Conventional and Advanced Magnetic Resonance Imaging Features of Supratentorial Extraventricular Ependymomas.

OBJECTIVE: To describe the magnetic resonance imaging characteristics of supratentorial ependymomas. METHODS: The magnetic resonance imaging and computed tomography imaging characteristics of 49 cases of supratentorial ependymomas were analyzed retrospectively. The location, size, degree of perilesional edema, gross appearance, computed tomography attenuation characteristics, T1 and T2 signal intensity characteristics, degree of diffusion restriction, presence of calcification, and hemorrhage were documented for each lesion. The intratumoral susceptibility scores (ITSS), apparent diffusion coefficient (ADC) values, relative cerebral blood volume, and choline/N-acetyl aspartate ratios were documented where available. RESULTS: The frontal lobe was the most common location with a mean size of 6.37 × 4.8 cm. Severe perilesional edema was evident in 30%. Heterogenous, solid-cystic appearance was present in 96% lesions, with 95% of extraventricular lesions extending from pial surface to the ventricular margin. Calcification was seen in 55% of cases. The ITSS was 3 in 85.7% of lesions. The mean ADC value calculated was 600 × 10 mm/s. The mean relative cerebral blood volume on dynamic susceptibility contrast perfusion was 4.83. The mean choline/N-acetyl aspartate ratio was 5.87. Leptomeningeal dissemination was demonstrable in 5 lesions. Four lesions were abutting the dura with frank dural invasion in one. One patient presented with disseminated disease without evidence of a primary lesion. CONCLUSIONS: A large lesion with relatively well-defined margins, heterogeneous solid cystic appearance, extending from the pial surface to the ventricular margin, presence of calcification, and ADC values approaching those of white matter should raise a suspicion of supratentorial ependymoma. High ITSS, MR perfusion parameters, and magnetic resonance spectroscopy characteristics are similar to those of other high-grade gliomas.