孤立性肺结节恶性概率估算临床预测模型的建立

目的：筛选恶性孤立性肺结节（solitary pulmonary nodules，SPN）的危险因素，构建判断SPN良恶性的临床预测模型。方法回顾性分析孤立性肺结节患者270例的临床资料及胸部影像学特征。结果270例患者中，肺癌110例（40．7％），良性病变160例（59．3％）。在SPN的影像学分析中，分叶、毛刺、胸膜牵拉、增强后强化、支气管充气征与结节恶性病变显著相关，P＜0．05，而边缘清晰、钙化、密度均匀均与良性病变显著相关，P＜0．05。在单因素分析中，恶性孤立性肺结节与患者年龄、吸烟史、影像学特征、结节最大径等因素显著相关，P＜0．05。在多因素分析中，患者年龄、恶性影像学特征及结节最大径是恶性SPN的独立危险因素（P＜0．01）。建立的SPN恶性概率估算的临床预测方程为：SPN恶性预测值P＝ex／（1＋ex），X＝－5．882＋0．050×年龄＋1．672×影像学良恶性＋0．123×结节最大径，其中e为自然对数。选取截断点为0．46，敏感性达82％，特异性达85％，阳性预测值80％，阴性预测值87％。临床预测模型的受试者工作特征曲线下面积为0．901。结论患者年龄、恶性影像学特征及结节最大径是判断SPN良恶性的独立影响因素，建立的数学预测模型的准确性较高，有较高的临床应用价值。

Establishment of Clinical Prediction Model to Estimate the Probability of Malignancy in Patients with Solitary Pulmonary Nodules

Objective To screen the clinical risk factors of lung cancer in the patients with solitary pulmonary nodules ( SPN) ,and build the clinical prediction model to estimate the probability of malignancy.Methods A retrospective analysis was performed on the clinical data and chest imaging characteristics of 270 patients with SPN.Results Among 270 patients,there had 110 (40.7%) cases of lung cancer,and 160 (59.3%) benign lesions.On the analysis of imaging characteristics,lobulation, spiculated sign, pleural indentation sign, contrast enhancement, air bronchogram sign were associated with lung cancer ( P <0.05).Nodules with clear boundary,calcification,homogeneous density were associated with benign lesions (P<0.05).Single factor analysis showed that age, smoking history, malignant imaging characteristics and diameter were significantly affected the judgment of SPN whether it was benign or malignant(P<0.05).The multivariate analysis revealed that age,malignant imaging characteristics and diameter were independent risk factors of lung cancer in the patients with SPN (P<0.01).The clinical pre-diction model to estimate the probability of malignancy as following:P=ex/(1+ex),X=-5.882+0.050* age+1.672*ima-ging characteristic+0.123* the maximum diameter,where the e is the base of the natural logarithm.The cut-off value was 0.46. The sensitivity was 82%,specificity 85%,positive balue 80%,and negative predictive value 87%.The area under the ROC curve for our model was 0.901.Conclusion Age,malignant imaging characteristics and diameter are independent risk factors of lung cancer in the patients with SPN.Our prediction model is accurate and sufficient to estimate the malignancy of patients with SPN.