联合双能量CT和传统影像特征鉴别高危和低危肺亚实性结节  

Combination of Dual-Energy CT and Traditional Image Features to Differentiate High-Risk and Low-Risk Pulmonary Subsolid Nodules

作　　者：何长久[1] 青浩渺[1] 李勇[1] 林礼波 解超莲 胡仕北[1] 周鹏[1] 刘杰克 HE Changjiu;QING Haomiao;LI Yong(Department of Radiology,Sichuan Clinical Research Center for Cancer,Sichuan Cancer Hospital&Institute,Sichuan Cancer Center,University of Electronic Science and Technology of China,Chengdu,Sichuan Province 610041,P.R.China)

机构地区：[1]四川省肿瘤医院·研究所,四川省肿瘤临床医学研究中心,四川省癌症防治中心,电子科技大学附属肿瘤医院影像科,成都610041

出　　处：《临床放射学杂志》2025年第5期850-855,共6页Journal of Clinical Radiology

基　　金：国家自然科学基金资助项目(编号:82202141);四川省自然科学基金资助项目(编号:2025ZNSFSC1765)。

摘　　要：目的探讨基于双能量CT(DECT)和传统影像特征的联合模型在鉴别高危和低危肺亚实性结节(SSN)中的价值。方法回顾性搜集行胸部增强DECT的306例SSN,根据病理结果将其分为低危组(原位腺癌、微浸润性腺癌和高分化浸润性非黏液腺癌;n=163)和高危组(中/低分化浸润性非黏液腺癌;n=143)。比较两组SSN的特征[性别、年龄、结节类型、直径、实性成分最大径、实性成分占比(CTR)、体积、CT值、动静脉期电子云密度(Rho_A、Rho_V)、碘含量(IC_A、IC_V)、标准化碘含量(NIC_A、NIC_V)、能谱曲线斜率(λ_(HU)_A、λ_(HU)_V)]。应用多因素Logistic回归构建基于DECT和传统影像特征的联合模型。应用曲线下面积(AUC)评估诊断效能。结果两组的结节类型差异有统计学意义(P<0.001)。高危组的年龄、直径、实性成分最大径、CTR、体积、CT值、Rho_A、Rho_V、IC_V和NIC_V均明显高于低危组,差异有统计学意义(P均<0.05)。高危SSN的独立预测因素包括直径(OR=1.114,P=0.004)、实性成分最大径(OR=1.129,P=0.003)和Rho_V(OR=1.007,P<0.001)。联合模型(特征:直径、实性成分最大径和Rho_V)的AUC(0.826,95%CI:0.779~0.867)显著高于直径(0.730,95%CI:0.676~0.779)、实性成分最大径(0.757,95%CI:0.705~0.804)和Rho_V(0.762,95%CI:0.710~0.808),差异均有统计学意义(P<0.001)。结论直径、实性成分最大径和Rho_V是高危SSN的独立预测因素,基于上述传统影像学特征和DECT特征构建的联合模型可能有助于鉴别高危和低危SSN。Objective To explore the value of a combined model based on dual-energy CT(DECT)and traditional image features in differentiating high-risk and low-risk pulmonary subsolid nodules(SSNs).Methods A total of 306 SSNs that underwent enhanced DECT of the chest were retrospectively collected and divided into a low-risk group(adenocarcinoma in situ,minimally invasive adenocarcinoma,and well-differentiated invasive nonmucinous adenocarcinoma;n=163)and a high-risk group(moderately/poorly differentiated invasive nonmucinous adenocarcinoma;n=143)according to pathological results.The characteristics of SSNs[including gender,age,nodule type,diameter,maximum diameter of solid component,consolidation-to-tumor ratio(CTR),volume,CT value,electron density(Rho_A,Rho_V),iodine concentration(IC_A,IC_V),normalized iodine concentration(NIC_A,NIC_V),slope of the spectral curve(λ_(HU)_A,λ_(HU)_V)in the arterial and venous phases]were compared.A combined model based on DECT and traditional image features was constructed using multivariable Logistic regression.The diagnostic performance was evaluated using the area under the curve(AUC).Results There was a significant difference in nodule type between the two groups(P<0.001).The age,diameter,maximum diameter of solid component,CTR,volume,CT value,Rho_A,Rho_V,IC_V,and NIC_V in the high-risk group were significantly higher than those in the low-risk group(all P<0.05).The diameter(OR:1.114,P=0.004),maximum diameter of solid component(OR:1.129,P=0.003),and Rho_V(OR:1.007,P<0.001)were independent predictors of high-risk SSNs.The AUC of the combined model(0.826,95%CI:0.779-0.867)was significantly higher than that of diameter alone(0.730,95%CI:0.676-0.779),maximum diameter of solid component alone(0.757,95%CI:0.705-0.804),and Rho_V alone(0.762,95%CI:0.710-0.808)(all P<0.001).Conclusion The diameter,maximum diameter of solid component,and Rho_V are independent predictors of high-risk SSNs.The combined model based on these features may help differentiate high-risk and low-risk SSNs.

关 键 词：肺亚实性结节 双能量CT 肺腺癌 危险级别 

分 类 号：R734.2[医药卫生—肿瘤]