Meningiomas have increasingly been detected incidentally due to advances in neuroimaging, such as CT and MR imaging.⁹ Neurological examinations and neuroradiological screening for indefinite complaints also contribute to this increase in detection, especially in advanced countries.¹⁷ Although most IDMs can be cured surgically,¹⁸ the morbidity and mortality associated with surgery itself cannot be determined as these tumors are generally considered asymptomatic and benign.⁹ Therefore, determining the proper management for these cases is critical. However, due to the lack of knowledge regarding the exact natural history of these tumors and their potential for growth, what to consider as “appropriate” treatment remains controversial.¹⁴

Several authors have investigated the natural history of IDMs.^{3,4,9,14,16,23,24} They have identified various factors predictive of tumor growth; however, these factors differ from study to study. Recently, we analyzed 70 IDMs using a volumetric method with a relatively long follow-up period. We concluded that most IDMs do not seem to grow for a certain period of time, and that they do not always grow exponentially but rather exhibit complex patterns of growth.⁶ In addition, the presentation of calcification on CT or MR imaging has a negative impact on its growth.

We further surveyed other factors predictive of IDM growth, and we identified tumor location, that is, non–skull base versus skull base, as a potentially useful clinicoradiological predictor of the growth behavior of IDMs. Here, we report our results and discuss biological evidence to support these findings through expansion of the theory involving symptomatic meningiomas.

Methods

A total of 110 patients with IDMs, who are being followed up at our institution, were included in this study. The tumor growth rate of individual IDMs was analyzed using a manual volumetric method. Then, to confirm the biological differences between skull base and non–skull base tumors, 210 symptomatic patients with Grade I meningiomas, based on the WHO 2007 classification, who were treated in Osaka University Hospital, were also included. This study was approved by the ethics review board of the university.

Results

Volumetric Comparison of Growth Rates and Patterns Between Skull Base and Non–Skull Base Meningiomas

Demographic data are summarized in Table 1. A total of 113 IDMs from 110 patients were included in this study; 2 patients had multiple tumors. There were 93 women and 17 men, and the mean age was 66.8 years (range 37–91 years). The mean follow-up period for sequential MR images was 46.9 months, ranging from 12 to 121 months. Table 2 shows the distribution of the cases based on location. Thirty-eight (34%) were located in the skull base and 75 (66%) were not.

TABLE 1:

Demographic data of 113 IDMs for volumetric analyses

Variable	All	Skull Base	Non–Skull Base
no. of cases (%)	113	38 (34)	75 (66)
mean age in yrs (range)	66.8 (37–91)	66.8 (40–84)	66.8 (37–91)
sex ratio (F:M)	93:17	29:9	64:8
mean follow-up period in mos (range)	46.9 (12–121)	40.3 (12–100)	50.3 (12–121)
mean initial tumor volume in cm3 (range)	9.79 (0.32–86.62)	12.22 (0.56–86.62)	8.56 (0.32–74.69)
no. w/o growth (%)	42 (37)
no. w/ growth (%)	71 (63)

TABLE 2:

Distribution of all IDM cases based on location

Location	No. of Cases
convexity	35
falx	17
petroclival	17
parasagittal	14
sphenoid wing	7
cavernous	6
intraventricular	4
planum sphenoidale	4
tentorium	4
olfactory	3
other	2

Of the 113 IDMs, 42 tumors (37%) showed no growth during the follow-up period and 71 (63%) showed growth. Interestingly, when all IDMs were divided into skull base and non–skull base categories, only 15 (39.5%) of 38 skull base meningiomas showed growth, whereas 56 (74.7%) of 75 non–skull base meningiomas showed growth (Fig. 1; p = 0.0004, Fisher exact test). Of the 38 skull base IDMs, 22 (57.9%) did not grow, and 1 reduced in volume.

Incidence of meningiomas showing growth and no growth in skull base and non–skull base tumors. Fifteen (39.5%) of 38 skull base tumors showed growth, whereas 56 (74.7%) of 75 non–skull base tumors showed growth (p = 0.0004, Fisher exact test).

As shown in Table 3, of the 71 IDMs that showed growth, no statistically significant difference was noted between the skull base and non–skull base groups in terms of mean age, sex, follow-up period, or initial tumor volume. However, the annual relative growth rate (p = 0.009) and the percentage of growth (p = 0.002) were much lower in the skull base group than in the non–skull base group. Accordingly, with slower growth rate, the tumor doubling time (p = 0.008) was much higher in the skull base group.

TABLE 3:

Comparison between skull base and non–skull base lesions in 71 growing IDMs^*

Variable	Skull Base (range)	Non–Skull Base (range)	p Value	Statistics Used
no. of cases	15	56
mean age in yrs	69.5 (40–83)	66.6 (37–91)	0.320	M-W
sex ratio (F:M)	11:4	48:6	0.207	Fisher
mean follow-up period in mos	49.3 (20–100)	53.1 (12–121)	0.485	M-W
mean initial tumor volume in cm3	14.71 (0.79–86.62)	8.11 (0.32–74.69)	0.163	M-W
absolute growth rate in cm3/yr	1.20 (0.02–7.85)	1.15 (0.04–11.33)	0.662	M-W
relative growth rate %/yr	6.84 (2–24)	13.78 (2–74)	0.009†	M-W
% growth	25.56 (15.57–43.27)	94.83 (16.11–1519.67)	0.002†	M-W
mean tumor doubling time	160.8 (39–383)	111.5 (15–420)	0.008†	M-W

* M-W = Mann-Whitney.

† Statistically significant.

To further exemplify the differential growth rate between the 2 groups, the case of a 67-year-old man with multiple meningiomas, who underwent follow-up for 25 months, is presented. The patient had 1 cavernous and 2 convexity meningiomas (Fig. 2). The doubling time for the skull base tumor was 261 months, whereas for the 2 convexity-located tumors, it was 162 and 116 months, with the latter showing growth of more than 15%.

Sequential volumetric data (A), the data for growth rates and results of regression analyses (B), and MR images (C) from a representative case of a 67-year-old man with multiple IDMs. Note that the percentage of growth of cavernous IDMs (red line) is smaller than those of convexity ones (gray and blue lines). Expo. = exponential; Judg. = judgment of growth pattern; p = p value; R² = coefficient of determination; S = statistically significant; TdT = tumor doubling time. Refer to the text for explanation of the dotted lines.

We also analyzed the growth pattern of these tumors (Table 4). Unexpectedly, 60% of skull base IDMs showed an exponential pattern of growth, whereas 32% of non–skull base IDMs showed an exponential pattern. Only 14% of non–skull base tumors showed no trend in growth pattern. In Fig. 2 (dotted lines), we can see that the cavernous tumor fitted better to an exponential curve and the 2 convexity tumors fitted better to a linear pattern using regression analysis. However, it is important to note that the R² for the 3 tumors statistically fit both linear and exponential curves (Fig. 2B). But as described in the methodology, in situations like this, the tumor will be categorized based on the pattern with a larger R².

TABLE 4:

Growth patterns of 71 growing IDMs

Pattern	No. of Cases (%)
	Skull Base	Non–Skull Base
no trend	0 (0)	8 (14)
exponential	9 (60)	18 (32)
linear	6 (40)	30 (54)

The number of patients who became symptomatic and underwent treatment during the follow-up period was also different between the 2 groups (Table 5). Only 1 (2.6%) of 38 patients with skull base IDMs underwent surgery, whereas 6 (8.0%) of 75 patients with non–skull base IDMs needed surgery or radiotherapy.

TABLE 5:

Summary of clinical courses of 113 IDMs

Course	No. of Cases (%)
	Skull Base	Non–Skull Base
remained asymptomatic	37 (97.4)	69 (92.0)
underwent surgery	1 (2.6)	4 (5.3)
underwent radiotherapy	0	2 (2.7)

Biological Differences in Skull Base and Non–Skull Base Meningiomas

To clarify the biological meaning of the volumetric results, we went back to the surgical cases seen in the same period. A summary of demographic data for the 210 consecutive symptomatic meningiomas is shown in Table 6. The mean age was 57.2 years (range 16–90 years). The incidence of male sex (25.7%) seemed higher in this series than in that of the IDMs, but it is comparable to the previously reported incidence.¹⁸ Histological subtypes were also verified and consisted of meningothelial (44.3%), transitional (21.0%), fibrous (19.0%), angiomatous (5.2%), psammomatous (5.2%), and other (5.2%) subtypes. There were 94 skull base and 116 non–skull base meningiomas. No statistical difference was noted according to the age or sex ratio. However, the mean MIB-1 index for skull base tumors was markedly low (2.09%) compared with that for non–skull base tumors (2.74%; p = 0.013; Table 6 and Fig. 3).

TABLE 6:

Biological comparison between skull base and non–skull base symptomatic meningiomas

Variable	All	Skull Base	Non–Skull Base	p Value (statistic)*
no. of cases	210	94	116
age in yrs				0.096 (unpaired t-test)
mean	57.2	55.5	58.6
range	16–90
sex ratio (F:M)	156:54	68:26	88:28	0.635 (Fisher)
MIB-1 index (%)				0.013 (unpaired t-test)†
mean	2.45	2.09	2.74
range	0.1–11.1

* The p values represent the differences between the skull base and non–skull base symptomatic meningiomas.

† Statistically significant.

Graph showing the MIB-1 index in surgical specimens from the skull base and non–skull base groups. The mean MIB-1 index for skull base tumors was markedly low (2.09%) compared with that for non–skull base tumors (2.74%; p = 0.013). Circles show the mean MIB-1 index, and bars show the 95% CI.

Because male sex is a well-known factor that affects the biological behavior of meningiomas in general, the difference between female and male sex was also analyzed. As expected, the highest mean MIB-1 index (3.16%) was seen in males with non–skull base tumors, and the lowest (1.82%) in females with skull base tumors. A statistically significant difference was observed between all male and all female cohorts (p = 0.045), as well as between male and female cohorts with skull base tumors (p = 0.021; Table 7).

TABLE 7:

Relation of MIB-1 index to location and sex

Variable	Male	Female	p Value (unpaired t-test)
no. of cases	54	156
mean age in yrs	59.59	56.42	0.13
mean MIB-1 (%)
all	2.99	2.27	0.045*
skull base	2.82	1.82	0.021*
non–skull base	3.16	2.61	0.249

* Statistically significant.

Discussion

Several reports have been published regarding the natural history of IDMs using various means of growth measurement, some using linear (diameter)^4,9,16,21,23 and some using volumetric^6,14,15,24 methods. According to these studies, the natural history of IDMs can be summarized in 3 points. First, most IDMs (or at least a subset of IDMs) may not grow for a certain period of time. Second, for tumors that grow, the rate of growth seems slow. In our previous report, the mean tumor volume doubling time in growing IDMs was 93.6 months (7.8 years). Finally, hyperintensity on T2-weighted imaging is positively correlated with growth, whereas the presence of calcification is negatively correlated with tumor growth. As to the impact of tumor location on growth behavior, probably because of the small number of patients, this has yet to be reported. In this paper, we found that IDMs in the skull base tended not to grow when compared with those in non–skull base locations. Furthermore, even if these tumors grow, the rate of growth was significantly lower in terms of the percentage growth, annual growth rate, and rates of experiencing symptoms and undergoing treatment.

In this report, we also found that there was a statistically significant difference in the MIB-1 index, which is thought of as a biological marker of cell proliferation, between skull base and non–skull base meningiomas. This is in line with previous studies indicating the existence of biological differences between the two.^8,10,11,20 Kasuya et al.⁸ reported that male sex, the absence of calcification on imaging, and non–skull base location are independent risk factors for a high MIB-1 index by logistic regression analysis among 342 consecutive surgical cases. More recently, McGovern et al.¹¹ made an important note that meningiomas with a non–skull base location are more likely to have a higher MIB-1 index and recur with a higher grade than those within the skull base. Sade et al.²⁰ found that the incidence of Grade II and III tumors is significantly higher outside the skull base (12.1%) than in the skull base (3.5%) from the records of 794 consecutive patients. Mahmood and colleagues¹⁰ likewise noted that among their 319 patients, 25 had Grade II and III tumors, of which, 7 (28%) were located at the skull base and 18 (72%) were outside the skull base. In the latest study from Kane et al.,⁷ based on their statistical observation of 378 surgical cases, they reported that anatomical location is a risk factor for Grade II and III meningiomas.

All these reports, including ours, seem to suggest that non–skull base meningiomas have a more aggressive behavior. In 2003, we proposed that loss of 1p is significantly correlated with malignant progression of meningiomas by analyzing 72 tumors, including WHO Grade II and III tumors, with fluorescence in situ hybridization and loss of heterozygosity analyses.¹³ This small study included 49 Grade I, 15 Grade II, and 8 Grade III tumors. Although the data were not shown in this paper, it was our impression that most Grade II and III tumors are found in a parasagittal (non–skull base) location. Therefore, we decided to categorize these tumors according to location, and we were able to obtain a statistically significant difference in the percentage of cells with 1p deletions (Table 8). We noted that skull base meningiomas harbored a significantly lower percentage of cells with 1p loss (20.31%) compared with non–skull base tumors (37.87%). This observation seems to further suggest that skull base tumors may indeed have fewer genetic aberrations and may have a less aggressive biological nature.

Regarding growth patterns, Nakasu et al.¹⁵ and our group⁶ reported that IDMs do not always grow exponentially and show various patterns of growth including a linear pattern. Based on the rule of proliferation kinetics, tumors with an exponential pattern of growth will maintain their cell-doubling time constant over time. In contrast, those with linear growth patterns will have a reduced cell-doubling time over time. In this study, 60% of skull base IDMs showed an exponential pattern of growth, whereas 33% of non–skull base IDMs showed an exponential pattern. However, these data must be cautiously interpreted because most of the tumors did fit both exponential and linear patterns statistically when regression analysis was performed. In fact, among the 63 tumors with either a linear or exponential growth pattern described in Table 4, only 1 tumor fit an exclusively linear pattern (data not shown). In other words, the data in Table 4 reflect the fact that we took the larger coefficient of determination (R²) in each tumor. Therefore, we believe that it is still too early to conclude that skull base tumors are more likely to present with an exponential pattern of growth. Longer follow-up periods and more cases are still warranted.

We believe that our findings may contribute to the understanding of the natural history of IDMs. It may also impact the way we currently manage IDMs. Those with a skull base location can be observed with sequential follow-up MR images more safely and longer than those with non–skull base tumors, except when they are located at the medial sphenoidal region adjacent to the optic nerve. We may be able to recommend a longer interval of follow-up MR imaging in skull base IDMs than in non–skull base tumors after confirming that IDMs tend not to grow during the early phase of follow-up.

For IDMs that became symptomatic or for uncomplicated symptomatic meningiomas requiring intervention, the findings in this paper may also be useful. As has been reported before, non–skull base meningiomas are more amenable to total resection and have better recurrence-free survival rates,¹¹ whereas total resection of skull base meningiomas may be limited by adjacent critical structures including the brainstem and cranial nerves. Black and colleagues¹ reported a series of 100 patients with skull base meningiomas who were treated with a combination of aggressive surgery and conformal radiation therapy. The authors found that this approach yielded an acceptable functional status in 99% of patients. McGovern et al.¹¹ also reported that adjuvant radiation therapy for skull base meningiomas improved the recurrence-free survival rate of subtotally resected skull base tumors to levels similar to those that were completely resected. Consistent with their suggestion, and taking into consideration the findings in this paper that skull base meningiomas have less aggressive behavior, it seems that maximal surgical reduction limited by the preservation of patient performance status, and the addition of postoperative radiotherapy or radiosurgery,^2,5 intensity-modulated radiotherapy,^12,19 and proton therapy,²² may be an acceptable option for these patients.

Conclusions

A sequential volumetric analysis of 113 IDMs revealed that skull base IDMs tended not to grow compared with non–skull base IDMs. Even when the skull base IDMs grow, the rate of growth is significantly lower than that for non–skull base tumors. Furthermore, a biological comparison of skull base and non–skull base surgically treated meningiomas in 210 consecutive patients showed that the mean MIB-1 index was significantly lower in skull base tumors. These findings may impact our understanding of the natural history of IDMs, as well as the strategies for management and treatment, not only of IDMs, but also of symptomatic meningiomas.

Disclosure

This work was supported in part by Grants-in-Aid No. 22103508 for Scientific Research of Computational Anatomy from the Ministry of Education, Science, Sports and Culture, Japan, to Naoya Hashimoto, M.D., Ph.D.

Author contributions to the study and manuscript preparation include the following. Conception and design: Hashimoto, Yoshimine. Acquisition of data: Hashimoto, Rabo, Okita, Kagawa, Fujimoto, Morii. Analysis and interpretation of data: Hashimoto, Kinoshita, Morii. Drafting the article: Hashimoto, Rabo. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Hashimoto. Statistical analysis: Hashimoto, Rabo. Administrative/technical/material support: Kagawa. Study supervision: Yoshimine.

Acknowledgment

The authors would also like to thank Ms. Mariko Kakinoki for her secretarial assistance.