Predictors of progression in radiation-induced versus nonradiation-induced pediatric meningiomas: a large single-institution surgical experience.

Abstract

The goal in this study was to outline unique differences between radiation-induced and nonradiation-induced pediatric meningiomas and to identify independent risk factors of tumor recurrence/progression. This is a retrospective cohort study of all pediatric meningiomas diagnosed and surgically treated at the authors' institution between 1993 and 2017. Multivariable Cox regression was applied to identify independent risk factors for tumor recurrence/progression. Thirty-five patients were identified. The primary etiology was nonradiation-induced (n = 24: n = 3 with neurofibromatosis type 2) or radiation-induced (n = 11: acute lymphoblastic leukemia [n = 5], medulloblastoma [n = 4], germ cell tumor [n = 1], and primitive neuroectodermal tumor [n = 1]) meningioma. The mean age at time of diagnosis was 10.7 ± 5.7 years for nonradiation-induced and 17.3 ± 3.5 years for radiation-induced meningiomas. Overall, 8/24 patients with nonradiation-induced meningioma experienced either recurrence or progression of the tumor. Of the 8 patients with tumor recurrence or progression, the pathological diagnosis was clear cell meningioma (n = 3: 2 recurrent and 1 progressive); grade I (n = 2 progressive); grade I with atypical features (n = 2: 1 recurrent and 1 progressive); or atypical meningioma (n = 1 recurrent). None of the patients with radiation-induced meningioma experienced recurrence or progression. Predictors of tumor recurrence/progression by univariate analysis included age at time of diagnosis ≤ 10 years (p = 0.002), histological subtype clear cell meningioma (p = 0.003), and primary etiology nonradiation-induced meningioma (p = 0.04), and there was a notable trend with elevated MIB-1 staining index (SI) (p = 0.09). There was no significant difference between nonradiation-induced and radiation-induced meningiomas (p = 0.258), although there was a trend between recurrent and nonrecurrent meningiomas (p = 0.09). Multivariate Cox regression, adjusted for length of follow-up, identified younger age at diagnosis (p = 0.004) and a higher MIB-1 SI (p = 0.044) as independent risk factors for recurrence. Elevated MIB-1 SI statistically correlated with atypia (p < 0.001). However, there was no significant statistical correlation between tumor recurrence/progression and atypia (p = 0.2). Younger patient age and higher MIB-1 SI are independent risk factors for recurrence. Atypia was not a predictor of recurrence.