摘要
目的:建立基于薄层CT图像的临床影像组学列线图评价其对肺结节良恶性鉴别的临床价值。方法:选自2018年3月至2020年9月间接受薄层CT检查并经病理证实的肺结节患者139例(良性65例,恶性74例)。从每个患者的胸部CT图像中提取影像组学特征。使用最小绝对收缩和lasso回归进行数据降维,选择有效特征并构建影像组学特征模型。结合独立的临床危险因素采用多元Logistic回归建立影像组学列线图。列线图的准确率和诊断效能在训练集中进行评估并在验证集中进行验证,最后通过决策曲线分析评价列线图的临床应用价值。结果:传统影像特征模型在训练集(AUC=0.86,95%CI 0.79~0.93)、验证集(AUC=0.79,95%CI 0.65~0.93)对肺结节良恶性诊断效能较差,混合模型在训练集(AUC=0.94,95%CI 0.90~0.99)、验证集(AUC=0.94,95%CI 0.88~1.00)中表现出更好的鉴别效能和病理符合率,决策曲线表明影像组学的加入有利于患者的预后。结论:基于薄层CT图像的临床影像组学列线图可以方便准确地判断肺结节恶性风险。
Objective:To establish a clinical radiomics nomogram based on thin-section CT images to evaluate its clinical value for identifying benign and malignant pulmonary nodules.Methods:Pulmonary nodules undergoing thin-layer CT examination and pathologically confirmed between March 2018 and September 2020.There were 139 patients(65 benign and 74 malignant).Imaging histological features were extracted from the chest CT images of each patient.Data were downscaled using minimum absolute shrinkage and lasso regression to select useful features and construct an imaging histology feature model.Multiple logistic regression combined independent clinical risk factors with building imaging histology column line graphs.The accuracy and diagnostic efficacy of the column line graphs were evaluated in the training set and validated in the validation set.Finally,the clinical application value of the column line graphs was evaluated by decision curve analysis.Results:The traditional image feature model had poor diagnostic efficacy for benign and malignant pulmonary nodules in the training set(AUC=0.86,95%CI 0.79~0.93),validation set(AUC=0.79,95%CI 0.65~0.93).The hybrid model had poor diagnostic efficacy for benign and malignant pulmonary nodules in the training set(AUC=0.94,95%CI 0.90~0.99),validation set(AUC=0.94,95%CI 0.88~1.00)showed better discriminatory efficacy and pathological compliance,and the decision curves indicated that the inclusion of imaging histology was beneficial for patient prognosis.Conclusion:Clinical imaging radiomics nomogram line drawings based on thin-section CT images can facilitate accurate determination of the risk of malignancy of pulmonary nodules.
作者
宋鑫洋
沈天赐
胡翔宇
王洋洋
杨峰
SONG Xinyang;SHEN Tianci;HU Xiangyu;WANG Yangyang;YANG Feng(Department of Radiology,Xiangyang No.1 People's Hospital,Hubei University of Medicine,Hubei Xiangyang 441100,China.;Department of Respiratory,Xiangyang No.1 People's Hospital,Hubei University of Medicine,Hubei Xiangyang 441100,China.)
出处
《现代肿瘤医学》
CAS
北大核心
2023年第8期1502-1506,共5页
Journal of Modern Oncology
基金
湖北医药学院研究生科技创新项目(编号:YC2022049)。
关键词
X射线计算机
肺结节
影像组学
X-ray computer
pulmonary nodules
imaging histology
