GENE-04. COMPREHENSIVE GENOMIC CHARACTERIZATION OF AGGRESSIVE MENINGIOMAS IDENTIFIES MOLECULAR SIGNATURES THAT PREDICT CLINICAL OUTCOMES

Abstract

BACKGROUND

Current pathological grading schemes do not fully predict meningioma behavior, and the molecular basis for meningioma is incompletely understood. Here, we perform whole exome sequencing (WES), DNA methylation arrays, RNA-seq, NanoString (NS), and immunohistochemistry using tissue microarrays (TMAs) on meningiomas to elucidate molecular drivers of aggressive behavior and identify clinically relevant biomarkers.

METHODS

We constructed a database containing clinical data and tissue from 283 patients who underwent resection of meningioma between 1990 and 2015. Histology was re-evaluated using current grading criteria. We conducted WES (n=24), methylation arrays (n=26), and RNA-seq (n=42) in a discovery cohort composed of 14 grade I, 23 grade II and 5 grade III meningiomas. We further carried out NS profiling of 282 cancer-associated transcripts (n=96), and TMAs (n=241) in an independent validation cohort composed of 132 grade I, 87 grade II and 22 grade III meningiomas. WES data was analyzed via the bcbio pipeline with the ensemble method and CNVkit, methylation data was processed in the minfi Bioconductor package, and RNA-seq data was analyzed with the DESeq2 Bioconductor package. Hierarchical clustering and statistics were performed in R. 

RESULTS

WES demonstrated increased somatic mutation burden was associated with decreased overall survival (OS) (p=0.005). Unsupervised hierarchical clustering of methylation arrays segregated meningioma samples into three groups with significant differences in OS (p=0.0465) and age (p=0.0123). RNA-seq showed enrichment of a FOXM1 mediated transcriptional network with an overrepresentation of cell division genes. High FOXM1 mRNA expression was associated with decreased local recurrence free survival (LRFS) (p=0.004). Consistently, high FOXM1 mRNA expression in NS samples was associated with decreased LRFS (p=0.001) and OS (p=0.006), and high FOXM1 protein expression was associated with decreased LRFS in TMA samples (p=0.049).

CONCLUSION

Comprehensive genomic characterization of aggressive meningiomas identifies molecular signatures associated with poor outcomes, suggesting novel prognostic markers and therapeutic targets.