PL02.3.A MOLECULAR CLASSIFICATION TO REFINE SURGICAL AND RADIOTHERAPEUTIC DECISION MAKING IN MENINGIOMA

Abstract BACKGROUND Treatment of the tumor and dural margin, most commonly with surgery and sometimes radiation, are cornerstones of therapy for meningioma. Molecular classifications have provided insights into the biology of disease. However, response to treatment remains heterogeneous. We compiled and generated clinical data on 2824 total meningiomas with molecular data on 1686 retrospective tumors and 100 prospective tumors from the RTOG-0539 phase-II clinical trial (NCT00895622) to define biomarkers of treatment response across molecular classifications. MATERIAL AND METHODS Data were collected from 10 different institutions. Meningiomas were re-graded in accordance with 2021 WHO classification based on the presence of CDKN2A/B deletion and TERT promoter mutation and stratified to Molecular Groups based on previously published methodologies utilizing DNA methylation alone or combined with RNA sequencing. Only molecularly-defined meningiomas based on the DKFZ central nervous system classifier were included in molecular classification and analysis. Propensity score matching (PSM) was used to mimic a randomized control trial for comparison of treatment arms (GTR vs STR, Simpson Grade 1/2 vs 3, Simpson Grade 1 vs 2, RT vs Observation). RESULTS Gross tumor resection (GTR) was associated with significantly longer progression-free-survival (PFS) across all Molecular Groups, and longer overall survival (OS) in Proliferative meningiomas (STR vs GTR HR 1.90, 95%CI 1.28-2.82, p=0.00155). Following PSM of key covariates including WHO grade, tumor location, and Molecular Group, the addition of dural margin treatment (Simpson grade 1 or 2) prolonged PFS compared to GTR alone (Simpson grade 3; HR 1.64, 95%CI 1.03-2.62, p=0.038). These results were reproduced when other molecular classifications and prognostic systems were utilized in PSM as well. When considering the effect of adjuvant radiotherapy (RT) following PSM, Immunogenic, NF2-wt, and Hypermetabolic meningiomas obtained a PFS benefit from the addition of RT but Proliferative meningiomas were RT-resistant. These findings were validated in meningiomas from the RTOG-0539 trial. To obtain an individualized prediction of RT-responsiveness, we leveraged RT-treated meningiomas from the RTOG-0539 trial (intermediate- and high-risk treatment arms in the trial) to build a DNA methylation and gene-expression model that was able to predict response to RT better than standard of care clinical covariates including WHO grad