摘要
目的通过超声、钼靶X线及临床特征综合分析,创立乳腺肿块良恶性的预测Logistic回归模型,以提高乳腺癌诊断的准确率。方法收集广州市增城区人民医院2015年5月-2017年5月住院治疗的女性乳腺疾病患者临床资料,其中乳腺癌患者138例,年龄(48.89±9.11)岁,乳腺良性肿块患者68例,年龄(49.29±9.63)岁;以良恶性病理结果为因变量(二分类),超声、钼靶X线及临床特征作为自变量,创立乳腺肿块良恶性预测的Logistic回归模型。结果病灶质地、钼靶肿块浸润或毛剌、超声形态、超声腋窝淋巴结肿大、钼靶钙化、血流信号和血流信号分级进入Logistic模型,回归模型预报准确率为92.58%,ROC曲线下面积为(0.906±0.018)。结论 Logistic回归筛选出的超声、钼靶X线及临床特征变量,有较准确的预报效应。
Objective In order to improve the diagnostic accuracy of breast cancer,to develop the Logistic regression model to forecast the mammary benign and malignant tumor by comprehensive analysis of ultrasound,X-ray mammography and clinical characteristics. Methods The clinical data of female patients with breast diseases who received hospital treatment in peoples hospital in Zengcheng district from May,2015 to May,2017 were collected. There were 138 breast cancer patients. The average age of them was（ 48. 89 ± 9. 11） years old. There were 68 mammary gland benign tumour patients. The average age of them was（ 49. 29 ± 9. 63）years old. The benign and malignant pathology results were regards as the variable of hospital treatment,and the ultrasound,X-ray mammography and clinical characteristics were regards as the independent variable to develop the Logistic regression model to forecast the mammary benign and malignant tumor. Results The lesions texture,tumor infiltration or burr,and calcification showed by X-ray mammography,shape of tumor,xillary lymph node enlargement,blood flow signals and blood flow grading showed by ultrasound were entered in Logistic regression model. The forecasting accuracy of the regression model was 92. 58%（ 212/229）. The area under ROC curve was（ 0. 906 ± 0. 018）. Conclusion The variables of ultrasound,X-ray mammography and clinical characteristics filtered by the Logistic regression model have accurate prediction effects.
作者
雷小艳
范振强
周颖兰
LEI Xiao Yan;FAN Zhen Qiang;ZHOU Ying Lan(Yong He Community Health Service Centers, Yongning Street, Zengcheng District, Guangzhou , Guangdong Province, 511356, China)
出处
《中国妇幼卫生杂志》
2018年第2期19-22,共4页
Chinese Journal of Women and Children Health
关键词
超声
钼靶X线
临床特征
LOGISTIC模型
乳腺肿瘤
ultrasound
X-ray mammography
clinical characteristics
logistic regression model
breast tumor
