【摘要】 目的 ：探讨基于MR成像的列线图模型对预测脑膜瘤WHO分级的应用价值。方法 ：回顾性分析2015年1月—2021年12月湖州市中心医院经手术病理证实为脑膜瘤的患者127例，按照7∶3的比例随机分为训练集与验证集。对常规MR特征进行统计学分析，筛选具有独立预测效能的MR特征；采用Analysis Kit软件进行统计学分析，并计算影像组学分数。对训练集纳入有独立预测效能的MR特征和影像组学分数共同构建列线图模型，使用验证集进行验证，并使用ROC曲线评价模型的准确性。结果：肿瘤形态、囊变坏死、瘤周水肿和影像组学分数共同构建了列线图模型，模型在训练集AUC值为0.965(95%CI:0.927～1.000)，在验证集AUC值为0.975(95%CI:0.931～1.000)。结论 ：基于MR成像的列线图模型具有较好的预测效能，可作为一种更佳的预测脑膜瘤WHO分级的量化工具辅助临床决策。更多还原

【Abstract】 Objective: To explore the value of nomogram model based on MRI in predicting the WHO grading of meningiomas. Methods: A retrospective analysis was performed on 127 patients with meningioma confirmed by surgery and pathology in Huzhou Central Hospital from January 2015 to December 2021. They were randomly divided into training set and validation set according to the ratio of 7 ∶3. The conventional MRI features were statistically analyzed to screen the MRI features with independent prediction efficiency. Statistical analysis was performed using Analysis Kit and radiomics scores were calculated. MRI features with independent predictive efficiency and radiomics score were incorporated to construct a nomogram model in the training set, and the validation set was used to verify it, the ROC curve was used to evaluate the accuracy of the model. Results: The tumor morphology, cystic necrosis, peritumoral edema, and radiomics score jointly constructed a nomogram model. The AUC value of the model in the training set was 0.965(95%CI: 0.927～1.000, and 0.975(95%CI: 0.931～1.000) in the validation set. Conclusion: The nomogram model based on MRI has good predictive efficiency, which can be used as a better quantitative tool for predicting WHO grade of meningiomas to assist clinical decision-making.更多还原