Radiographic Prediction of Meningioma Grade and Genomic Profile

Background: The clinical management of meningioma is guided by tumor grade and biological behavior. In particular, the profile of chromosomal aberrations in meningiomas has been shown to associate with recurrence risk. However, both tumor grade and genomic profiling can only be determined following surgical resection and histopathologic analysis at present. Pre-operative determination of these variables may influence clinical decision-making. We investigated the association between imaging features extracted from pre-operative gadolinium-enhanced T1-weighted MRI and meningioma grade and genomic profile.

Methods: 175 meningioma patients (103 low-grade and 72 high-grade) with pre-operative contrast-enhanced T1 MRI were assessed for 15 radiomic (quantitative) and 10 semantic (qualitative) imaging features, as well as their pathologic grade, subtype, and genomic profile. The area under the receiver operator characteristics curve (AUC) and odds ratios (OR) were used to assess radiomic and semantic features. Random forest classifiers were developed on a training dataset and validated on an independent dataset to assess predictive power of imaging features for biologic variables. Overall burden of chromosomal alterations as well as specific copy number changes associated with grade I versus grade II-III meningiomas were analyzed.

Results: Twelve radiographic features (eight radiomic and four semantic) were significantly associated with meningioma grade, each with high predictive power. High-grade tumors exhibited necrosis or hemorrhage, intratumoral heterogeneity, and non-spherical shape compared with low-grade tumors. Combining radiomic and semantic features further increased their classification power (AUC = 0.86). Among grade I meningiomas, non-spherical (irregular) shape was significantly associated with chromosome 22 loss (p = 0.006). Among high grade meningiomas, tumors with venous sinus invasion demonstrated higher chromosomal alterations.

Conclusion: We found a radiographic signature for meningioma grade and genomic profile using standard pre-operative contrast-enhanced MRI. This strong link between the radiographic phenotype of a tumor and its pathology may influence the decision to observe a tumor or to pursue surgery and earlier consideration of adjuvant therapies. Our study highlights the potential clinical impact of integrative imaging analysis in guiding meningioma management.