MENINGIOMA GRADING - AN ANALYSIS OF HISTOLOGIC PARAMETERS

Histologic grading of meningiomas has prognostic and sometimes therape utic implications, but diagnostic criteria for atypical meningioma are vague, and the significance of brain invasion in the determination of malignancy remains controversial. We reviewed our experience with 581 patients whose meningiomas were resected at Mayo Clinic during the ye ars 1978 through 1988. All patients were followed until death or a med ian of 9.0 years. Ten histologic parameters were assessed and compared with recurrence-free survival. On univariate analysis, six variables were associated with recurrence, although most were statistically sign ificant only in the subset of patients having undergone gross total tu mor resection. On multivariate analyses, the most significant paramete rs were histologic brain invasion (when assessable) and maximal mitoti c rate of at least four per 10 high-power fields (HPF). Also significa nt were combinations of at least three of the following four parameter s: hypercellularity, architectural sheeting, macronucleoli, and small cell formation. Proposed grading criteria based on these findings yiel ded 81% classic, 15% atypical, and 4% brain invasive meningiomas with respective 5-year recurrence rates of 12%, 41%, and 56%. There was no association between histologic grade and either extent of surgical res ection or patient age. However, male sex was associated with high-grad e (atypical/brain invasive) tumors. Too few frankly anaplastic meningi omas were encountered for statistical analysis. Brain invasion and an increased mitotic index (at least four per 10 HPF) are the most powerf ul histologic factors prognostic for recurrence in meningiomas. We pro pose an objective definition for atypical meningioma based on our data . Because the difference in recurrence rates for brain invasive and at ypical meningiomas was not statistically significant, it could not be determined whether brain invasion alone warrants a designation of mali gnancy. Likewise, we were unable to determine what constitutes histolo gic anaplasia due to the rarity of such cases.