目的应用logistic回归对肺良 性肿瘤形态学征象进行分析。方法 收集 52例肺良性肿瘤及30例恶性肿瘤，对形 态学征象进行单因素X2 检验及相关因素 logistic回归分析，形态学征象包括边缘 征、钙化征、胸膜继发改变、肿瘤位置、 分叶征、肿瘤形状及囊变征。结果 经过 单因素X2 检验，得到肺良性肿瘤形态学征 象中，边缘征、钙化征、胸膜继发改变、 肿瘤所在位置、分叶征、肿瘤形状及囊变 征的X2 值分别为26.801、7.329、0.008、 0.444、11.134、11.134及0.596，其中只 有边缘征、钙化征、分叶征及肿瘤形状的 P值<0.05，有统计学差异；进行相关因素 logistic回归分析后，得到边缘征的Wald 统计量值最大，相伴概率最小，分别为 9.779和0.002。结论 肺良性肿瘤形态学 征象中边缘征、钙化征、分叶征及肿瘤形 状对鉴别诊断有影响，其中边缘征影响程 度最高。

Objective To analyse morphological signs of pulmonary benign tumor with logistic regression. Methods 52 patients with pulmonary benign tumor and 30 cases malignant tumor are collected, Single factor X2-test and related factors logistic regression of morphological signs are carried out, Morphological signs include edge character, calcification, pleural secondary change, tumor location, lobulation, tumor shape and cystic degeneration. Results Through the single factor X2-test, morphological signs of the edge character, calcification, pleural secondary change, tumor location, lobulation, tumor shape and cystic degeneration in pulmonary benign tumor,the X2 is respectively 26.801, 7.329, 0.008, 0.444, 11.134, 11.134 and 0.596, only the edge character, calcification, lobulation and tumor shape with P<0.05, existing statistically significant.after logistic regression analysis for the related factors, to get Wald statistic value maximum and concomitant probability to a minimum of the edge character, it is respectively 9.779 and 0.002. Conclusion Morphological signs of edge character, calcification, lobulation and tumor shape in pulmonary benign tumor could influence antidiastole, and the most valuable sign is edge character.