基于高分辨CT测量参数对肺孤立性GGO结节病理类型的预测价值分析  

High Resolution CT Measurements of Solitary Pulmonary Nodular Ground Glass Opacity Predicts its Pathological Subtype

作　　者：屈艳娟[1] 徐丽莹[1] 曹毅媛[1] 鲁植艳[1] 廖美焱[1] 田志雄[1] QU Yanjuan;XU Liying;CAO Yiyuan(Department of Radiology,Zhongnan Hospital of Wuhan University,Wuhan,Hubei Province 430071,P. R. China)

机构地区：[1]武汉大学中南医院医学影像科

出　　处：《临床放射学杂志》2018年第11期1839-1843,共5页Journal of Clinical Radiology

摘　　要：目的探讨基于术前肺孤立磨玻璃样结节(nGGO)高分辨CT(HRCT)测量参数计算Ki67增值指数(PI)预测值对其病理类型的预测价值。方法搜集223例病理确诊为腺癌的肺孤立性nGGO的HRCT资料。测量nGGO直径、体积(Vol)、最大(Max)CT值、平均(Avg)CT值、CT值标准差(Std)以及术后Ki67PI,通过多元线性回归构建CT测量参数对Ki67PI的预测模型,计算Ki67PI的预测值;然后通过受试者工作特征曲线(ROC)分析比较Ki67PI预测值对nGGO病理类型浸润前腺癌(PIA)、微浸润腺癌(MIA)及侵袭性腺癌(IAC)的鉴别价值。结果基于多元线性回归得到方程:Ki67PI预测值=0.014×Avg+0.024×Std+10.030,计算所得Ki67PI预测值在PIA、MIA及IAC间的差异显著(P<0.001)。ROC曲线分析显示,Ki67PI预测值的曲线下面积(AUC)对鉴别PIA和MIA(0.855∶0.770,P=0.066)或是MIA和IAC(0.741∶0.825,P=0.117)均与Ki67PI实测值相当。结论基于nGGO术前HRCT测量参数计算所得Ki67PI预测值能准确鉴别病灶病理类型,诊断效能与实测Ki67PI相当。Objective To investigate the prediction value of Ki67 proliferation index ( PI) Prediction calculated frompreoperative high resolution CT ( HRCT) measurements of solitary pulmonary nodular ground glass opacity ( nGGO) on itspost-operative pathological subtype. Methods HRCT of 223 patients with solitary lung nGGOs which were confirmed ad-enocarcinoma postoperatively were collected. Diatmeters,Volume ( Vol) ,the maximum ( Max) ,average ( Avg) and stand-ard deviation ( Std) of CT attenuation,as well as the postoperative Ki67 PI of nGGOs were measured. Multivariate linear re-gression was performed to establish an equation for calculating Ki67 PI Prediction from CT measurements of nGGO. There-after,the receiver operating characteristic curve ( ROC) analysis were performed to compare Ki67 PI Prediction with the ac-tual Ki67 PI in differentiating pathological subtypes of nGGO,namely,preinvasive adenocarcinoma ( PIA) ,minimally inva-sive adenocarcinoma ( MIA) and invasive adenocarcinoma ( IAC) . Results The multivariate linear regression equationwas established as: Ki67-PI Prediction =0. 014 × Avg + 0. 024 × Std + 10. 030,which were significantly different betweenPIA and MIA ( P <0. 001) ,as well as between MIA and IAC ( P <0. 001) . The ROC analysis showed that the area undercurve ( AUC) of the Ki67-PI Prediction was close to the actual Ki67-PI either in differentiating MIA from PIA ( 0. 855∶0. 770,P =0. 066) ,or in distinguishing IAC from MIA ( 0. 741∶ 0. 825,P =0. 117) . Conclusion The Ki67 PI Predictioncalculated from preoperative HRCT measurements of nGGO allows for accurate differentiation of pathological subtypes of nG-GOs,with the diagnostic performance being close to the actual one.

关 键 词：磨玻璃样影 腺癌 肺 计算机体层摄影 

分 类 号：R734.2[医药卫生—肿瘤] R730.44[医药卫生—临床医学] R563