光谱CT多参数功能成像预测磨玻璃密度肺腺癌侵袭性的价值

Predicting the invasiveness of lung adenocarcinoma manifesting as ground-glass opacity using spectral CT multi-parameter functional imaging

目的探讨光谱CT定量参数预测表现为磨玻璃结节(GGN)肺腺癌侵袭性的价值。方法回顾性分析2022年3至10月在郑州大学第一附属医院经手术切除且病理证实为肺腺癌的129例患者的临床及影像资料,其中男45例,女84例,年龄33~81岁。按照病理结果129例患者分为微浸润腺癌(MIA)组64例和浸润性腺癌(IAC)组65例。所有患者均在术前2周内接受光谱CT增强扫描,在后处理工作站上重建碘密度图、有效原子序数(Zeff)图、电子云密度(ED)图,测量并计算相关光谱参数,包括标准化碘浓度(NIC)、动脉增强分数(AEF)、Zeff及ED;分析病灶的常规CT特征,包括最大径、CT值、结节类型、结节边缘、分叶征、毛刺征、空泡征、胸膜凹陷征、异常血管征及空气支气管征。采用独立样本t检验、Mann-Whitney U检验、χ^(2)检验比较2组临床特征、常规CT特征及光谱CT参数。采用多因素logistic回归分析评估肺腺癌侵袭性的独立危险因素,并构建模型。采用受试者操作特征(ROC)曲线评估各参数及模型预测肺腺癌侵袭性的效能。结果MIA组与IAC组的最大径、CT值、结节类型、边缘、毛刺征、胸膜凹陷征、空气支气管征、静脉期NIC、AEF、静脉期Zeff、动脉期ED、静脉期ED差异均有统计学意义(P<0.05)。多因素logistic回归分析显示,最大径(OR=1.183,95%CI 1.062~1.318)、CT值(OR=1.004,95%CI 1.001~1.007)、静脉期NIC(OR=1.185,95%CI 1.083~1.298)、AEF(OR=0.975,95%CI 0.957~0.994)、静脉期Zeff(OR=0.031,95%CI 0.005~0.196)是肺腺癌侵袭性的独立影响因子。以最大径和CT值建立常规CT模型,以静脉期NIC、AEF、静脉期Zeff建立光谱CT模型,以上参数联合建立联合模型。常规CT模型、光谱CT模型及联合模型预测肺腺癌侵袭性的ROC曲线下面积分别为0.828、0.854和0.902。结论双层探测器光谱CT的定量参数可以作为预测表现为GGN肺腺癌侵袭性的指标,AEF具有最高的诊断效能。光谱CT联合常规CT特征可以进一步提高诊断效能。

Objective To explore the clinical value of quantitative parameters on spectral CT in predicting the invasiveness of lung adenocarcinoma manifesting as ground-glass nodules(GGN).Methods The clinical and imaging data of 129 patients with pathologically confirmed lung adenocarcinoma who were surgically resected in the First Affiliated Hospital of Zhengzhou University from March to October 2022 were retrospectively analyzed,including 45 males and 84 females,aged from 33 to 81.According to the pathological results,they were divided into the minimally invasive adenocarcinoma(MIA)group(n=64)and the invasive adenocarcinoma(IAC)group(n=65).All patients underwent enhanced spectral CT within two weeks before surgery.The iodine density map,Z-Effective(Zeff)map,and electron density(ED)map were reconstructed on the post-processing workstation,and the spectral parameters,including normalized iodine concentration(NIC),arterial enhancement fraction(AEF),Zeff,and ED were measured and calculated.Conventional CT features were analyzed,including maximum diameter,CT value,nodule types,margin,lobulation sign,spiculation sign,bubble sign,pleural retraction sign,abnormal vascular sign,and air bronchial sign.The clinical features,conventional CT characteristics and spectral CT parameters of two groups were compared using the independent sample t test,the Mann-Whitney U test,and theχ^(2) test.Multivariate logistic regression analysis was used to evaluate the independent risk factors of lung adenocarcinoma invasiveness,and the model was constructed.Receiver operating characteristic(ROC)curve analysis was performed to evaluate the performance of parameters and models in predicting the invasiveness of lung adenocarcinoma.Results There were significant differences between the MIA group and IAC group in maximum diameter,CT value,nodule type,margin,spiculation sign,pleural retraction sign,air bronchial sign,venous phase NIC,AEF,venous phase Zeff,arterial phase ED,venous phase ED(P<0.05).Multivariate logistic regression analysis showed that the maximum diameter(OR=1.183,95%CI 1.062-1.318),CT value(OR=1.004,95%CI 1.001-1.007),venous phase NIC(OR=1.185,95%CI 1.083-1.298),AEF(OR=0.975,95%CI 0.957-0.994),venous phase Zeff(OR=0.031,95%CI 0.005-0.196)were independent influence factors for the invasiveness of lung adenocarcinoma.The conventional CT model was established with the maximum diameter and CT value,and the spectral CT model was established with venous phase NIC,AEF,and venous phase Zeff.The combined model was established with all the parameters above.Areas under the ROC curve of the conventional CT model,the spectral CT model,and the combined model for predicting the invasiveness of lung adenocarcinoma were 0.828,0.854,and 0.902,respectively.Conclusion The quantitative parameters of double-layer detector spectral CT can be used as an indicator to predict the invasiveness of lung adenocarcinoma manifesting as GGN,and AEF has the highest diagnostic efficacy.Spectral CT combined with conventional CT features can further improve the diagnostic efficiency.