ABSTRACT
BACKGROUND: Many meningiomas are identified by imaging and followed, with an assumption that they are WHO Grade I tumors. The purpose of our investigation is to find clinical or imaging predictors of WHO Grade II/III tumors to distinguish them from Grade I meningiomas. METHODS: Patients with a pathologic diagnosis of meningioma from 2002-2009 were included if they had pre-operative MRI studies and pathology for review. A Neuro-Pathologist reviewed and classified all tumors by WHO 2007. All Brain MRI imaging was reviewed by a Neuro-radiologist. Pathology and Radiology reviews were blinded from each other and clinical course. Recursive partitioning was used to create predictive models for identifying meningioma grades. RESULTS: Factors significantly correlating with a diagnosis of WHO Grade II-III tumors in univariate analysis: prior CVA (p = 0.005), CABG (p = 0.010), paresis (p = 0.008), vascularity index = 4/4: (p = 0.009), convexity vs other (p = 0.014), metabolic syndrome (p = 0.025), non-skull base (p = 0.041) and non-postmenopausal female (p = 0.045). Recursive partitioning analysis identified four categories: 1. prior CVA, 2. vascular index (vi) = 4 (no CVA), 3. premenopausal or male, vi < 4, no CVA. 4. Postmenopausal, vi < 4, no CVA with corresponding rates of 73, 54, 35 and 10% of being Grade II-III meningiomas. CONCLUSIONS: Meningioma patients with prior CVA and those grade 4/4 vascularity are the most likely to have WHO Grade II-III tumors while post-menopausal women without these features are the most likely to have Grade I meningiomas. Further study of the associations of clinical and imaging factors with grade and clinical behavior are needed to better predict behavior of these tumors without biopsy.
