非实性肺结节独立危险因素分析及良恶性预测模型建立

Analysis of independent risk factors for non-solid pulmonary nodules and establishment of a benign and malignant predictive model

目的拟筛选影响非实性肺结节的独立危险因素,建立其良恶性预测模型。方法回顾性分析2015—2020年于江苏省镇江市第一人民医院诊断的309例非实性肺结节患者的临床资料,根据结节良恶性将其分为观察组(恶性,61例)和对照组(良性,248例)。采用单因素及二分类logistic回归分析恶性结节的危险因素,建立预测模型。采用受试者操作特征曲线分析该模型的诊断价值。结果观察组年龄、癌胚抗原(CEA)水平、结节直径、实性部分长径、平均CT(m-CT)值及钙化、血管征、空泡征、分叶征、毛刺征占比均高于对照组,差异有统计学意义(P<0.05)。多因素分析结果显示,高CEA水平、实性部分长径较长、高m-CT值及具有血管征、空泡征、分叶征、毛刺征均为非实性肺结节恶性的独立危险因素(OR>1,P<0.05)。预测模型为P=e^(x)/(1+e^(x)),x=-6.078+(2.002×毛刺征)+(3.438×空泡征)+(3.877×血管征)+(0.622×CEA)+(0.149×实性部分长径)+(0.003×|m-CT|)+(5.494×分叶征)。该模型曲线下面积为0.882(P<0.05),截断值为0.842。结论本研究建立的非实性肺结节的良恶性预测模型具有一定的诊断价值,对其良恶性诊断具有较大的临床运用前景。

Objective To screen independent risk factors for non-solid pulmonary nodules and establish a predictive model for benign and malignant pulmonary nodules.Methods The clinical data of 309 patients with non-solid pulmonary nodules diagnosed in the First People’s Hospital of Zhenjiang,Jiangsu Province from 2015 to 2020 were retrospectively analyzed,and they were divided into observation group(malignant,61 cases)and control group(benign,248 cases)according to benign and malignant of the nodules.Single factor and binary logistic regression were used to analyze the risk factors in malignant nodules,and a prediction model was established.The diagnostic value of the model was analyzed by the receiver operation characteristic curve.Results Age,carcinoembryonic antigen(CEA)level,nodule diameter,solid part length diameter,mean CT(m-CT)value,and the proportions of calcification,vascular sign,vacuole sign,lobulation sign,and spicule sign in the observation group were higher than those in the control group,and the differences were statistically significant(P<0.05).Multivariate analysis showed that high CEA level,long solid part length diameter,high m-CT value,and vascular sign,vacuolar sign,lobulation sign,and spicule sign were independent risk factors for malignant non-solid pulmonary nodules(OR>1,P<0.05).The prediction model was P=e^(x)/(1+e^(x)),x=-6.078+(2.002×spicule sign)+(3.438×vacuole sign)+(3.877×vascular sign)+(0.622×CEA)+(0.149×solid part length diameter)+(0.003×|m-CT|)+(5.494×lobulation sign).The area under the curve of the model was 0.882(P<0.05),and the cut-off value was 0.842.Conclusion The benign and malignant prediction model of nonsolid pulmonary nodules is established in this study,which has certain diagnostic value and great clinical application prospect for benign and malignant diagnosis.