Diagnostic value of six MRI features for central neurocytoma

作者：Li, Xiaodan; Guo, Liuji; Sheng, Sen; Xu, Yikai; Ma, Lichao; Xiao, Xiang; Si, Zhiguang; Chen, Yanping; Wu, Yuankui^*

来源：European Radiology, 2018, 28(10): 4306-4313.

DOI：10.1007/s00330-018-5442-y

• Cerebral ventricle neoplasms
• Neurocytoma
• Magnetic resonance imaging
• Diagnostic imaging
• ROC curve

ObjectivesTo increase our understanding of the imaging features of central neurocytoma (CN) and improve the preoperative MRI diagnosis accuracy.MethodsPreoperative MR images of 30 CNs and another 68 intraventricular non-CN tumours were analysed by one experienced neuroradiologist retrospectively to identify previously reported features and new features of CN. Six blinded radiologists independently reviewed all these MRI images, and scored all characteristic features on a five-point scale. Diagnostic value was assessed by the area under the receiver operating characteristic curve (AUC); sensitivity, specificity and accuracy were also calculated.ResultsIn addition to the scalloping' sign, broad-based attachment' sign and soap-bubble' sign, three new MRI features of CN were identified, including the peripheral cysts' sign, fluid-fluid level' sign and the gemstone' sign. The scalloping sign showed the highest AUC value (0.82), followed by the peripheral cysts sign (0.75) and broad-based attachment sign (0.75). The scalloping sign exhibited the highest specificity (82%), followed by the fluid-fluid level sign (79%) and gemstone (78%) sign. The broad-based attachment sign (85%) was the most sensitive feature, followed by the soap-bubble sign (84%) and peripheral cysts sign (77%).ConclusionThere are six characteristic MRI features that help to improve the preoperative diagnostic accuracy of CN.Key Points center dot This study is the largest magnetic resonance imaging (MRI) cohort on central neurocytoma (CN).center dot Three new features helpful for the diagnosis of CN were reported.center dot Diagnostic value of six MRI features of CN was preliminarily determined.

• 出版日期2018-10
• 单位南方医科大学