肺结节CT可视影像学特征评分预测病理性质的诊断价值  

The value of CT visual score in predicting pathology of pulmonary nodules

作　　者：甘雄辉 潘永军 罗艳梅 韦乐心 韩秋丽[1] GAN Xiong-hui;PAN Yong-jun;LUO Yan-mei(Department of Radiology,Liu Zhou People Hospital,Guangxi 545006,China)

机构地区：[1]广西柳州市人民医院放射科,广西545006

出　　处：《放射学实践》2023年第6期709-714,共6页Radiologic Practice

基　　金：广西壮族自治区卫生健康委员会计划项目(Z-B20221376)。

摘　　要：目的:探讨肺结节可视影像学特征评分预测病理性质的诊断价值。方法:回顾性分析2020年1月-2021年12月经手术病理证实的肺实性结节、纯磨玻璃结节、混合磨玻璃结节共185例的CT资料。按照病理性质分为良性病变、浸润前病变、微浸润病变、侵润病变。良性病变为炎性结节,浸润前组包括AAH及AIS,微侵润组为MIA,浸润组为IAC。4组临床资料计量数值符合正态分布的采用单因素方差分析,不符合正态分布的采用Kruskal-wallis H检验。筛选有统计意义的变量进行有序logistic回归分析,再筛选公认的恶性征象分叶征、毛刺征、胸膜凹陷征、空泡征、大小作为构建logistic回归模型,利用ROC曲线下的面积(AUC)比较logistic回归模型、大小及评分分值诊断浸润病变效能。结果:单因素方差分析中良性组、浸润前组、微侵润组、浸润组年龄差异无统计学意义(P>0.05),良性组评分分值2.1±0.87,浸润前组评分分值2.6±1.77,微侵润组2.6±1.34,浸润组评分分值4.9±1.72,差异具有统计学意义(P<0.05);Kruskal-wallis H检验性别、大小、分叶征、毛刺征、胸膜凹陷征、空泡征、血管支气管扭曲征差异有统计学意义(P<0.05),而血管集束征差异无统计学意义(P>0.05);有序logistic回归分析显示分叶、毛刺、大小为浸润病变的独立危险因素,ROC下曲线面积logistic回归模型AUC=0.880,敏感度0.823,特异度0.819;评分分值AUC=0.882,评分分值临界值是3.25,敏感度0.832,特异度0.791;大小AUC=0.770;大小临界值11.7 mm,敏感度0.743,特异度0.708。结论:肺结节可视影像学特征评分分值和logistic回归模型诊断效能相当,可作为预测肺结节浸润病变的一种补充手段。Objective:To explore the value of visual score in predicting pathology of pulmonary nodules.Methods:185 cases of CT data of pulmonary nodules from January 2020 to December 2021 were analyzed retrospectively,which were confirmed by surgery and pathology.According to the pathology,these cases was divided into benign lesions,pre-invasive lesions,minimally invasive lesions and invasive lesions.The benign lesions were mainly inflammatory nodules,Pre-invasive lesions included AAH and AIS.The data of four groups consistent with the normal distribution were analyzed with one-way ANOVA,otherwise,Kruskal-Wallis H test was used.lobulation sign、spiculation、pleural indentation sign、vacuole sign、the size were screened by ordered Logistic regression analysis and followed by establishing the logistic regression model.Receiver operating characteristic curve(ROC)was performed to evaluate the diagnostic performance of the logistic regression model,the size and scores.Results:There was no significant difference in age among the four groups(P>0.05).The CT visual scores of benign group(2.1±0.87),pre-invasive group(2.6±1.77),minimally invasive group(2.6±1.34)and invasive group(4.9±1.72)were significantly different(P<0.05).There were significant difference in lobular sign,spiculation,pleural indentation sign,vacuole sign,vascular-bronchial distortion sign and gender among the four groups(P<0.05),except vascular bundle sign.Logistic regression analysis showed that the lobulation sign,spiculation and the size were independent risk factors for predicting the pathological invasiveness.AUC of logistic regression model was 0.880,sensitivity 0.823,specificity 0.819;AUC of the CT visual score was 0.882.The cut off value of the score was 3.25,the sensitivity was 0.832 and the specificity was 0.791;The AUC of the size was 0.770,the cut value of the size was 11.7mm,the sensitivity was 0.743 and the specificity was 0.708.Conclusion:The efficiency of CT visual score of pulmonary nodules was a good prediction method,which can be used as a

关 键 词：肺结节 肿瘤分级 体层摄影术 X线计算机 诊断 

分 类 号：R521.6[医药卫生—内科学] R730.4[医药卫生—临床医学] R814.4R44