Diffusion-weighted MR imaging: diagnosing atypical or malignant meningiomas and detecting tumor dedifferentiation

BACKGROUND AND PURPOSE: Atypical and malignant meningiomas are uncommon tumors with aggressive behavior and higher mortality, morbidity, and recurrence compared with benign tumors. We investigated the utility of diffusion-weighted (DW) MR imaging to differentiate atypical/malignant from benign meningiomas and to detect histologic dedifferentiation to higher tumor grade.MATERIALS AND METHODS: We retrospectively compared conventional and DW MR images (b-value 1000 s/mm²) acquired on a 1.5T clinical scanner between 25 atypical/malignant and 23 benign meningiomas. The optimal cutoff for the absolute apparent diffusion coefficient (ADC) and normalized ADC (NADC) ratio to differentiate between the groups was determined by using receiver operating characteristic (ROC) analysis.RESULTS: Irregular tumor margins, peritumoral edema, and adjacent bone destruction occurred significantly more often in atypical/malignant than in benign meningiomas. The mean ADC of atypical/malignant meningiomas (0.66 ± 0.13 × 10^?3 mm²/s) was significantly lower compared with benign meningiomas (0.88 ± 0.08 × 10^?3 mm²/s; P < .0001). Mean NADC ratio in the atypical/malignant group (0.91 ± 0.18) was also significantly lower than the benign group (1.28 ± 0.11; P < .0001), without overlap between groups. ROC analysis showed that ADC and NADC thresholds of 0.80 × 10^?3 mm²/s and 0.99, respectively, had the best accuracy: at the NADC threshold of 0.99, the sensitivity and specificity were 96% and 100%, respectively. Two patients had isointense benign tumors on initial DW MR imaging, and these became hyperintense with the decrease in ADC and NADC below these thresholds when they progressed to atypical and malignant meningiomas on recurrence.CONCLUSIONS: ADC and NADC ratios in atypical/malignant meningiomas are significantly lower than in benign tumors. Decrease in ADC and NADC on follow-up imaging may suggest dedifferentiation to higher tumor grade.

Meningiomas comprise 14%–20% of all intracranial tumors, with a higher incidence of up to 35.2% among Asians and Africans.¹ Although they are generally benign tumors, up to 10% of meningiomas are atypical or malignant, characterized by nuclear disorganization, necrosis, prominent nucleoli, and increased mitoses on histology. Because of their aggressive behavior, atypical/malignant meningiomas are associated with high morbidity and mortality and may invade the adjacent bone and brain parenchyma. They are also more prone to recur in 29%–41% of patients than typical meningiomas, where the recurrence rate is between 7%–20%.² Although typical extra-axial benign meningiomas are easily diagnosed, distinction from more malignant histologic grades by CT or conventional MR imaging is difficult.³ Neuroimaging features, such as heterogeneous appearance, heterogeneous enhancement, marked perilesional edema, irregular cerebral surface, mushrooming on the outer edge of the lesion, and bone destruction, are not unique or reliable for diagnosing atypical/malignant meningiomas.^4–8 A diagnostic method that can differentiate between benign and atypical/ malignant meningiomas would, therefore, be desirable for surgical planning.Diffusion-weighted (DW) MR imaging has been used to study primary brain tumors, including histologic grading of gliomas and response to treatment.^9–16 Only a few studies have evaluated DW MR imaging for grading meningiomas, and although some have found that apparent diffusion coefficient (ADC) of atypical/malignant meningiomas was significantly lower than benign meningiomas,^2,17 others have not duplicated these results.¹⁶ Furthermore, the accuracy and threshold ADC to distinguish between benign and atypical/malignant meningiomas has not been established. There is also a paucity of literature on the DW MR imaging appearance of malignant transformation of benign meningiomas to higher grade tumors. In this study, we compared DW MR imaging between benign and atypical/malignant groups of meningiomas to estimate the cutoff ADC value for optimal tumor grading and describe imaging features of dedifferentiation of benign meningiomas.