Neurology India Medknow Publications on behalf of the Neurological Society of India ISSN: 0028-3886 EISSN: 1998-4022 Vol. 51, Num. 3, 2003, pp. 336-340

Neurology India, Vol. 51, No. 3, July, 2003, pp. 336-340

Diagnostic validity of the Ki-67 labeling index using the MIB-1 monoclonal antibody in the grading of meningiomas

Devaprasath A, Chacko G
Neuropathology, Department of Neurological Sciences, Christian Medical College and Hospital Vellore - 632004

Correspondence Address:
Neuropathology, Department of Neurological Sciences, Christian Medical College and Hospital, Vellore - 632004
geetachacko@cmcvellore.ac.in

Code Number: ni03111

ABSTRACT

Background: About 10% of meningiomas behave aggressively and are graded atypical or malignant with important therapeutic and prognostic implications. Routine histological parameters are inconsistent in the assessment of their aggressive behavior. Aims: The aim of this study was to find a threshold level of the MIB-1 labeling index (MIB-1 LI) with the highest diagnostic validity in predicting histological atypia in a meningioma. Setting and Design: This was a retrospective study of all atypical and malignant meningiomas diagnosed at our center between January 1995 and June 2000 and which were identified from the General Pathology Registry. Material and Methods: These meningiomas were assessed histologically with respect to the individual criteria of atypia. They were categorized according to the WHO 2000 classification as benign, atypical and anaplastic meningiomas, WHO Grades I, II and III respectively and by immunohistochemical analysis using the MIB-1 monoclonal antibody. Statistical Analysis: The diagnostically useful cut-off level for the prediction of atypia was estimated by calculating the sensitivity and specificity of the MIB-1 LI at various levels and a receiver operated characteristic (ROC) analysis was performed. The correlation between the individual histological parameters was studied and the MIB-1 LI was obtained using Fisher's exact test. Results: Of the 40 meningiomas studied 21 were benign, 16 atypical and 3 anaplastic. Atypical tumors had a higher MIB-1 LI than benign tumors, with diagnostic validity highest at a threshold of 7%, with a sensitivity of 0.86 and a specificity of 0.93, giving a likelihood ratio of 17. The MIB-1 LI correlated well with mitotic activity and the other individual criteria in the WHO 2000 definition of atypia in a meningioma. MIB-1 LI did not, however, correlate well with brain invasion. Conclusion: The MIB-1 LI has the highest validity in the diagnosis of atypia in meningiomas at a threshold level of 7%. The MIB-1 LI used in conjunction with histological features can help in making a recommendation regarding potentially aggressive behavior in meningiomas.

INTRODUCTION

The majority of meningiomas are considered benign and slow-growing tumors, however, about 20% of them can have a more aggressive course manifesting usually as local recurrences.[1]

Decisions regarding patient management rely on a variety of clinical, radiological and histological prognosticators. Of these, the extent of surgical resection plays the most important role, as tumors that are incompletely resected often recur regardless of other considerations.[2] Nonetheless, some histological features are associated with tumor recurrence despite gross total resection (GTR). The WHO 2000 classification provides histological criteria grading meningiomas as benign, atypical or malignant with reasonably good prognostic correlation.[3] However, the histological diagnosis of atypia remains controversial with significant inter-observer variability that creates difficulties in assessing the prognosis and postoperative management of patients. A number of studies have examined histological features of aggressive meningiomas in an attempt to define a phenotype that is predictive of such behavior.[3] Although certain features have been identified as being more frequently observed in these tumors, unfortunately not all aggressive tumors demonstrate these characteristics. Borderline cases exist in all grading schemes and there still remains a group of cases that show clinically aggressive behavior despite being histologically benign (WHO Grade I) and as many as 7-20% of benign (WHO Grade I) tumors are known to recur.[3]

Ancillary techniques would therefore result in improved diagnostic and prognostic accuracy. Several studies have attempted to characterize the behavior of meningiomas by measuring cell proliferation indices and tumor markers. These include the percentage of S-phase or S+G2M phases, Bromodeoxyuridine (BrdU) Labeling index (LI), Ki-67 or MIB-1 Labeling index (MIB-1 LI), and counts of Argyrophilic Nucleolar Organizer Regions (AgNORS).[4],[5],[6],[7],[8],[9],[11] Of these, the MIB-1 LI has been noted to have strong associations with recurrence-free survival (RFS) and histological grade in meningiomas.[2],[7],[8],[9],[10] The aim of our study was to find a threshold level of the MIB-1 LI with the highest diagnostic validity in predicting histological atypia in a meningioma. Our focus was on the use of the MIB-1 LI as an adjunct to histology.

MATERIAL AND METHODS

All meningiomas diagnosed as atypical or malignant between January 1995 and June 2000 were identified from the General Pathology registry. Tumors for which blocks or slides were unavailable were excluded from the study. Forty cases were analyzed histologically and immunohistochemically. Slides of the 40 cases were studied with respect to the individual criteria of atypia. They were categorized according to the WHO 2000 classification as benign, atypical and anaplastic meningiomas, WHO Grades I, II and III respectively.3 Individual histological features were recorded as present, even if seen focally. According to the WHO criteria, brain invasion is not considered a feature of atypia and is defined as irregular groups of tumor cells infiltrating the adjacent cerebral parenchyma without an intervening layer of leptomeninges, often causing reactive astrogliosis. All cases were examined for evidence of brain invasion.

Immunohistochemical staining by the avidin-peroxidase technique using monoclonal antibody to MIB-1 (DAKO Patts, Denmark) was performed as detailed below. Representative 5µ sections of each case were mounted on poly-L-Lysine coated slides and incubated overnight at 370C. Sections of tonsil were used as positive control. Negative controls were achieved by omitting the primary antibody. Sections were deparaffinized followed by rehydration in decreasing ethanol concentrations and placed in distilled water. Antigen retrieval was done by pressure cooking.13 10mM sodium citrate (pH 6.0) was boiled in a pressure cooker. The slides were placed in the boiling buffer and the lid of the pressure cooker closed. The cooker was allowed to come to full pressure. After 2 minutes of boiling at full pressure, the cooker was cooled by quenching in cold running tap water. The slides were then removed quickly and placed in distilled water for 2 changes and allowed to cool to room temperature, and then transferred to 0.05 M Tris (Aldrich) buffered saline (pH 7.4), care being taken to ensure that the slides did not dry. The slides were individually drained of buffer, by gentle tapping, followed by careful wiping of excess liquid around the section. The sections were then covered with normal human pooled serum (1:5 dilution, Institutional Blood Bank) and incubated for 15 minutes. All excess liquid was drained off the slide by gentle tapping and the sections covered with the diluted MIB-1 monoclonal antibody (DAKO Patts, Denmark), at a dilution of 1:75 and incubated for 30 minutes at room temperature (240C). The slides were then rinsed in Tris-buffered saline 3 times for 5 minutes each. The sections were then drained and covered by diluted secondary antibody, biotinylated rabbit anti mouse (1:200 dilution, DAKO Patts, Denmark) and incubated for 30 minutes at room temperature. The slides were then rinsed in Tris-buffered saline 3 times for 5 minutes each. Endogenous peroxidase was blocked with 0.5% hydrogen peroxidase (Qualigens) in methanol by incubating the slides covered with solution for 30 minutes. The sections were rinsed in Tris-buffered saline 3 times, for 5 minutes each. The sections were then drained and covered with peroxidase conjugated avidin (1:200 dilution, DAKO Patts, Denmark) and incubated for 30 minutes. The slides were rinsed with 3 changes of Tris-buffered saline for 5 minutes each. The slides were then developed using freshly prepared diaminobenzidine tetrahydrochloride solution (DAKO Patts, Denmark) containing hydrogen peroxide, for 10 minutes. At this point positive controls were checked to ascertain the end of incubation. The sections were counterstained with Harris Hematoxylin for 10 seconds. The sections were then dehydrated, cleared, and mounted with DPX as mounting medium. Cells with brown nuclei were considered immunopositive for MIB-1. The cell counts were performed in regions of maximum immunoreactivity using a Leitz dialux microscope (Leitz Westlar, Germany) under high power objective (magnification 400x), without knowledge of the histological grade. An average of 1000 cells were counted and the MIB-1 LI determined and expressed as the percentage of labeled nuclei. In order to estimate a diagnostically useful cut-off level for the prediction of atypia, the sensitivity and specificity of the MIB-1 LI at various levels was calculated and a receiver operated characteristic (ROC) analysis performed. The correlation between the individual histological parameters studied and the MIB-LI was obtained using Fisher's exact test.

RESULTS

There were a total of 223 meningiomas diagnosed between January 1995 and June 2000, of which 40 had been previously diagnosed as atypical or malignant meningiomas. On reviewing the slides using the criteria recommended by the WHO 2000 classification, 21 (52.5%) of these 40 were reclassified as benign, WHO Grade I tumors, 16 (7.17%) as atypical WHO Grade II tumors and 3 (1.34%) as anaplastic WHO Grade III meningiomas. The general and histopathological features of the tumors studied are summarized in [Table-1]. The three grades of tumor were distributed using arbitrary cut-offs of the MIB-1 LI as shown in [Table-2].

The ROC analysis [Figure-1] revealed highest validity at a cut-off value of 7% with a sensitivity of 0.86 and a specificity of 0.93 giving a likelihood ratio of 17. Tumors with a MIB-1 LI of >7% were 17 times more likely to be atypical than benign. The distribution of meningiomas at the estimated MIB-1 LI cut-off level of 7% is represented pictorially in [Figure-2]. Three cases that were histologically WHO Grade I had MIB-1 LI of > 7%. The first patient underwent a total excision of a parietal convexity meningioma. Histologically, the mitotic index was 2/10hpf. There were no other features to suggest histological atypia. MIB-1 LI was 7.5%. At the 6-month follow-up there was no residual tumor seen on imaging studies with contrast. A recurrence of tumor was seen on imaging studies at the two-and-a half-year follow-up. The second case was that of a right frontal convexity mass that underwent total excision. The mitotic index was 1/10hpf. No histological features of atypia were seen. The MIB-1 LI was 9.2%. No recurrence was noted at the last follow-up, 2 years postoperatively. The third patient underwent near total excision of a right pterional mass and was given radiosurgery for residual tumor. Histologically, there was evidence of necrosis and the mitotic index was 2-3/10hpf. The MIB-1 LI was 12.3%. A large recurrence was seen four-and-a-half years after the first surgery. The patient underwent a second surgery. The MIB-1 LI at this time was 18.1%. Radiation therapy and chemotherapy were given after the second surgery. At the last follow-up, 12 months after the second surgery, there was no recurrence of tumor. The correlation of MIB-1 with individual histological characteristics using the Fisher exact test [Table-3] shows a good correlation between MIB-1 LI and the 4 parameters used in the WHO grading system, namely, mitotic index, increased cellularity, sheet-like architecture and small cell formation. There was a poor correlation with certain other histological parameters, such as necrosis and cellular pleomorphism. There was also a poor correlation noted with brain invasion. A histologically stratified distribution of brain invasion with MIB-1 LI at the cut-off level of 7% demonstrates that the MIB-1 levels of brain invasive tumors vary according to their histological grade; 6 of 8 brain invasive meningiomas having a MIB-1 LI > 7% were WHO grade II or III while 9 of 10 brain invasive meningiomas with MIB-1 LI < 7% were grade I tumors ([Table-4] and [Figure-3].

DISCUSSION

Meningiomas are benign indolent neoplasms, with a subset of them being aggressive with a higher tendency to recur despite radical surgery. Certain histological characteristics have stronger associations with a decreased recurrence-free survival.[2],[10] The histological diagnosis of atypia remains largely subjective, with significant inter-observer variability that creates difficulties in assessing the prognosis and postoperative management of patients.[2] In meningiomas, the MIB-1 LI has been correlated with clinical tumor recurrence.[11] We therefore studied individual histological criteria of atypia and compared them to the MIB-1 LI. An objective method of counting positively stained nuclei provides a more reproducible method for assessing aggressive behavior as an adjunct to histology. Our results showed lower MIB-I LI in benign and higher labeling indices in atypical meningiomas in accordance with earlier quoted studies.[7],[8],[9],[10],[11],[12] An ROC analysis revealed highest validity for the diagnosis of atypia at a MIB-I LI of 7%. The occurrence of MIB-1 LI of >7% in a few benign WHO grade I meningiomas suggests that in some cases, focal areas of histological atypia may not be included in a biopsy. As was to be anticipated, although few in number, the frankly anaplastic meningiomas had an extremely high MIB-1 LI. High MIB-1 LI (>7%), were found to have statistically significant correlations with an increased mitotic index (>4 / 10 HPF), uninterrupted patternless or sheet-like growth, increased cellularity and small cells with a high nuclear: cytoplasmic ratio, all of which are WHO criteria for the diagnosis of atypia. The correlation of necrosis with MIB-1 LI only neared statistical significance (P=0.0576). Certain studies suggest that geographic necrosis is indicative of infarction while micronecrosis reflects aggressiveness better.[2] The WHO guidelines do not differentiate between these two types of necrosis, probably explaining why it did not correlate as well with a high MIB-1 LI. Higher MIB-1 LI correlate with increased risk of recurrence, but the specific cut-off levels and counting techniques have varied considerably between studies. High proliferative activity may be focal, therefore it is not possible to establish universal values for determining the risk of recurrence.[2],[7],[8],[9],[10] Perry et al[11] demonstrated an association between MIB LI of 4.2% or greater and decreased recurrence-free survival in 425 gross totally resected meningiomas. The authors stated that MIB-1 is a useful ancillary study for routine evaluation of meningiomas, particularly those with borderline atypia. They admitted, however, that a MIB LI of 4.2% may not be applicable to other institutions owing to variability in staining and counting methodologies among different laboratories. The difference in cut-off values could be attributed to the use of an electronic image analyzer by Perry et al [11] as opposed to the manually counted MIB-1 LI in our study. Furthermore, we studied the validity of the MIB-1 LI in predicting histological atypia and not its association with a decreased recurrence-free survival, as was done by Perry et al.[11] In the present study brain invasion did not correlate well with the MIB-1 LI. Although Perry et al have reported brain invasive tumors to behave in a manner similar to atypical tumors they did not differentiate between those that were otherwise histologically bland and those that had histological features of atypia.[11] A limitation of our study was the lack of clinical correlation. However, since meningiomas may recur many years after curative surgery,[9] long follow-up periods of more than 10 years would be required to be very certain that the meningioma in question is truly non-recurrent. This was outside the scope of the present study. Besides, the MIB-1 LI has been showed to have strong associations with recurrence-free survival (RFS) in meningiomas.[2],[7],[8],[9],[10]

There are a few problems inherent in the study of the recurrence of meningiomas. The extent of initial resection, probably the most influential determinant of recurrence, is difficult to document since the surgeon's estimate of the extent of resection cannot account for microscopic residual tumor. Another problem is that the definition of what constitutes recurrence varies between studies; some define recurrence as the performance of a second surgery, others have included radiographic enlargement.[9] Yet another problem is that clinical follow-up periods vary between studies, and the length of the postoperative follow-up is important, especially, as was mentioned earlier, in the case of non-recurrent tumors. While we do find that MIB-1 LI is a useful objective method, at no point can it be considered a supplement for histological investigations. However, as the criteria for histological grading are subject to considerable inter-observer variation,[3] the MIB-1 LI would be a useful adjunct to histological grading. Tumors that do not have any histological features of malignancy but have a high score should be reviewed carefully. Further research is indicated to assess the significance of brain invasion in predicting biological behavior.

CONCLUSION

This study indicates that at our center, the MIB-1 LI has highest validity at a level of 7% in the diagnosis of histological atypia in meningiomas and has a good correlation with individual WHO histological features of atypia. Our results suggest that MIB-1 LI has a role in the detection of atypical meningiomas as focal areas of atypia may not be represented in a biopsy owing to sampling errors. The MIB-1 LI used in conjunction with histological features can help in making a recommendation regarding potentially aggressive behavior in meningiomas.

ACKNOWLEDGMENTS

The authors are grateful to Ms Janet Paul for technical assistance. Supported by the Fluid research grant of the Christian Medical College Hospital, Vellore.

REFERENCES

Copyright 2003 - Neurology India. Also available online at http://www.neurologyindia.com

The following images related to this document are available:

Photo images