120例胶质瘤及瘤周水肿MRI影像组学在评估肿瘤复发中的价值

目的 探讨胶质瘤及瘤周水肿(PTE)MRI影像组学在评估肿瘤复发中的价值。 方法 选取山东大学齐鲁医院2013年1月至2020年12月经术后病理证实的胶质瘤患者120例,包括55例复发和65例无复发患者,根据术前T₂WI和T₁WI增强图像对肿瘤和PTE进行三维容积感兴趣区勾画,并按照8∶2的比例分为训练组和验证组,分析两者及联合的组学特征与肿瘤复发的关系。使用受试者工作特征(ROC)曲线下面积(AUC)与准确性矩阵,比较和评价不同影像组学模型的训练结果。 结果 对于PTE,K临近法(KNN)分类器预测效能最好:训练组AUC值、敏感度、特异度分别为0.910、0.84、0.88,验证组分别为0.916、0.82、0.93。对于肿瘤,逻辑回归(LR)分类器预测效能最好:训练组AUC值、敏感度和特异度分别为0.777、0.69、0.67,验证组分别为0.758、0.82、0.92。当肿瘤+PTE联合时,逻辑回归(LR)分类器预测效能最好:训练组AUC值、敏感度、特异度为0.977、0.88、0.89,验证组则为0.841、0.73、0.83。 结论 胶质瘤PTE和肿瘤影像组学特征在预测胶质瘤术后复发方面具有一定的价值,其中PTE的KNN组学模型效能最佳。

Value of MRI radiomics of glioma and peritumoral edema in evaluating tumor recurrence

Objective To explore the value of MRI radiomics of glioma and peritumoral edema(PTE)in evaluating the postoperative recurrence. Methods A total of 120 patients with glioma confirmed by postoperative pathology during Jan. 2013 and Dec. 2020 were retrospectively selected, including 55 cases with recurrence and 65 cases without recurrence. The tumor and PTE were delineated by three-dimensional volumetric regions of interest based on the preoperative T₂WI and contrast enhanced T1WI images, which were divided into training group and validation group according to the ratio 8∶2. The relationship between tumor recurrence and the radiomic characteristics was analyzed. The receiver operating characteristic(ROC)curve was drawn, and the area under curve(AUC)and accuracy matrix were used to compare and evaluate the results of different radiomic models. Results For PTE, KNN classifier had the best prediction performance(AUC=0.910, sensitivity=0.84, specificity=0.88), while in the validation group, the AUC, sensitivity, and specificity were 0.916, 0.82 and 0.93, respectively. For tumor, LR classifier had the best prediction performance: the AUC, sensitivity and specificity of the training group were 0.777, 0.69 and 0.67, respectively, while in the validation group, they were 0.758, 0.82 and 0.92, respectively. In the model of tumor connected with PTE, LR classifier had the best prediction performance: the AUC, sensitivity and specificity of the training group were 0.977, 0.88 and 0.89, respectively, while in the validation group, they were 0.841, 0.73 and 0.83, respectively. Conclusion The MRI radiomic features of PTE and glioma are valuable to predict postoperative recurrence, and the KNN model of PTE has the best diagnostic efficacy.