摘要
目的探讨基于增强CT图像的影像组学模型在预测肺癌肿瘤增殖抗原Ki-67表达水平中的价值。方法收集2014年1月—2018年12月接受胸部CT增强扫描并在检查后2周内经病理证实、行Ki-67表达水平检测的282例肺癌病人影像与临床相关资料。按7∶3比例将病人分成训练组(n=197)和验证组(n=85)。Ki-67指数≤40%为低表达,>40%为高表达。使用影像组学与人工智能整体解决方案应用平台A.K软件进行影像特征提取。采用logistic回归建立基于影像组学标签与临床因素预测肺癌病人Ki-67低、高表达的列线图模型。结果列线图模型在训练组和验证组的受试者工作特征曲线下面积为0.89和0.87;校准曲线和临床决策曲线均显示模型预测效果良好。结论基于增强CT影像组学标签的列线图模型可用于预测肺癌病人Ki-67的表达水平。
Objective To investigate the value of aradiomicsmodel based on contrast-enhanced computed tomography(CT)images in predicting the expression level of the tumor proliferation antigen Ki-67 in lung cancer.Methods Imaging and clinical data were collected from 282 lung cancer patients who underwent contrast-enhanced chest CT scans from January 2014 to December 2018 and were confirmed by pathology and tested for the expression level of Ki-67 within 2 weeks after examination.The patients were divided into training group with 197 patients and validation group with 85 patients at a ratio of 7∶3.Ki-67 index≤40%was defined as low expression,and>40%was defined as high expression.Based on the integrated solution of radiomics and artificial intelligence,A.K software was used for the extraction of image features.The logistic regression analysis was used toestablish a nomogram model for predicting the low or high expression of Ki-67 in lung cancer patients based on a combination of radiomics labels and clinical factors.Results The nomogram model had an area under the receiver operating characteristic curve of 0.89 in the training group and 0.87 in the validation group,and both the calibration curve and the clinical decision curve showed that the model had a good predictive effect.ConclusionThe nomogram model based on radiomics labels established using contrast-enhanced CT images can be used to predict the expression level of Ki-67 in lung cancer patients.
作者
刘世合
付青
郝大鹏
娄和南
季维娜
张传玉
LIU Shihe;FU Qing;HAO Dapeng;LOU Henan;JI Weina;ZHANG Chuanyu(Department of Radiology,The Affiliated Hospital of Qingdao University,Qingdao 266067,China)
出处
《青岛大学学报（医学版）》
CAS
2024年第2期288-292,共5页
Journal of Qingdao University(Medical Sciences)
基金
青岛市市南区科技计划项目(2020-2-005-YY)。
