Diffusion tensor magnetic resonance imaging for predicting the consistency of intracranial meningiomas.

BACKGROUND: The ability of preoperative MRI-sequences to predict the consistency of intracranial meningiomas has not yet been clearly defined. We aim to demonstrate that diffusion tensor imaging (DTI) improves the prediction of intracranial meningiomas consistency.
METHODS: We prospectively studied 110 meningioma patients operated on in a single center from March 1st to the 25th of May 2012. Demographic data, location and size of the tumor, peritumoral edema, T1WI, T2WI, proton density weighted (PDWI), fluid-attenuated inversion recover (FLAIR) sequences, and arterial spin labeling (ASL) perfusion were studied and compared with the gray matter signal to predict meningioma consistency. Diffusion tensor imaging (DTI) with fractional anisotropy (FA) and mean diffusivity (MD) maps were included in the preoperative MRI. Meningioma consistency was evaluated by the operating surgeon who was unaware of the neuroradiological findings.
RESULTS: In univariate analysis, meningioma size (diameter > 2 cm) and supratentorial or sphenoidal wing location were more frequently associated with hard-consistency meningiomas (p < 0.05). In addition, isointense signal on MD maps (p = 0.009), hyperintense signal on FA maps, and FA value > 0.3 (p = 0.00001) were associated with hard-consistency tumors. Age and sex, T1WI, T2WI, PDWI, FLAIR, or ASL perfusion sequences and peritumoral edema were not significantly associated with meningioma consistency. In logistic regression analysis, the most accurate model (AUC: 0.9459) for predicting a hard-consistency meningioma shows that an isointense signal in MD-maps, a hyperintense signal in FA-maps, and an FA value of more than 0.3 have a significant predictive value.
CONCLUSIONS: FA value and MD and FA maps are useful for prediction of meningioma consistency and, therefore, may be considered in the preoperative routine MRI examination of all patients with intracranial meningiomas.