实用医学杂志 ›› 2023, Vol. 39 ›› Issue (1): 119-123.doi: 10.3969/j.issn.1006⁃5725.2023.01.021

• 医学检查与临床诊断 • 上一篇    下一篇

恶性乳腺非肿块型病变超声特征分析及列线图预测模型构建

吴小茜1 江燕辉1 刘美伶1 谢静雯1 陈淼2 张建兴2    

  1. 1 广州中医药大学第二临床医学院(广州510120);2 广东省中医院超声影像科(广州510120)

  • 出版日期:2023-01-10 发布日期:2023-01-10
  • 基金资助:
    广东省高等教育教学改革项目(编号:A3⁃0402⁃21⁃427⁃001Z04)

Analysis of ultrasound characteristics of malignant breast non⁃mass⁃like lesions and development of a no⁃ mogram

WU Xiaoxi*,JIANG Yanhui,LIU Meiling,XIE Jingwen,CHEN Miao,ZHANG Jianxing.   

  1. The Second Clinical Medical College,Guangzhou University of Chinese Medicine,Guangzhou 510120,China

  • Online:2023-01-10 Published:2023-01-10

摘要:

目的 筛查恶性乳腺非肿块型病变(non⁃mass⁃like lesions,NML)的危险因素,构建 NML 病变 恶性风险预测列线图模型。方法 回顾性分析 577 例乳腺 NML 病变患者的临床资料及超声图像特征,筛查恶性 NML 病变的危险因素,构建列线图预测模型,绘制该模型的 ROC 曲线和校准曲线并行 Hosmer⁃ Lemeshow 拟合优度检验。结果 年龄、病变可触及、微钙化、后方回声衰减、结构扭曲、血流信号丰富和腋 窝淋巴结异常是恶性 NML 病变的危险因素。基于上述变量构建的列线图预测模型 AUC 0.869,灵敏度 74.5%,特异度为 86.8%;校准曲线显示模型预测概率与实际概率之间具有良好的一致性,平均绝对误 =0.007;Hosmer⁃Lemeshow 拟合优度检验显示差异无统计学意义。结论 基于患者临床资料和超声图像 特征构建的 NML 恶性风险预测模型具有较高的临床应用价值、校准度和稳定性,可以协助超声医生对 NML 的诊断及危险分层。

关键词:

非肿块型病变, 列线图, 乳腺癌, 超声

Abstract:

Objective The research aimed to analysis risk factors for malignant breast non⁃mass⁃like lesions(NML),and use these factors to construct a nomogram. Methods The clinical data and ultrasound image characteristics of 577 patients with NML were retrospectively analyzed. The risk factors of malignant NML lesions were screened,and the nomogram was constructed. The ROC curve and calibration curve of the model were drawn and the Hosmer⁃Lemeshow test was performed. Results Age,palpable lesions,microcalcification,acoustic shad⁃ owing,architectural distortion,rich blood flow and axillary enlarged node were risk factors for malignant NML lesions. The nomogram based on the above variables had an AUC of 0.869,a sensitivity of 74.5% and a specificity of 86.8%. The calibration curves showed good agreement between the predicted and actual probabilities,with the mean absolute error = 0.007. And the Hosmer ⁃ Lemeshow test showed there was no statistical difference. Conclusion The nomogram based on the clinical data of patients and ultrasound features has high clinical applica⁃ tion value,calibration degree and stability,which can assist sonographers in the diagnosis and risk stratification of NML.

Key words:

NML, nomogram, breast cancer, ultrasound