孤立性肺结节影像学特征的logistic回归分析

目的采用logistic回归分析对良、恶性孤立性肺结节（SPN）的临床及影像征象进行恶性风险因素分析，以提高早期肺癌的检出率。方法回顾分析43例SPN病人的临床及影像资料，先进行单因素分析，筛选出良、恶性组差异有统计学意义的因素，代入二元logistic回归分析进行多因素分析，得出良、恶性SPN的独立影响因素。结果43例病人共43个SPN，良性组18例（41.9%），恶性组25例（58.1%）。孤立性肺结节良恶性的单因素分析显示，病人为女性、CT影像中结节有分叶、毛刺征、血管集束征、支气管征、胸膜凹陷征和无平直边的结节恶性率较高（P < 0.05~P < 0.01）；多因素logistic回归分析结果显示，毛刺征为恶性SPN的独立危险性因素（P < 0.05），有平直边为良性SPN的独立保护性因素（P < 0.05）。结论良、恶性SPN的各影像征象存在重叠，毛刺征为恶性SPN的独立危险因素，而有平直边为良性结节的独立保护性因素。

Logistic regression analysis of imaging features of solitary pulmonary nodule

ObjectiveTo investigate the clinical characteristics and imaging signs of malignant solitary pulmonary nodule(SPN) to analyze the malignant risk factors, in order to improve the early detection rate of lung cancer.MethodsThe clinical and imaging data of 43 patients with SPN were analyzed using single factor analysis, and the statistical significance factor between benign and malignant groups was screened.These factors were multiplely analyzed using binary logistic regression analysis, and the independent influencing factors in benign and malignant SPN were screened.ResultsAmong 43 patients with SPN, 18 benign cases(41.9%) and 25 malignant cases(58.1%) were identified.The single factor analysis of benign and malignant solitary pulmonary nodules showed that the malignant rates of nodules with lobulation, burr sign, vascular cluster sign, bronchial sign, pleural indentation sign and no straight edge in CT images in female patients were higher(P < 0.05 to P < 0.01).Multiple-factor logistic regression analysis showed that blur sign was the independent risk factor of malignant SPN(P < 0.01), and the straight edge was the independent protective factor of benign SPN(P < 0.05).ConclusionsThe imaging features of benign and malignant SPN overlap.The blur sign is the independent risk factor of malignant SPN, and the straight edge is the independent protective factor of benign nodules.