MRI影像组学列线图术前预测子宫内膜癌淋巴结转移  

MRI Radiomics Nomogram for Preoperative Prediction of Lymph Node Metastasis in Endometrial Cancer

作　　者：方如旗[1] 翁淑萍[1] 陈霞平[1] 刘凯丽 杨敏芳 郑贤应 FANG Ruqi;WENG Shuping;CHEN Xiaping;LIU Kaili;YANG Minfang;ZHENG Xianying(Department of Radiology,Fujian Maternity and Child Health Hospital,Fuzhou 350001,China;不详)

机构地区：[1]福建省妇幼保健院影像科,福建福州350001 [2]福建省妇产医院影像科,福建福州350011

出　　处：《中国医学影像学杂志》2023年第11期1172-1177,共6页Chinese Journal of Medical Imaging

基　　金：福建省卫生健康科研人才培养项目(2019-ZQN-26);福建省科技创新联合资金项目(2020Y9146);福建省自然科学基金面上项目(2023J011215)。

摘　　要：目的构建并验证MRI影像组学列线图术前预测子宫内膜癌淋巴结转移的价值。资料与方法回顾性分析2014年9月—2022年5月福建省妇幼保健院195例子宫内膜癌患者的临床及影像资料,利用MaZda软件获取影像组学特征,建立影像组学得分,利用多因素Logistic回归建立淋巴结转移预测列线图,采用受试者工作特征曲线评价模型的预测效能,并在测试集中进行验证。结果训练集134例,其中淋巴结转移43例;验证集61例,其中淋巴结转移28例,术前MRI诊断淋巴结转移的敏感度为32.4%(23/71),特异度为96.0%(119/124)。多因素Logistic回归分析显示影像组学得分、MRI提示宫颈间质浸润、Ki-67、MRI提示淋巴结转移为独立危险因素,依此构建的预测模型在训练组及验证组的曲线下面积分别为0.897、0.840,Hosmer-Lemeshow检验显示模型在训练组、验证组拟合优度良好(P=0.689、0.167)。结论MRI影像组学列线图提高了子宫内膜癌淋巴结转移的诊断效能,有助于临床制订个体化治疗策略。Purpose To develop and validate a nomogram based on MRI radiomics for preoperative prediction of lymph node(LN)metastasis in endometrial cancer.Materials and Methods A total of 195 consecutive endometrial cancer patients in Fujian Maternity and Child Health Hospital from September 2014 to May 2022 were retrospectively enrolled,the clinical and MRI data were analyzed,the radiomic features were extracted by MaZda software to obtain radiomics signature.Multivariable Logistic regression analysis was performed to develop the predicting model of LN metastasis and obtain a nomogram.The performance of the nomogram was assessed by receiver operating characteristic curve analysis in training group,and external validation was also assessed in testing group.Results A total of 134 patients were enrolled in training group,in which 43 cases were LN positive,and 61 patients were included in testing group,in which 28 cases were LN positive.LN metastasis of sensitivity and specificity was 32.4%(23/71),96.0%(119/124)via conventional MRI,respectively.Multivariable Logistic regression analysis showed that the radiomics signature,cervical stromal invasion reported by MRI,Ki-67 expression and LN metastasis reported by MRI were the independent predictors for LN status.The area under the curve of training and testing group were 0.897 and 0.840,respectively,via nomogram based on Logistic regression.The model showed good calibration in both training and testing group via Hosmer-Lemeshow test(P=0.689,0.167).Conclusion The MRI radiomics nomogram may improve preoperative identification of LN status and help in clinical decision-making in endometrial cancer.

关 键 词：子宫内膜肿瘤 磁共振成像 扩散加权成像 淋巴结转移 影像组学 列线图 

分 类 号：R445.2[医药卫生—影像医学与核医学] R737.33[医药卫生—诊断学]