1.Department of Radiology, The First Affiliated Hospital of Nanjing Medical University;2.Department of Neurosurgery, The First Affiliated Hospital of Nanjing Medical University
目的：分析常规MRI特征在弥漫性星形细胞瘤IDH基因突变预测中的临床应用价值。 方法：回顾性分析经手术病理证实的85例弥漫性星形细胞瘤患者（58例IDH突变型，27例IDH野生型）的临床及术前常规MRI图像资料。由两名中枢神经影像医师独立依据伦勃朗视觉感受图像（VASARI）特征集的标准进行影像学特征的定量分析。应用Mann-Whitney U检验、卡方检验或Fisher精确检验评价IDH突变型组与野生型组影像学特征的统计学差异。运用Logistic回归分析及受试者工作特征（ROC）曲线分析评价影像学特征对于预测弥漫性星形细胞瘤IDH突变的诊断效能。 结果：两名中枢神经影像医师VASARI特征评分的一致性较好。年龄、肿瘤定位、脑室受累、深部脑白质受累、跨中线、T2/FLAIR不匹配征对于弥漫性星形细胞瘤IDH基因突变的预测具有统计学意义（P＜0.05）。其中年龄、肿瘤定位、T2/FLAIR不匹配征为独立预测危险因素，ROC曲线下面积（AUC）分别为0.779、0.705和0.840；联合三者对预测弥漫性星形细胞瘤IDH基因突变的诊断效能最高，AUC为0.927。 结论：术前常规MRI特征有助于非侵入性的预测弥漫性星形细胞瘤IDH基因突变。
Objective: To analyze the clinical value of conventional MRI features in predicting IDH gene mutation in diffuse astrocytoma. Methods: 85 patients with pathologically confirmed grade Ⅱ diffuse astrocytoma (27 cases with IDH-wild type, 58case with IDH-mutant) who underwent preoperative conventional MRI examination were enrolled in this retrospective study. Two neuroradiologists who were blind to the histopathological results independently conducted the quantitative analysis of MRI features according to the VASARI (visually accessible rembrandt images) scoring system. Mann-Whitney U test, Chi-square test or Fisher's exact test was used to compare the statistical differences in imaging features between IDH-mutant group and IDH-wild group. Logistic regression analysis and receiver operating characteristic (ROC) curve analysis were performed to evaluate the diagnostic efficiency of imaging characteristics in predicting the IDH mutation of diffuse astrocytoma. Results: The two radiologists had a good agreement on the scores of conventional MRI features. Age, tumor location, ependymal invasion, deep white matter invasion, tumor crosses midline and T2/FLAIR mismatch signs are statistically significant in predicting IDH gene mutation in diffuse astrocytoma (P<0.05). Age, Tumor Location and T2/FLAIR mismatch signs were independent predictive risk factors, and the area under curve (AUC) were respectively 0.779, 0.705 and 0.840. The combination of the three parameters has the highest diagnostic efficiency in predicting IDH gene mutation in diffuse astrocytoma, with an AUC of 0.927. Conclusion: Conventional MRI features before surgery are helpful to non-invasive predict IDH gene mutation in diffuse astrocytoma.