基于肺磨玻璃结节CT征象的诊断模型列线图评估肺癌浸润性

目的:探讨CT影像征象鉴别磨玻璃结节(GGNs)型肺腺癌浸润程度的价值.方法:搜集2017年6月-2019年10月在本院行手术切除的166例GGNs样肺癌患者的病例资料,其中浸润前组24例、浸润组142例.比较两组间影像学特征的差异,通过二项logistic回归分析筛选浸润性肺癌的危险因素并建立联合诊断模型,利用ROC...

Evaluation on the invasion of lung cancer by diagnostic model nomogram based on the CT characteristics of pulmonary ground glass nodules

Objective:To explore the value of CT imaging characteristics in differentiating the infiltration degree of lung adenocarcinoma manifested as ground glass nodules(GGNs).Methods:From June 2017 to October 2019,the clinical data of 166 patients with pulmonary carcinoma appearing as ground-glass nodules(GGNs)confirmed by pathology were collected in this study.166 GGNs were divided into pre-invasion group(n=24)and invasion group(n=142).The difference of imaging characteristics between the two groups were compared and diagnostic model was then established by univa-riate analysis and multivariate logistic regression analysis.Receiver operating characteristic(ROC)curve were used to evaluate the diagnostic efficacy of CT features.Nomogram was used to quantify the independent risk factors.Results:The proportion of pure GGNs was higher in pre-invasion group.The imaging features of the two groups showed that the diameter of the pre-invasion group was smaller with lower density.The invasion group was more prone to appear as pleural changes,air bronchogram,vascular notch sign,vessel changes,spiculation and lobulation.Multivariate logistic regression analysis showed that vascular changes[odds ratio(OR)=4.319,P=0.01],spiculation(OR=3.664,P=0.016)and vascular notch sign(OR=3.36,P=0.047)were independent risk factors.ROC analysis revealed the AUC of the diagnostic model was 0.805(95%CI:0.713~0.898;P<0.05),with sensitivity of 62.0%and specificity of 87.5%.Nomogram analysis showed that C index of the three imaging features was 0.828(95%CI:0.732~0.924).Conclusion:Invasive GGNs are significantly associated with vascular changes,spiculation and vascular notch sign.The invasion diagnostic model based on CT imaging features is of good diagnostic efficacy.Nomogram is helpful for quantitative diagnosis.