Logistic回归分析在超声多参数鉴别乳腺良、恶性肿块中的应用

目的：研究高频超声多参数鉴别乳腺良、恶性肿块的价值。方法：记录145个乳腺肿块二维灰阶声像图参数和彩色多普勒血流参数,获取各肿块病理确诊结果,Logistic回归统计分析多参数综合鉴别诊断乳腺肿块良恶性的价值。结果：单独二维灰阶声像图参数的Logistic回归模型综合鉴别乳腺肿块性质的准确率为82.1%,二维灰阶声像图结合彩色多普勒多参数的Logistic回归模型综合鉴别诊断乳腺良恶性肿块的准确率为93.3%。结论：应用二维灰阶声像图结合彩色多普勒多参数的Logistic回归分析能提高乳腺良、恶性肿块鉴别的准确性。

Logistic regression used in the analysis of multiple parameters of high frequency ultrasonography in differentiation of benign from malignant breast mass

Objective To investigate the differential diagnostic value between benign and malignant breast mass by application of high frequency ultrasonography.Methods Characteristics of 145 breast masses on 2-dimenslonal B-ultrasound and color Doppler was recorded.Meanwhile their pathological results examinations were performed.Differentiation between benign and malignant masses was made using Logistic regression of multiple parameters of ultrasonography.Results The accuracy of 2-dimenslonal B-ultrasound applying Logistic regression of multiple parameters to diagnose breast masses was 82.1%,that combined with color Doppler was 93.3%.Conclusion These findings indicate that application of Logistic regression of multiple parameters on 2-dimenslonal B-ultrasound combined with color Doppler ultrasound can improve the accuracy in differential diagnosis of mammary masses.