821 Association Between Tumor Mutations and Meningioma Recurrence in Grade I/II Disease

Abstract

INTRODUCTION: A subset of grade I and II meningiomas, in contrast to their predicted favorable oncologic course, recur. Recent interest has been paid to the development of predictive measures to identify which tumors may necessitate more aggressive initial management or more frequent post-resection surveillance. METHODS: Using our tissue bank of 255 meningiomas with linked targeted next-generation sequencing, we identified 173 with at >90 days of post-surgical follow-up. We computed progression free survival cox-regression models for each of the 5% most prevalent tumor mutations corrected for tumor grade, recurrence status, and resection status. Grade III tumors were excluded due to issues with the proportional hazards assumption. We then built a multi-gene model by including all individual gene predictors with a p-value of less than 0.20. Starting with that model, we performed automated backwards selection. RESULTS: In the single gene analysis, ATM (HR = 5.22; 95%CI: 1.81 – 15.08), CREBBP (HR = 2.71; 95% CI = 1.15 – 6.38), and POLE (HR = 0.56; (95%CI = 0.32 – 0.97) were significantly associated with alterations in disease progression. In the multi-gene model, only POLE remained a significant predictor of recurrence after adjusting for the same clinical covariates and the genes ATM, ATR, CREBBP, RNF43, and SETD2. Backwards selection identified recurrence status, resection status, and mutations in ATM (HR = 8.14; 95%CI = 2.65 – 24.99) and POLE (HR = 0.45; 95% CI = 0.25 – 0.78) as predictive of recurrence. CONCLUSIONS: There is an association between ATM, CREBBP and POLE with progression free survival after accounting for tumor grade, resection extent, and recurrence status. In a muiti-gene model, POLE was the only gene found to alter survival. Backwards variable selection identified ATM as deleterious and POLE as protective.