超声多参数回归模型评估乳腺癌腋窝淋巴结转移的价值

目的：建立以乳腺癌腋窝淋巴结超声诊断特征为变量的Logistic回归模型，评价常规超声及实时弹性成像技术对乳腺癌腋窝淋巴结有无转移鉴别诊断的价值。方法对病理证实的112例乳腺癌患者的113个腋窝淋巴结的常规超声及弹性成像各因素进行综合分析，建立二分类Logistic回归模型，通过绘制ROC曲线评价模型的诊断效能。结果在纳入的113个淋巴结中，术后病理证实28个无转移，85个有转移。经过前进法Logistic回归分析，筛选出弹性应变率比值（ SR）、弹性评分及形态等3个指标对乳腺癌腋窝淋巴结转移诊断有统计学意义的特征变量。 Logistic回归模型对乳腺癌腋窝淋巴结转移预报的正确率为93．8％， ROC曲线下面积为0．962。结论二分类Logistic回归模型对腋窝淋巴结性质有较好的诊断效能，弹性成像提高了常规超声诊断腋窝淋巴结的准确度。

Evaluation of axillary lymph node metastasis of breast cancer via an ultrasonic multi parameter regression model

Objective To evaluate conventional ultrasound combined with real-time elastography in differential diagnosis of metastatic axillary lymph nodes of breast cancer with a logistic regression model.Methods Conventional ultrasound and real-time elastography were performed in 112 breast cancer patients with 113 axillary lymph nodes.All the cases were confirmed with pathological examinations after surgery. A binary logistic regression model based on ultrasonographic features was developed.A receiver operator characteristic ( ROC) curve was constructed to assess the performance of the model.Results There were 28 nodes with no metastasis and 85 with metastasis in a total of 113 axillary lymph nodes.Three ultrasonographic features including the elastic strain ratio(SR), elastic score and shape were finally selected into the logistic regression model .The correct rate of the logistic regression model for predicting axillary lymph node metastasis was 93.8%, and the area under ROC curve was 0.962.Conclusion The binary classification logistic regression model has a good diagnostic efficacy in the diagnosis of benign and malignant axillary lymph nodes,while real-time elastography can improve the accuracy of conventional ultrasound in the diagnosis of axillary lymph nodes.