基于常规超声与超声弹性成像技术的列线图预测乳腺癌腋窝淋巴结转移的价值

The application value of a nomogram based on conventional ultrasound and ultrasound elastography for predicting axillary lymph node metastasis of breast cancer

目的探讨基于常规超声与超声弹性成像(UE)技术的列线图预测乳腺癌(BC)腋窝淋巴结(ALN)转移的价值。方法选取2012年1月至2020年10月广州中医药大学第一附属医院诊治的224例女性右侧BC患者作为研究对象。根据ALN术后或穿刺活检的病理结果是否存在转移灶将患者分为转移组(n=128)和非转移组(n=96)。所有患者淋巴结术前均行常规超声及UE检测;乳腺肿瘤及右侧ALN均由术后或穿刺活检的病理诊断结果证实。观察并记录患者ALN短径、长径、短径/长径、距离皮肤深度、皮质厚度和弹性评分,并比较两组数据的差异。采用Logistic回归分析判定转移的独立预测因素,构建预测转移的列线图模型,绘制受试者工作特征(ROC)曲线评估各预测因素和列线图模型的预测能力。结果 ALN短径、短径/长径、皮质厚度和弹性评分是BC患者发生ALN转移的独立预测因素,其中皮质厚度(≥2.92mm)和弹性(≥3分)的预测能力较强。构建的列线图预测转移的ROC曲线下面积(AUC)为0.996(0.995~0.998),预测能力优良。结论常规超声和UE技术均可预测BC患者是否发生ALN转移,基于两种超声构建的列线图模型可作为量化工具用于BC患者ALN转移的预测,有助于术前治疗方案的制定,提高患者预后水平。

Objective To explore the value of nomogram based on conventional ultrasound and ultrasound elastography(UE) in predicting breast cancer(BC) axillary lymph node(ALN) metastasis. Methods Female patients with right BC who were treated in the First Affiliated Hospital of Guangzhou University of Chinese Medicine from January 2012 to October 2020 were selected as the research objects. The study subjects were divided into metastatic group and non-metastatic group according to whether ALN metastasis occurred. The lymph nodes of all patients underwent routine ultrasound and UE detection before surgery;breast tumors and right ALNs were confirmed by pathological diagnosis results after surgery or needle biopsy. Observe and record the ALN short diameter, long diameter, short diameter/long diameter, skin depth, cortical thickness and elasticity score of the study subjects, and compare the difference between the data of two groups. Logistic regression analysis was used to determine the independent predictors of metastasis, construct a nomogram model for predicting metastasis, and receiver operating curve were used to evaluate the predictive ability of each predictor and nomogram model. Results ALN short diameter, short diameter/long diameter, cortical thickness and elasticity score were independent predictors of ALN metastasis in BC patients, among which cortical thickness(≥2.92 mm) and elasticity(≥3 points) had strong predictive abilities. The area under ROC curve(AUC) of the nomograms for predicting metastasis was 0.996(0.995-0.998), indicating excellent prediction ability. Conclusions Both conventional ultrasound and UE techniques can predict whether ALN metastasis occurs in BC patients. The nomogram model based on the two ultrasounds can be used as a quantitative tool to predict ALN metastasis in BC patients, which was helpful for the formulation of preoperative treatment plans, and improving the prognosis of patients.