Acta Neurochir (Wien) (2004) 146: 37–44 DOI 10.

1007/s00701-003-0173-4

Clinico-Pathological Study The Ki-67 proliferation antigen in meningiomas. Experience in 600 cases
F. Roser1, M. Samii1;2 , H. Ostertag3 , and M. Bellinzona1
1 2

Department of Neurosurgery, Klinikum Hannover Nordstadt, Hannover, Germany International Neuroscience Institute, Klinikum Hannover Nordstadt, Hannover, Germany 3 Institute of Pathology, Klinikum Hannover Nordstadt, Hannover, Germany Published online December 22, 2003 # Springer-Verlag 2003

Summary
Background. Meningiomas are mostly benign tumours that can be cured by surgical resection. Because meningiomas tend to recur, long term management in patients with subtotal tumour resection remains controversial. Previous studies have shown that the proliferation potential of meningiomas by Ki-67 labelling indices (LI) might predict their natural history. The purpose of this study was to analyse the reliability of Ki-67-labelling index in predicting the behaviour of meningiomas, and to help the neurosurgeon in establishing better follow up criteria and long term management strategies for these patients. Method. From 1990 to 2000 1328 meningiomas have been operated in our Neurosurgical Department. A total of 600 tumours were examined immunohistochemically using the Mib-1 monoclonal antibody. Clinical charts of the patients including surgical, histological and follow up records, as well as imaging studies were analysed retrospectively. Ki67 LI were correlated with neuroradiological findings, 3D volumetric studies, histological subtype, recurrence-free survival, grade of resection, consistency of tumour tissue, location, osseous involvement, en plaque appearance, vascularity and progesterone-receptor status. Findings. Among the 600 patients analysed, there were 66% females (mean LI 3.8%) and 34% males (mean LI 5.7%), including 20 neurofibromatosis-type-2 (NF-II) patients with a mean LI of 5.2%. Histological grading revealed 91% WHO I meningiomas (mean LI 3.28%), 7% WHO II (mean LI 9.95%) and 2% WHO III (mean LI 12.18%). Labelling indices in recurrent meningiomas increased from initial resection to a fourth local resection. A significant correlation between negative progesteron-receptor status and high tumour vascularity with high Ki67 LI was seen. Ki67 was not a statistically significant predictor of survival time in totally excised WHO I meningiomas. Interpretation. Mib-1 is one important tool in addition to routine histological evaluation, but a combination of clinical factors and particularly the extent of surgical resection, along with the biological features of the tumour, should influence the decision of the neurosurgeon to the patient follow up. Keywords: Meningioma; Mib-1; Ki-67; recurrence; proliferation.

Introduction Although generally considered benign, the biological behaviour of meningiomas varies considerably. Not only histopathological criteria but also the surgical grade of resection determines the recurrence free-survival in patients with histologically comparable meningiomas [3, 18, 34, 48]. Between 7 and 32% of benign meningiomas recur after total resection, and between 19 and 50% after subtotal removal [11, 20]. Because of their clinical behaviour with a tendency to recur and unfavourable outcome after repeated operations, meningiomas cannot be classified as a benign entity despite their pathological classification. Decisions regarding patient management therefore rely on a variety of clinical, radiological and pathological prognosticators. Growth potential of meningioma is variable and even the microscopic morphological classification of the World Health Organization (WHO) cannot predict the clinical behaviour of these tumours [25]. Quantifying their proliferative potential may help to predict the biological behaviour of individual tumours of comparable histology. The prognostic significance of various proliferative indices in meningiomas has already been assessed by other authors, and it has been suggested that the tumour proliferative potential can predict the patient’s clinical course [1, 2, 11, 16, 20, 23, 28, 32, 33, 35, 40, 42, 44]. The nuclear antigen Ki-67 expressed by proliferating cells has become available for routinely

38

F. Roser et al. considered significant at P-value < 0.05. The distribution of the recurrence-free survival times were estimated using the Kaplan-Meier method. Pearson’s regression analysis was performed to determine the correlation between mean Ki-67 LI and other influencing factors. Multivariate analysis (ANOVA) was also performed. Volume measurements were performed with the Image J program (Scion Image, based on NIH Image. Beta Release 4.0.2) scanning adjacent axial, coronal and sagittal slices of post contrast CT=MRI scans throughout the tumour. The growth rates were determined according to the absolute growth rate (cm3=year) calculation [dV (latest-initial)=t]. Recurrence was defined as radiologically detected evidence of regrowth regardless of symptoms.

processed paraffin section [8, 15]. The Mib-1 antibody detects an epitope on the Ki-67 antigen, a nuclear protein present only during active phases of the cell cycle (G1, S, G2 and M). Several studies were done to investigate how Ki-67 labelling indices could help to predict recurrences. However, a lack of adequate case numbers, inconsistent statistical methods and clinical considerations precludes a comparison of these studies. The aim of the present study was to assess if and how the neurosurgeon can rely on the Ki-67 labelling index to use the best follow-up criteria and therapeutic options. Therefore we compared selected histopathological and macroscopical features as well as survival times in meningioma patients. Further comparisons were made for each patient in the recurrent group between the initially resected tumour and all local recurrences to determine whether recurrent meningioma tends to become histologically more aggressive.

Results The mean Mib-1 SI in the 184 male patients was 5.78% (SD 6.31), whereas that of the 370 female patients was 3.87% (SD 6.314). Taking the WHO classification into consideration, this sex-related difference was not significant (Table 1). There were no higher Mib-1 scores observed in younger patients ( < 37 years mean Ki-67 LI 4.2%, SD 4.19) whereas patients with NF-II show significantly higher mean Ki-67 LI (6.2%, SD 3.95). All NF-II cases were excluded from the following statistical evaluations due to their different tumorigenesis. In this series of 600 retrospectively evaluated meningiomas no tumour was embolized pre-operatively. After classifying 580 tumours into 6 typical categories of meningioma appearance, we found no statistically significant relationship between the proliferation index and tumour location (P ¼ 0.273) (Table 2).

Materials and methods
Between 1990 and 2000, 1328 meningiomas have been removed in our Neurosurgical Department. Six hundred paraffin blocks of 554 patients were selected and the paraffin blocks retrieved from the archives. Among our patients there were 184 males and 370 females (ratio 1:2), ranging in age from 15 to 94 (mean age of 64). All patients were followed for a median of 74 months (ranging from 3–228 months) or until death. The material included 20 patients with NF-II disease (for a total of 28 tumours), with a mean age of 32.3 years at presentation (16–60) and a male to female ratio of 1:1. Clinical information of each patient was obtained through review of medical records, as well as follow-up examinations with clinical and neuroradiological evaluation or through detailed questionnaires with radiological reports of the latest MRI findings. Paraffin sections were stained with hematoxylin-eosin, and neoplastic areas were delineated. They were categorized into meningioma subtypes according to the new WHO classification [25]. Immunohistochemistry was performed using 2 mm thick paraffin sections. In short, paraffin sections were dewaxed and stained with Anti-Mib-1 purchased from DAKO [Copenhagen, Denmark], according to standard protocols using the ChemMate Detection Kit Alkaline Phospatase from DAKO [Copenhagen, Denmark] and NeoFuchsin as chromogen. Sections were counterstained with Hemalaun and coverslipped with KP tape [KliniPath BV, Netherlands]. Positive control specimens were run in every essay using glioblastoma tissue. Also, negative controls were run in every staining session. In each case, the entire section was systematically examined using an optical grid on high power fields (400 Â , ZEISS microscope) for the presence of immunoreactivity. In every slide the area of densest staining (‘hot spot’), was searched for and counting was performed in 10 contiguous fields. The average of the results in these fields determined the proliferative index (LI). Only unquestionably stained nuclei were accepted as positively stained. All slides were routinely examined by one pathologist (HO). Randomly chosen stains were also counted by independent researchers to check the interindividual difference of interpretation of the mean Ki-67 LI, and revealed no statistically significant differences. Statistical analysis was performed using the SPSS 10.0 (SPSS Inc.) software program for windows. Several parameters shown in the result section were analysed with Student’s t-test or log-rank tests. Mean value, standard deviation (SD) and P-value were calculated. Differences were

Table 1. Mean Ki-67 LI in meningiomas No. of cases WHO I – Female – Male WHO II – Female – Male WHO III – Female – Male Ki-67 (range) Ki-67 (mean þ SD) 3.54 Æ 4.97 3.30 Æ 4.44 4.37 Æ 4.80 11.9 Æ 8.25 12.27 Æ 8.51 11.67 Æ 8.07 18.2 Æ 9.53 19.0 Æ 10.53 15.8 Æ 9.17

526 370 156 45 18 27 9 3 6

0–58

3–30

5–30

Table 2. Mean Ki-67 LI in different locations (n ¼ 580) Region Convexity CPA Sphenoid wing Frontobasal Petroclival Spinal Number of cases 178 144 105 67 54 32 Mean Ki-67 5.02 4.78 4.27 3.57 3.84 4.65 SD 5.36 6.75 6.13 3.84 3.15 6.62

The Ki-67 proliferation antigen in meningiomas

39

WHO I meningiomas had a 3.54% Ki-67 LI (range 0–58%, SD 4.97). Ki-67 LI in WHO I tumours was significantly lower than in WHO II (atypical) and WHO III (anaplastic), (P < 0.0001). Data are summarized in Table 1. When subdividing WHO I meningiomas according to the WHO histological classification we did not find a significant Ki-67 LI difference in 379 meningotheliomatous (mean Ki-67 LI 3.28%), 80 fibrous (mean Ki-67 LI 3.95%), 54 transitional (mean Ki-67 LI 2.88%), 11 psammomatous (mean Ki-67 LI 1.1%), 7 angiomatous (mean Ki-67 LI 3.0%), 7 clear cell (mean Ki-67 LI 0.7%) or 5 microcystic meningiomas (mean Ki-67 LI 4.4%) (P ¼ 0.087). Patients with first time treated meningiomas (n ¼ 463) were compared with patients with local recurrence of a previous surgically treated meningioma (n ¼ 117). First time treated meningiomas had a mean Ki-67 LI of 3.9% vs. 6.91% in recurrent meningiomas [P < 0.0001]). In 25 patients with WHO I, 6 patients with WHO II and 3 patients WHO III we measured the mean Ki-67 LI from the initial tumour operation to its second, third and in 5 cases to its fourth local recurrence and we saw a progression of the Ki-67 LI in local recurrences in each patient (Fig. 1). There was a transformation from benign to atypical meningioma from second to third local recurrence in 16% of the cases, whereas no dedifferentiation from WHO I to  III or WHO II to WHO III was seen. Tumour size had a mean value of 43.65 cm3 ranging from 3–372 cm3 ; progesterone receptors showed a mean index of 23.8%, ranging from 0–85%); intra-osseous involvement was seen in 167 cases. Male gender (increased time-to-recurrence), vascularity (increased vascularity – increased time-torecurrence) and progesterone-receptor status (high PR-Status – increased time-to-recurrence) have been found to be significantly related to time-to-recurrence (ANOVA, p< 0.0005).

We then looked at correlations of these significant variables to see which ones might be redundant in any model to account for these parameters. Increased vascularity was associated with significant increases in progesterone staining and decreases in WHO index and Ki67 labelling. In addition, increased WHO index was associated with decreased PR-Status and increased Ki-67 LI. Finally, decreased progesterone staining was associated with increased Ki-67 LI. In 38 patients we saw a significant correlation between the neuroradiologically assessed tumour growth and the proliferative activity of the tumour (Fig. 2). Analysis of the dependence of time-to-recurrence and grade of resection was performed. Separation of all WHO I meningiomas with recurrences (n ¼ 120) in four different resection grades according to the Simpson classification (S ) showed that time-to-recurrence in completely resected meningiomas depend on the

Fig. 2. Correlation of mean Ki-67 LI and neuroradiological growth rate in WHO I meningiomas (n ¼ 38; r ¼ 0.34; p ¼ 0.0001)

Fig. 1. Individual proliferation in local recurrence of meningioma patients

Fig. 3. Recurrence free survival for WHO I and S I resected meningiomas

40

F. Roser et al.

resection grade (S I 67.3 month, S II 50.0 month, S III 46.3 month and S IV 40.3 month). The mean Ki-67 LI was independent of the resection grade (S I 5.75%, S II 4.2%, S III 6.76% and S IV 7.99%). For survival analysis WHO II and  III meningiomas, as well as partially resected meningiomas, were excluded. In WHO I and S I resected meningiomas we did not find any significant difference in survival according to different mean Ki-67 LI (n ¼ 169, Pearson correlation R ¼ 0.176, P ¼ 0.36). A cut-off point of 4% was set as the mean value of the mean Ki-67 indices in the analysed group (Fig. 3). Conclusion Meningiomas are mostly benign tumours that usually do not invade the brain parenchyma. The WHO grading system aiming to describe different types of tumours, frequently fails to determine the clinical behaviour of meningiomas. Even in cases of complete removal according to the Simpson classification the chance of recurrence is high [48]. Pathological specimens in those cases do not show either atypical features or histopathological signs of increased biological activity. Proliferation markers like BrdU or AgNOR have been used to describe the clinical course in meningioma patients [5, 27, 45]. New biochemical markers like Topo-isomerase II-, telomerase or apoptotic fragmentation [13, 29, 31, 50] reflect meningioma behaviour in conjunction with morphological grading, but they all describe the biological activity in atypical and anaplastic meningiomas better than predicting the recurrence potential in benign meningiomas. The Mib-1 monoclonal antibody, staining the Ki-67 antigen, can be used on paraffin embedded specimens. Its reliability in analysing meningioma growth and recurrence has been shown by several authors [2, 33, 39, 40, 44]. The purpose of this study was to evaluate the value of the Ki-67 antigen in predicting the behaviour of meningiomas. Our data suggest that there is no statistically significant correlation between mean Ki-67 LI and recurrence free survival in patients harbouring a benign meningioma and who underwent radical tumour resection (Fig. 3). We excluded atypical and anaplastic meningiomas from survival analysis because aggressive behaviour of these meningiomas with dedifferentiation at time of recurrence is well described through morphological studies. In terms of true recurrence, incompletely resected meningiomas with infiltrated dura left in place or subtotally removed tumour tissue, which will certainly

regrow over time, would not fit into the concept of recurrence after radical resection and had been excluded from the analysis as well. Even with these stringent selection criteria a large sample could be meaningfully analysed as we disposed of a very large number of cases from the outset. Møller separated different resection grades and histology before survival analysis in 25 cases and could not give any significant prediction with Ki-67 immunohistochemistry as well [35]. Abramovich reported the absence of a significant difference in proliferation in 59 meningioma patients who had been radically operated upon. He documented higher indices in the recurrent group and recognized a clear overlap of the index range between the groups [2]. These and our results differ from most of the other studies showing a strong correlation between mean Ki67 LI and recurrence free survival. Ohta et al. showed a correlation with recurrence free interval in 42 patients, but the follow-up time was 60 month and the statistics were performed without dividing the meningiomas according to the WHO classification [42]. Perry et al. showed values for interpretation of borderline atypical meningiomas in 425 meningiomas through multivariate analysis. Brain invasion, mitosis count > 3=10 and a mean Ki-67 LI > 4.2% correlated to decreased recurrent free survival. Based on ‘‘gross total resection’’ the statistics summarize all grades of surgical excision, comparing a disproportionate group of 389 patients with Ki-67 LI < 4.2% versus 33 patients with Ki-67 LI > 4.2%. WHO I meningiomas alone show an even more disproportionate ratio of 325 vs. 15 patients [44]. Perry et al. states that the Ki-67 LI influences survival, but he takes all WHO grades together and makes no mention of Ki-67 LI influence on survival with WHO I alone. Our statistically more homogenous group has 169 patients with Ki-67 LI < 4.2% versus 23 patients with Ki-67 LI > 4.2%, all with WHO I and S I resection grade in survival analysis. Matsuno et al. and Nakasu et al. reported a 3.2% and 3% cut-off point for higher recurrence tendency. However 20% to 50% of the recurrent meningioma group were atypical, and all non-recurrent meningiomas were WHO I. Survival analysis was performed taking all patients together without dividing for surgical resection or histological grade [32, 40]. The small number of patients in each group and the uneven distribution of histological grades might have additionally influenced the mean LI and the survival analysis. Failing to separate different histopathological and resection grades expose these studies to the criticism

The Ki-67 proliferation antigen in meningiomas

41

that Ki67 may simply describe a well known phenomenon, in other words that more aggressive histology and less aggressive regression correlate to shorter disease free survival times. The strength of our data lies furthermore in the fact, that all meningiomas were treated surgically by one experienced neurosurgeon (MS), morphologically diagnosed by one pathologist (HO) and stained by the same method over a period of ten years. Definition of radical tumour removal and interpretation of positive tissue staining are therefore easier to achieve. The proliferation index did not show any differences concerning sex or age analysis, but in NF-II patients showed significant higher levels. This may reflect differences in molecular biology between sporadic and NF-II meningiomas and may be related to an earlier onset, multiplicity or more aggressive behaviour of NF-II tumours [4]. The proliferation index in 45 atypical and 12 anaplastic meningiomas in our series was higher than in benign meningiomas and reflects the aggressive histopathology described as increased vascularity and mitosis, a loss of architectural pattern, prominent nucleoli, nuclear polymorphism and necrosis [1, 12, 19, 25] (Table 1). Although embolisation of meningiomas has been performed for many years as a preoperative adjunct to reduce tumour vascularity and facilitate surgical excision, it may potentially cause an erroneous diagnosis of a high-grade lesion due to the fact that it may produce tumour necrosis. Increase in MIB-1 labelling indices in those tumours exhibiting necrotic foci has been demonstrated, but it did not have any prognostic significance [36, 41, 43]. Genetic alterations can explain biological progression. Next to chromosome 22 anomalies, deletions of the short arm of chromosome 1 have previously been described as the most frequent alterations detected by cytogenetic analysis of meningiomas [6]. Deletion of 1p and therefore enzyme activity loss of tissue non-specific alkaline phosphatase (ALPL) has been proposed to be associated with the development of atypical and anaplastic meningiomas [38]. The frequency of loss of heterogeneity on chromosomes 1p, 10q and 22 increases with tumour grade, which would support the concept that aggressive meningiomas develop through tumour progression [47, 51]. Attempts were also made to predict recurrence in benign meningiomas, combining cytogenetics and histology through a multimodal approach [24]. Some investigators reported a higher recurrence of meningiomas at distinct locations (parasagittal, sphenoid

ridge) while others did not. [10, 19, 34]. In our study there was no correlation between tumour location and mean Ki-67 LI. The recurrence appears to be influenced by accurate microsurgical removal, and this may be location dependent. Literature data support the view that tumour doubling time and mean Ki-67 LI in subtotal resected meningiomas correlate [21, 39]. In these studies initial and recurrent tumour were not necessarily from the same patient, therefore no conclusion about increased growth fraction in the individual patient could be made [33]. We found a strong correlation between neuroradiologically measured growth rates and mean Ki-67 LI, and for the first time increased growth rates and mean Ki-67 LI in local recurrencies in the individual patient. Even though we analysed a higher number of cases for this purpose, case numbers were still too small to statistically separate different stages of surgical resection [39]. The definition of tumour recurrence in a meningioma is controversial throughout the literature. The term ‘‘radical removal’’ is a surgical definition, therefore subjective so that tumour recurrence may in fact be a regrowth. Post-operative MRI studies with contrast medium can minimise this error and show the actual extent of tumour removal. Regrowth after surgical intervention can be silent for a long period of time until the tumour reaches a substantial size while others will be brought to attention sooner after compressing important structures like brain stem, optic chiasm or sagittal sinus in parasagittal meningiomas [46]. Besides the present study, we found another series where recurrence is defined based on radiological evidence of regrowth [40]. Other investigators have stated that a radiologically detected regrowth that does not require intervention does not count as true recurrence [13]. This is correct in terms of clinical follow-up, but statistical analysis of morphological features needs an accurate definition of recurrence. The significant correlation of high vascularity and high Ki-67 LI suggests the neo-angeogenetic capability of meningiomas and might account for higher proliferation in these tumours as reflected by Ki-67 [7, 17]. It has been shown that high progesterone receptor status in benign meningiomas is associated with lower recurrence rate and vice versa [9, 14]. Our data support this negative correlation of high proliferation index in meningiomas and low progesterone receptor level showing ‘‘protection’’ against recurrence. In our series we observed elevated proliferative activity in recurrent meningiomas throughout the histological

42
Table 3. Meningioma recurrence determined by mean Ki-67 LI. Review of the literature Literature No. of patients 600 139 59 22 60 37 66 85 425 127 160 F=U time (month) 73.5 81 109 64 96 120 – 50.4 106.8 38 36 Mean Ki-67 LI (%) initial 3.9 2.06 1.5 1.8 1.06 3.8 7.9 0.6 1.5 1.6 1.12

F. Roser et al.

Mean Ki-67 LI (%) recurrence 6.9 4.37 4.7 3.0 3.12 7.1 7.42 1.1 4.2 3.6 4.04

Present study 2002 Nakasu et al., Am J Surg Pathol 2001 Abramovich et al., Arch Path Lab Med 1999 Nakaguchi et al., Cancer 1999 Karamitopolou et al., Human Pathology 1998 Abramovich et al., Human Pathology 1998 Madsen et al., Clin Neuropathology 1997 Møller et al., J Neuro-Oncol 1997 Perry et al., Cancer 1997 Matsuno et al., Acta Neuropath 1996 Kolles et al., Acta Neurochir 1995

grouping, like most other studies had shown [1, 2, 32, 33, 35, 40, 42]. A decrease of the LI has only been reported by Madsen et al. [30] (Table 3). Caution is recommended when using the proliferation index as the sole prognostic indicator or as a substitute for morphological diagnosis due to the overlap in each group (WHO I mean Ki-67 LI 0–30%, WHO II mean Ki-67 LI 3–35% or WHO III mean Ki-67 LI 5–58%) [1, 22, 26, 28]. This might be due to the heterogeneity of biological activity within the tumour tissue [2, 46]. Proliferating cells in all histological grades were found to be distributed heterogeneously throughout the tumour, especially in recurrences [42]. It is debatable that the focal accumulation of proliferation may affect tumour recurrence for the ‘‘highest area’’ counting method [40]. We used the ‘‘highest area’’ counting method to minimize missing of focal accumulation of biological activity within the meningioma tissue, as it is recommended by Nakasu et al. [40]. This is an important issue when considering modern neurosurgical techniques, where only small parts of tumour with unknown precise location of the tissue reaches the pathologist. Indeed it is debatable whether recognized areas of high mitotic activity in meningiomas determined with either counting method, reflect the proliferation status of the whole tumour. Studies assessed the expression of leucocyte integrins and macrophage-associated antigens in meningiomas to detect any influence on proliferation capability. Evidence of a correlation between the Ki-67 proliferation index and macrophage infiltration could not be provided, although it has been suggested that the expression of leucocyte antigens could play a role in the attraction of immunocompetent cells in the stroma of meningiomas. As proliferation cells in meningiomas are spread heterogeneously within the tumour and macrophage= lymphocyte infiltration has been reported to be mainly at the tumour margins, only double labelling methods

could give insight into this phenomenon and exclude errors in statistical evaluation [37]. In our series more than 80 patients have a mean Ki-67 LI of 1% with recurrence times ranging from 11 to 148 month. Therefore a cut-off value over which a tumour becomes suspicious cannot be given. Precise values from different laboratories are not applicable to other institutions because of differences in methodology, counting procedures and interpretation of the results as reflected by the relatively wide range of initial and recurrent Ki-67 LI determined by several investigators (Table 3). Despite our single center study individual set cut-off points even for our Neurosurgical Department would be questionable due to the heterogeneity of meningiomas in the different groups (resection grade, histology, Ki-67). In histopathological borderline cases, with some but not convincing aspects of atypia, the Ki-67 LI, combined with the routine histopathological workup can provide more insight in to the behaviour of a meningioma, particularly in the presence of high vascularity, low PRstatus, subtotal resection and recurrence [44]. High scores are worrisome and should lead to a more close follow up for evidence of recurrent tumour, but the confidence in high LI should not interfere with the decisions for treatment plans as some authors recommend [32, 40, 49]. Further studies are needed to establish new indicators that can describe meningioma behaviour in a reliable and predictive way.

References
1. Abramovich CM, Prayson RA (1998) MIB-1 labeling indices in benign, aggressive, and malignant meningiomas: a study of 90 tumors. Hum Pathol 29: 1420–1427 2. Abramovich CM, Prayson RA (1999) Histopathologic features and MIB-1 labeling indices in recurrent and nonrecurrent meningiomas. Arch Pathol Lab Med 123: 793–800

The Ki-67 proliferation antigen in meningiomas 3. Adegbite AB, Khan MI, Paine KW, Tan LK (1983) The recurrence of intracranial meningiomas after surgical treatment. J Neurosurg 58: 51–56 4. Antinheimo J, Haapasalo H, Haltia M, Tatagiba M, Thomas S, Brandis A, Sainio M, Carpen O, Samii M, Jaaskelainen J (1997) Proliferation potential and histological features in neurofibromatosis 2-associated and sporadic meningiomas. J Neurosurg 87: 610–614 5. Bagni A, Botticelli A, Martinelli A, Azzoni P, Trentini GP (1991) AgNOR counts in recurrent and non-recurrent meningiomas. Histopathology 19: 465–467 6. Bello MJ, de Campos JM, Kusak ME, Vaquero J, Sarasa JL, Pestana A, Rey JA (1994) Allelic loss at 1p is associated with tumor progression of meningiomas. Genes Chromosomes Cancer 9: 296–298 7. Bitzer M, Opitz H, Popp J, Morgalla M, Gruber A, Heiss E, Voigt K (1998) Angiogenesis and brain oedema in intracranial meningiomas: influence of vascular endothelial growth factor. Acta Neurochir (Wien) 140: 333–340 8. Boker DK, Stark HJ, Gullotta F, Nadstawek J, Schultheiss R (1990) Immunohistochemical demonstration of the KI-67-antigen in paraffin- embedded tumor biopsies. Clin Neuropathol 9: 51–54 9. Brandis A, Mirzai S, Tatagiba M, Walter GF, Samii M, Ostertag H (1993) Immunohistochemical detection of female sex hormone receptors in meningiomas: correlation with clinical and histological features. Neurosurgery 33: 212–217 10. Christensen D, Laursen H, Klinken L (1983) Prediction of recurrence in meningiomas after surgical treatment. A quantitative approach. Acta Neuropathol (Berl) 61: 130–134 11. Crompton MR, Gautier-Smith PC (1970) The prediction of recurrence in meningiomas. J Neurol Neurosurg Psychiatry 33: 80–87 12. de la Monte SM, Flickinger J, Linggood RM (1986) Histopathologic features predicting recurrence of meningiomas following subtotal resection. Am J Surg Pathol 10: 836–843 13. Falchetti ML, Pallini R, Larocca LM, Verna R, D’Ambrosio E (1999) Telomerase expression in intracranial tumours: prognostic potential for malignant gliomas and meningiomas. J Clin Pathol 52: 234–236 14. Fewings PE, Battersby RD, Timperley WR (2000) Long-term follow up of progesterone receptor status in benign meningioma: a prognostic indicator of recurrence? J Neurosurg 92: 401–405 15. Gerdes J, Becker MH, Key G, Cattoretti G (1992) Immunohistological detection of tumour growth fraction (Ki-67 antigen) in formalin-fixed and routinely processed tissues. J Pathol 168: 85–86 16. Hsu DW, Efird JT, Hedley-Whyte ET (1998) MIB-1 (Ki-67) index and transforming growth factor-alpha (TGF alpha) immunoreactivity are significant prognostic predictors for meningiomas. Neuropathol Appl Neurobiol 24: 441–452 17. Izycka-Swieszewska E, Rzepko R, Borowska-Lehman J, Baranowska E, Warzocha D (1999) Recurrent meningiomas – the immunohistochemical analysis of angiogenesis and cellular proliferation. Preliminary study. Folia Neuropathol 37: 179–184 18. Jaaskelainen J (1986) Seemingly complete removal of histologically benign intracranial meningioma: late recurrence rate and factors predicting recurrence in 657 patients. Surg Neurol 26: 461–469 19. Jaaskelainen J, Haltia M, Servo A (1986) Atypical and anaplastic meningiomas: radiology, surgery, radiotherapy, and outcome. Surg Neurol 25: 233–242 20. Jellinger K, Slowik F (1975) Histological subtypes and prognostic problems in meningiomas. J Neurol 208: 279–298 21. Kakinuma K, Tanaka R, Onda K, Takahashi H (1998) Proliferative potential of recurrent intracranial meningiomas as evaluated by labelling indices of BUdR and Ki-67, and tumour doubling time. Acta Neurochir (Wien) 140: 26–31

43
22. Karamitopoulou E, Perentes E, Diamantis I, Maraziotis T (1994) Ki-67 immunoreactivity in human central nervous system tumors: a study with MIB 1 monoclonal antibody on archival material. Acta Neuropathol (Berl) 87: 47–54 23. Karamitopoulou E, Perentes E, Tolnay M, Probst A (1998) Prognostic significance of MIB-1, p53, and bcl-2 immunoreactivity in meningiomas. Hum Pathol 29: 140–145 24. Ketter R, Henn W, Niedermayer I, Steilen-Gimbel H, Konig J, Zang KD, Steudel WI (2001) Predictive value of progression-associated chromosomal aberrations for the prognosis of meningiomas: a retrospective study of 198 cases. J Neurosurg 95: 601–607 25. Kleihues P, Burger PC, Scheithauer B (1993) The new WHO classification of brain tumors. Brain Pathol 3: 255–268 26. Kolles H, Niedermayer I, Schmitt C, Henn W, Feld R, Steudel WI, Zang KD, Feiden W (1995) Triple approach for diagnosis and grading of meningiomas: histology, morphometry of Ki-67= Feulgen stainings, and cytogenetics. Acta Neurochir (Wien) 137: 174–181 27. Kunishio K, Ohmoto T, Matsuhisa T, Maeshiro T, Furuta T, Matsumoto K (15-4-1994) The significance of nucleolar organizer region (AgNOR) score in predicting meningioma recurrence [see comments]. Cancer 73: 2200–2205 28. Langford LA, Cooksley CS, DeMonte F (1996) Comparison of MIB-1 (Ki-67) antigen and bromodeoxyuridine proliferation indices in meningiomas. Hum Pathol 27: 350–354 29. Langford LA, Piatyszek MA, Xu R, Schold SC Jr, Wright WE, Shay JW (1997) Telomerase activity in ordinary meningiomas predicts poor outcome. Hum Pathol 28: 416–420 30. Madsen C, Schroder HD (1997) Ki-67 immunoreactivity in meningiomas – determination of the proliferative potential of meningiomas using the monoclonal antibody Ki-67. Clin Neuropathol 16: 137–142 31. Maier H, Wanschitz J, Sedivy R, Rossler K, Ofner D, Budka H (1997) Proliferation and DNA fragmentation in meningioma subtypes. Neuropathol Appl Neurobiol 23: 496–506 32. Matsuno A, Fujimaki T, Sasaki T, Nagashima T, Ide T, Asai A, Matsuura R, Utsunomiya H, Kirino T (1996) Clinical and histopathological analysis of proliferative potentials of recurrent and non-recurrent meningiomas. Acta Neuropathol (Berl) 91: 504–510 33. Matsuno A, Nagashima T, Matsuura R, Tanaka H, Hirakawa M, Murakami M, Tamura A, Kirino T (1996) Correlation between MIB-1 staining index and the immunoreactivity of p53 protein in recurrent and non-recurrent meningiomas. Am J Clin Pathol 106: 776–781 34. Mirimanoff RO, Dosoretz DE, Linggood RM, Ojemann RG, Martuza RL (1985) Meningioma: analysis of recurrence and progression following neurosurgical resection. J Neurosurg 62: 18–24 35. Moller ML, Braendstrup O (1997) No prediction of recurrence of meningiomas by PCNA and Ki-67 immunohistochemistry. J Neurooncol 34: 241–246 36. Morimura T, Takeuchi J, Maeda Y, Tani E (1994) Preoperative embolization of meningiomas: its efficacy and histopathological findings. Noshuyo Byori 11: 123–129 37. Mosnier JF, Perret AG, Scoazec JY, Brunon J (2000) Expression of beta2 integrins and macrophage-associated antigens in meningeal tumours. Virchows Arch 436: 131–137 38. Muller P, Henn W, Niedermayer I, Ketter R, Feiden W, Steudel WI, Zang KD, Steilen-Gimbel H (1999) Deletion of chromosome 1p and loss of expression of alkaline phosphatase indicate progression of meningiomas. Clin Cancer Res 5: 3569–3577 39. Nakaguchi H, Fujimaki T, Matsuno A, Matsuura R, Asai A, Suzuki I, Sasaki T, Kirino T (15-5-1999) Postoperative residual tumor growth of meningioma can be predicted by MIB-1 immunohistochemistry. Cancer 85: 2249–2254

44
40. Nakasu S, Li DH, Okabe H, Nakajima M, Matsuda M (2001) Significance of MIB-1 staining indices in meningiomas: comparison of two counting methods. Am J Surg Pathol 25: 472–478 41. Ng HK, Poon WS, Goh K, Chan MS (1996) Histopathology of postembolized meningiomas. Am J Surg Pathol 20: 1224–1230 42. Ohta M, Iwaki T, Kitamoto T, Takeshita I, Tateishi J, Fukui M (15-12-1994) MIB1 staining index and scoring of histologic features in meningioma. Indicators for the prediction of biologic potential and postoperative management. Cancer 74: 3176–3189 43. Paulus W, Meixensberger J, Hofmann E, Roggendorf W (1993) Effect of embolisation of meningioma on Ki-67 proliferation index. J Clin Pathol 46: 876–877 44. Perry A, Stafford SL, Scheithauer BW, Suman VJ, Lohse CM (1-61998) The prognostic significance of MIB-1, p53, and DNA flow cytometry in completely resected primary meningiomas. Cancer 82: 2262–2269 45. Shibuya M, Hoshino T, Ito S, Wacker MR, Prados MD, Davis RL, Wilson CB (1992) Meningiomas: clinical implications of a high proliferative potential determined by bromodeoxyuridine labeling. Neurosurgery 30: 494–497 46. Siegers HP, Zuber P, Hamou MF, van Melle GD, de Tribolet N (1989) The implications of the heterogeneous distribution of Ki-67 labelled cells in meningiomas. Br J Neurosurg 3: 101–107 47. Simon M, von Deimling A, Larson JJ, Wellenreuther R, Kaskel P, Waha A, Warnick RE, Tew JM Jr, Menon AG (15-10-1995) Allelic losses on chromosomes 14, 10, and 1 in atypical and malignant meningiomas: a genetic model of meningioma progression. Cancer Res 55: 4696–4701

F. Roser et al.: The Ki-67 proliferation antigen in meningiomas 48. Simpson D (1957) The recurrence of intracranial meningiomas after surgical treatment. J Neurol Neurosurg Psychiatry 20: 22–39 49. Takeuchi H, Kubota T, Kabuto M, Kitai R, Nozaki J, Yamashita J (1997) Prediction of recurrence in histologically benign meningiomas: proliferating cell nuclear antigen and Ki-67 immunohistochemical study. Surg Neurol 48: 501–506 50. Tanaka M, Kubo O, Tajika Y, Takahashi T, Suzuki S, Takakura K (1999) [Immunohistochemical evaluation of intracranial recurrent meningiomas: correlation of topoisomerase II alpha expression and cell proliferative potential]. No To Shinkei 51: 1033–1039 51. Weber RG, Bostrom J, Wolter M, Baudis M, Collins VP, Reifenberger G, Lichter P (23-12-1997) Analysis of genomic alterations in benign, atypical, and anaplastic meningiomas: toward a genetic model of meningioma progression. Proc Natl Acad Sci USA 94: 14719–14724

Comment
This work on a large series of patients confirms the value of the Ki-67 labelling index for the classification of meningiomas. However, it can not serve as a predictive factor on its own. N. de Tribolet

Correspondence: Florian Roser M.D., Department of Neurosurgery, Klinikum Hannover Nordstadt, Haltenhoffstr. 41, 30167 Hannover, Germany. e-mail: f.roser@gmx.de