Susceptibility-Weighted MRI for Predicting NF-2 Mutations and S100 Protein Expression in Meningiomas

Area of Science:

• Neuroimaging
• Oncology
• Radiology

Background:

• Meningiomas exhibit varied disease courses influenced by S100 protein expression and neurofibromatosis type 2 (NF-2) mutations.
• Non-invasive biomarkers are crucial for predicting meningioma behavior and guiding treatment strategies.

Purpose of the Study:

• To investigate the utility of susceptibility-weighted MRI (SWI) features for non-invasively identifying NF-2 copy number loss and S100 protein expression in meningiomas.
• To correlate SWI findings with tumor characteristics and genetic alterations.

Main Methods:

• Retrospective analysis of 99 patients with S100 data and 92 with NF-2 data, using preoperative 3T cranial MRI.
• Extraction of morphological, radiomic (Pyradiomics), and deep learning (CNN) features from SWI.
• Statistical analysis including Mann-Whitney U tests and logistic regression, with CNN features used for classification.

Main Results:

• NF-2 copy number loss correlated with higher-grade tumors, "en plaque" growth, and calcification.
• S100 protein expression was associated with elevated SWI signal intensity and decreased intratumoral entropy.
• Deep learning models achieved AUCs of 0.85 for S100 expression and 0.74 for NF-2 loss.

Conclusions:

• SWI is a promising imaging technique for non-invasively detecting NF-2 copy number loss and S100 protein expression in meningiomas.
• SWI features reveal underlying neovascularization and microcalcification, aiding in the characterization of meningioma subtypes.