摘要
目的探讨基于增强CT图像的影像组学术前预测胃癌淋巴结转移的价值。方法回顾性分析经手术后病理证实的胃癌患者259例,将其随机分为训练集(n=183)和验证集(n=76)。在CT增强静脉期图像上对肿瘤区域进行分割,使用AK软件计算提取396个影像组学特征,在训练集中采用最小冗余最大相关方法(mRMR)及最小绝对收缩和选择算子(LASSO)来选择最有预测价值的影像组学特征,使用多元逻辑回归分析构建影像组学预测模型,最后使用受试者工作特征(ROC)曲线,在训练集和验证集中分别通过曲线下面积(AUC)评估该模型预测胃癌淋巴结转移的效能。结果在提取的396个影像组学特征中,最终筛选出12个对胃癌淋巴结转移最有预测价值的影像组学特征构建影像组学模型,其在训练集和验证集中的AUC分别为0.82和0.78,其准确度、敏感度、特异度分别为0.79、0.74、0.83及0.76、0.63、0.85。结论基于增强CT图像的影像组学对术前预测胃癌淋巴结转移具有潜在的价值。
Objective To explore the value of radiomics based on enhanced-CT images in predicting lymph node metastasis of gastric cancer.Methods A total of 259 patients with gastric cancer confirmed by pathology after operation were retrospectively analyzed,which were divided into two cohorts namely,training(n=183)and validation(n=76)cohort.Tumors were segmented by using AK software on CT venous phase images.Total 396 radiomic features in the segmented images were calculated,then fifiltered and minimized by minimum redundancy maximum relevance(mRMR)and least absolute shrinkage and selection operator(LASSO)regression to select optimal radiomic features based on its correlation with lymph node metastasis of gastric cancer,and then develop radiomics model,which by using multivariable logistic regression analysis.The receiver operating characteristic(ROC)curves were generated and the areas under the curves(AUC)were reckoned to predict lymph node metastasis in both training and validation cohorts.Besides,we performed 100 times LGOCV to verify the reliability of our results.Results In 396 extracted radiomic features,12 were selected to develop radiomics model,its AUC in the training cohort and validation cohort were 0.82 and 0.78 respectively,and its accuracy,sensitivity and specificity were 0.79,0.74,0.83 and 0.76,0.63 and 0.85 respectively.Conclusion The radiomics based on enhanced-CT images has potential to facilitate prediction of preoperative lymph node metastasis in gastric Cancer.
作者
高玉青
王小雷
徐鹤
李淑华
赵灿灿
赵德雷
段绍峰
谢宗玉
GAO Yu-qing;WANG Xiao-lei;XU He;LI Shu-hua;ZHAO Can-can;ZHAO De-lei;DUAN Shao-feng;XIE Zong-yu(Department of Radiology,The First Affiliated Hospital of Bengbu Medical College,bengbu 233000,Anhui Province,China;General Electric(GE)Healthcare,Shanghai 21000,China)
出处
《中国CT和MRI杂志》
2022年第11期140-142,共3页
Chinese Journal of CT and MRI
基金
安徽省教育厅自然科学基金重点项目(KJ2019A0402)
蚌埠医学院自然科学重点项目(2020byzd082)
蚌埠医学院自然科学重点项目(BYKY2019124ZD)。
