Apparent Diffusion Coefficient (ADC) and First-Order Histogram Statistics in Differentiating Malignant Versus Benign Meningioma in Adults

  Apparent diffusion coefficient (ADC) measured by diffusion-weighted MRI and first-order histogram (FOH) extracted features, as markers of tumor heterogeneity, have been implicated in differentiating grade of the intracranial tumors. To examine whether ADC, normalized ADC (NADC), and FOH features such as entropy, kurtosis, and uniformity can differentiate benign vs. malignant meningioma. Patients and MRI with diffusion-weighted (DW) imaging sequence of 62 patients with histologically-proven meningioma (37 benign and 25 malignant/atypical) were included. After co-registration of ADC maps to their corresponding anatomical MRI (post-contrast T1-weighted [T1C] images) and delineation of the tumors border by selecting regions of interest (ROIs) on T1C images, a mask of tumor was created and overlaid on the corresponding ADC map. Then, FOH features were extracted. Mean (± standard deviation [SD]) ADC values in benign and malignant subgroups were respectively 1.05 (± 0.23) and 0.99 (± 0.29) 10-3 mm2/s, and P = 0.69. NADC ratios were not statistically significant between benign (0.5 ± 0.09) and malignant (0.5 ± 0.07) meningioma groups (P = 0.89). Mean values of entropy (6.36 vs. 6.44), kurtosis (5.77 vs. 5.45), and uniformity (536.8 vs. 304.18) were comparable between benign and malignant meningioma subgroups. Receiver operating characteristic (ROC) curve analysis did not yield the significant area under the curve results to show acceptable diagnostic accuracy for any of the measured variables. ADC, NADC, and statistical features of tumor heterogeneity by FOH method measured by DW-MRI were not able to differentiate benign vs. malignant/atypical meningiomas.