The long-term follow up showed that 5 or 10 years without recurrence is not a safe interval for ruling out future recurrences. In this series, 25 years were necessary to reach a limit after which tumors did not recur. Thus, the slow growth of meningiomas necessitates a follow-up of >10 years in each patient to get reliable data on the efficiency of different treatments. Several reasons for the difference between our findings and those of previous reports are possible.
This analysis was based on actual long-term observations. It differs from most previous reports (1, 8, 22) in that we did not use a life table analysis to estimate recurrence rates but reported the figures from patients followed until 5, 15, and 25 years after surgery. Extrapolations and assumptions regarding when to expect a tumor recurrence were thus avoided. Avoiding life table analysis introduces another problem. Some patients died because of unrelated causes or were lost to follow-up before the respective end points were reached. We therefore chose to show figures for the detected recurrences and an estimated maximum of possible recurrences.
Another explanation of the different results in previous reports as compared with this study might be that the operations reported as Simpson Grade 1 and 2 operations were radical to a higher degree in our study. A recurrence in this setting indicates that tumor was left at surgery despite the surgeon's conviction that removal was radical. We compared the case load of meningioma operations in different institutions reporting long-term recurrence rates. Table 6 shows the recurrence rates in the different series. A strong negative correlation is detectable between the number of meningioma operations per year in the reported series and the number of tumor recurrences. It is probable that the degree of radicalness in operations reported as "radical" may have differed according to the surgeon's experience with a large number of similar operations. This argument is corroborated by the similarity of progression in operations known to be subtotal (i.e., Simpson Grade 4 and 5 operations). The finding would support the view that the amount of surgical experience of a certain lesion within the prevalent surgical paradigm at any given time determines long- and short-term outcomes of patients with cranial base meningiomas.
| Grade | |||||
|---|---|---|---|---|---|
| Series (Ref. No.) | op/yrb | 1 | 2 | 3 | 4 |
| (% Recurrence) | |||||
| Simpson, 1957 (27) | 10 | 9 | 19 | 29 | 44 |
| Yamashita et al., 1980 (32) | 9 | 33 | 27 | 44 | |
| Adgebite et al., 1983 (1) | 5 | 14 | 18 | 52 | |
| Chan and Thompson, 1984 (8) | 14 | 11 | 22 | 33 | |
| Mirimanoff et al., 1985 (22) | 18 | 7 | 37 | ||
| Jääskeläinen, 1986 (15) | 3 | ||||
| Present series | 30 | 3 | 4 | 25 | 45 |
Surgical radicalness
The operative notes that were analyzed showed that the surgeons had conscientious attitudes regarding the radical removal of tumors. Radical removal was regarded as a necessary goal, and the measures taken to achieve this goal were carefully described. Still, recurrence rates were high in the long term, and the consequences of the recurrences were worse than would be expected when dealing with benign tumors. The obvious and rational explanation of tumor recurrence is that tumor removal was not radical (17, 27). The difficult tumor locations that exert surgical challenges today must have precluded radical removal with the surgical techniques that were available. The importance of radical tumor removal, including its dural and bony attachments, was noticed as early as 1957 (27). Modern surgical techniques of extensive extradural bone work before intradural tumor removal lower future tumor recurrence rates (2, 25, 26). The findings of late recurrences of anterior clinoid and subfrontal meningiomas especially stress that bony and dural tumor attachments in such locations, which are usually accessible, should be removed. Data from this old series are therefore highly relevant for planning surgery today. It is, however, not always feasible to remove tumor that is invading central structures or to be sufficiently radical in the cranial base, especially considering that microscopic tumor clusters can be found at a distance from the main tumor (6).
Radiotherapy
Radiotherapy is widely used to treat surgically inaccessible tumor remnants, malignant meningiomas, or diffusively growing recurrent tumor. It seems to improve outcome in invasive and/or aggressive meningiomas (4, 21, 29). Radiotherapy seems to have improved the results for the patients in this study. Ten-year results of meningiomas invading the central cranial base were better in patients who had received radiotherapy than in those who were not treated. A longer follow-up, however, revealed that progressive tumor growth and bad outcomes also resulted for patients who had been treated with radiotherapy. The present data, therefore, support the use of radiotherapy for palliation of unresectable meningiomas, but the long-term results are not in agreement with reports claiming cure of central tumors with diagnostic biopsy and radiotherapy (10). Medical palliation of unresectable meningioma has also been tried but has achieved only limited success (13).
Tumor biology
Tumor biology is another factor related to recurrences. Twenty-six percent of recurrent tumors showed histological signs of aggressive behavior, such as atypia, necroses, or invasions that were not detectable in the original tumors (16, 18, 28, 30). The finding agrees with previous reports of increased histological signs of aggressive behavior in recurrent tumors and indicate selection of cell clones with a propensity to grow faster than the original tumor (1, 8, 16, 27, 28, 32). Cranial base meningiomas in general were, however, reported to be biologically aggressive in lower proportions than meningiomas in other locations (3, 30). Our findings of only 2 to 3% anaplastic/histologically aggressive tumors are in agreement. The histological appearance of a tumor is not sufficient to predict its clinical behavior. A comprehensive management of these tumors could be improved with refined diagnostic means to evaluate the proportion of dividing cells, receptors, etc. (11, 12, 16, 20, 28). Still, the main problem with recurrent tumor did not appear to be linked to aggressive behavior or an increased recurrence potential but to continued and, with time, invasive growth of histologically benign tumors.
Received, September 1, 1995.
Accepted, January 16, 1996.
Reprint requests: Tiit Mathiesen, M.D., Department of Neurosurgery, Karolinska Hospital, S-171 76, Stockholm, Sweden.
Some of the limitations of this study must be pointed out. The grading of the degree of resection was based on the surgeon's impression and not on postoperative imaging. We all know from current imaging experience that surgeons' estimates of resection can be heavily flawed. Second, recurrence and progression were considered on the basis of appearance of symptoms and not radiological examination. This once again depends on the location of the tumor, the attitude of the patient and physician, and the frequency of follow-up. Third, 10.8% of the patients died perioperatively. Comparisons of series and results should take this into consideration. Perhaps some of these patients would have lived longer if they had been only observed or treated by radiotherapy. Last, the patients' quality of life was not studied. This has been a problem with almost all studies that have looked at meningioma recurrence. It is obviously better for a patient to live for 5 years in a good condition than for 20 years in a poor condition.
Notwithstanding the above limitations, this study is extremely important. It will serve as a benchmark for other surgeons to compare their results, until a better quality study, satisfying the above criteria, is published.
The optimal philosophy of treating cranial base meningiomas is still elusive, although we have made much progress with regard to their surgical treatment and adjuvant therapies (particularly radiosurgery). We would ideally like to see a complete resection of the tumor, without permanent morbidity. However, as the treatment becomes more radical, the perioperative morbidity and mortality rates may increase. Removing the tumor partially and then radiating the remaining tumor is another option. However, radiotherapy does present some risks. Many of these tumors will recur, and surgical treatment of previously irradiated tumors is much more hazardous, with higher risk of morbidity and mortality.
Therefore, all studies that report the efficacy of radiotherapy in controlling tumors (including radiosurgery) and studies that report surgical results must address the total management morbidity, i.e., the initial perioperative morbidity, the quality of life during follow-up, and the morbidity of treating recurrences. Obviously, randomized, controlled studies are ideal, but in their absence, the reports should carefully describe features about the tumors that would allow comparisons.
This study reiterates the now well-established fact that the extent of tumor resection is the most important variable with regard to the recurrence of cranial base meningiomas. It also points to the value of long-term follow-up. These two considerations are very important in our current environment of managed care, in which less is considered more and many difficult lesions are operated on by surgeons with limited experience.
Laligam N. Sekhar
William Monacci
Washington, District of Columbia
This article provides significant data that will deeply impact our understanding of meningiomas and their management. Not since Simpson's milestone article in 1957 (4) has the effect of surgical treatment on the recurrence of meningiomas been this unequivocally documented. Clearly, benign meningiomas must be totally removed, including any tumor involving bone and dura, and this excision must be achieved at the first operation. The means to do this for meningiomas of the cranial base are available today with the advent of microsurgery and cranial base surgery.
Conservative attitudes, which still prevail in many centers around the world, are based on the concern regarding morbidity and the complexity of the surgical endeavors required to achieve Grade 1 removal for cranial base meningiomas. Furthermore, cases have been followed for years after partial removal and the meningiomas have remained dormant with no evidence of progression. Absent, and sorely needed, was a study that addressed a sizable number of cranial base meningiomas with long-term follow-ups, hence, defining the natural history of cranial base meningiomas. This article provides the answers.
For obvious reasons, the authors report on symptomatic recurrences. In further perspective or contemporary studies, particularly with the wide availability and the high resolution of magnetic resonance imaging, radiologically revealed recurrences should be documented. They will be detected earlier and, undoubtedly, will be much more frequent than indicated by symptomatic recurrences.
Because of the special conditions surrounding cranial base meningiomas, Kobayashi et al. (3) introduced a modified Simpson grading system. The Kobayashi modification is helpful, and we have adopted it for our reporting on cavernous sinus meningiomas (2). We encourage all neurosurgeons to integrate this considerably improved system into their record keeping and data reporting.
Of particular interest is the finding on radiation therapy. This study not only has demonstrated, as have earlier ones, that radiation may delay symptomatic recurrence, but it also has confirmed that longer follow-ups substantiate the need for radical removal and that radical removal should be attempted at first operation. A surgeon's desire and determination to achieve safe total removal should not be hampered, especially by the potential effects of radiation. This article particularly demonstrates the ineffectiveness of radiation on long-term follow-up and that long-term radiation effects are not insignificant (1).
I reemphasize the authors' point regarding bone removal. Cranial base tumors frequently involve bone. What is called hyperostosis is actually involvement of bone by tumor, which has been documented repeatedly. This issue became especially relevant in regard to frontal basal meningiomas, olfactory groove meningiomas, and tuberculum sella meningiomas, all of which have an excellent potential for permanent curative removal. Conservative management of the involved bone at the cranial base has been the main reason for recurrences that typically occur 10 years later and further involve the cranial base and the paranasal sinuses. We have managed several of these recurrent tumors, and it is worthwhile to stress this point.
Ossama Al-Mefty
Little Rock, Arkansas
The authors report data regarding the long-term results of patients who underwent surgery for the treatment of cranial base meningiomas. It is important to note that they used clinical recurrences as the basis for their data. The series covers patients treated between 1947 and 1982. It is likely that many of the recurrences that caused death and disability would have been detected much earlier and treated if routine follow-up computed tomography or magnetic resonance imaging had been available. Therefore, most of the results, although of historical interest, have limited application to the treatment of these tumors at the present time.
There is, however, one group that provides important data. It is likely that a Simpson Grade 1 operation (complete removal with resection of dura and abnormal bone) has not changed a great deal over the years, with the exception that microsurgical techniques make complete removal more assured and have likely increased the number of patients in whom this is possible. For Simpson Grade 1 operations, the authors report, for all sites, a recurrence rate of 3.5% at 5 years, 7 to 10% at 15 years, and 13 to 16% at 25 years. In addition to the references cited by the authors, the importance of the length of follow-up has also been emphasized by Philippon (1) who found that one-third of recurrences occurred greater than or equal to 10 years after operation. The authors' data on this and the other Simpson grades indicate that all patients who have undergone surgery for cranial base meningiomas need to be followed indefinitely with computed tomography or magnetic resonance imaging every 1 to 2 years, even when there has been an apparent total removal of the cranial base meningioma. This data will be important in documenting for insurance companies the need for these routine follow-up studies.
Robert G. Ojemann
Boston, Massachusetts
Return to the Neurosurgery Home Page
