基于CT影像组学预测肺癌椎体转移的研究
The Study of CT-based Radiomics for the Prediction of Vertebral Metastasis of Lung Cancer
李倩 1张堃 2王曼 1袁佳莹2
作者信息
- 1. 湖南省脑科医院放射科
- 2. 湖南中医药大学第一附属医院放射科(湖南长沙 410007)
- 折叠
摘要
目的 探讨基于CT平扫图像的影像组学方法在预测肺癌椎体早期转移中的应用价值.方法 回顾性分析2016年1月至2019年7月间于我院就诊的肺癌椎体转移病例.按照2:1将样本随机分为训练集(n=80)和测试集(n=40).采用3D-SLICER软件进行图像分割与特征提取,使用绝对最小收缩和选择算子(LASSO)回归对影像组学特征进行降维,采用多因素回归分析潜在的独立临床预测因子,分别建立影像组学回归模型及临床-影像组学联合模型,并构建临床-影像组学列线图.结果 影像组学预测模型、临床-影像组学联合模在训练集的AUC值分别为0.786(0.685-0.887)、0.819(0.728-0.911),在测试集的AUC值分别为0.750(0.589-0.911)、0.831(0.669-0.992).结论 基于CT平扫图像的影像组学方法可有效预测肺癌椎体早期转移的风险,构建临床-影像组学列线图可作为辅助诊断肺癌椎体早期转移的可视化工具,为肺癌骨转移的临床决策提供支持.
Abstract
Objective To investigate the value of radiomics based on plain CT images in predicting early vertebral metastasis of lung cancer.Methods Patients with vertebral metastasis of lung cancer in our hospital from January 2016 to July 2019 were retrospectively analyzed.The samples were randomly divided into training set(n=80)and test set(n=40)according to 2:1.3d-SLICER software was used to segment images and extrac features.The absolute minimum shrinkage and selection operator(LASSO)was used to reduce radiomics features.Multiple factors regression analysis was used to analysis potential independent clinical predictors.Build radiomics logistic regression model,clinical-radiomics model,and the clinical-radiomics nomogram.Results The AUC values of the radiomics model and the clinical-radiomics model in the training set were 0.786(0.685-0.887)and 0.819(0.728-0.911),respectively.The AUC values in the test set were 0.750(0.589-0.911)and 0.831(0.669-0.992),respectively.Condlusion The radiomics based on PLAIN CT can effectively predict the risk of early vertebral metastasis of lung cancer,and the clinical-radiomics nomogram can be used as a visual tool to assist in the diagnosis of early vertebral metastasis of lung cancer.
关键词
影像组学/计算机断层扫描/肺癌/脊柱转移瘤Key words
Radiomics/Computed Tomography/Lung Cancer/Metastatic Tumor of Spine引用本文复制引用
出版年
2024
NETL
NSTL
万方数据