瘤内瘤周超声影像组学预测乳腺浸润导管癌Ki-67状态
Prediction of Ki-67 status in breast invasive ductal carcinoma using intratumoral and peritumoral ultrasomics
师琳 1钟李长 1谷丽萍1
作者信息
- 1. 上海交通大学医学院附属第六人民医院超声医学科,上海 200233
- 折叠
摘要
目的:探讨瘤内联合瘤周超声影像组学在预测乳腺浸润导管癌Ki-67表达水平中的应用价值.方法:选取接受乳腺手术且病理证实的300例乳腺浸润导管癌患者进行分析,其中包括Ki-67低表达组(Ki-67<30%)169例和高表达组(Ki-67≥30%)131例.选择二维超声图像上病变的最大截面来勾画感兴趣区域,同时自动适形向外扩展5mm获得瘤周区域,提取瘤内和瘤周的影像组学特征.将本研究病例按7∶3随机分组分为训练组(n=210)和验证组(n=90).所有特征经过Z-score标准化、t检验、Pearson相关系数及最小绝对收缩和选择算法进行特征筛选,得到最佳特征组合;最后利用随机森林(RF)模型进行乳腺浸润导管癌Ki-67表达水平的分类,建立瘤内、瘤周以及瘤内联合瘤周超声影像组学模型,借助接受者操作特征曲线评估模型对预测乳腺浸润导管癌Ki-67表达水平的诊断效能.结果:基于瘤内联合瘤周超声影像组学特征的RF模型在预测乳腺癌Ki-67表达状态方面表现更佳,模型在训练组和验证组的曲线下面积分别为0.899(95%CI:0.860～0.939)和0.832(95%CI:0.746～0.917),优于瘤内或瘤周影像组学单一预测效能,差异有统计学意义(P<0.05).结论:基于瘤内联合瘤周超声影像组学特征构建的RF模型,在预测乳腺浸润性导管癌Ki-67表达水平方面表现出良好的诊断效能,有望作为一种非侵入性工具,为乳腺浸润性导管癌的个性化治疗策略提供有用信息.
Abstract
Objective To explore the application value of intratumoral and peritumoral ultrasomics in predicting the expression level of Ki-67 in breast invasive ductal carcinoma.Methods A retrospective study was conducted on 300 patients who underwent breast surgery and were pathologically confirmed to be breast invasive ductal carcinoma,including 169 cases in low expression group(Ki-67<30%)and 131 cases in high expression group(Ki-67≥30%).The largest cross-section of the lesion on the two-dimensional ultrasound images was selected to delineate the region of interest which was automatically expanded outward by 5 mm to obtain the peritumoral area,and the intratumoral and peritumoral ultrasomics features were extracted.The patients were randomly divided into training group(n=210)and validation group(n=90)in a ratio of 7:3.All features were screened through Z-score standardization,t test,Pearson correlation coefficient and least absolute shrinkage and selection operator algorithm to obtain the optimal feature combination.Random forest(RF)model was used to classify the expression level of Ki-67 in breast invasive ductal carcinoma,and 3 models(intratumoral,peritumoral and combined models)were established,whose diagnostic efficacies in predicting the expression level of Ki-67 in breast invasive ductal carcinoma were evaluated using receiver operating characteristic curve.Results The RF model constructed based on the combination of intratumoral and peritumoral ultrasomics features performed better in predicting the Ki-67 expression status in breast cancer,with the AUC values of 0.899(95%CI:0.860-0.939)in training group and 0.832(95%CI:0.746-0.917)in validation group,demonstrating its superior diagnostic efficacy over the intratumoral and peritumoral radiomics alone(P<0.05).Conclusion A RF model constructed based on intratumoral combined with peritumoral ultrasomics features which has good diagnostic performance in predicting the expression level of Ki-67 in breast invasive ducts has the potential to serve as a non-invasive tool,providing useful information for personalized treatment strategy of breast invasive ductal carcinoma.
关键词
乳腺癌/Ki-67水平/超声影像组学/瘤内/瘤周/机器学习Key words
breast cancer/Ki-67 level/ultrasomics/intratumoral/peritumoral/machine learning引用本文复制引用
出版年
2024
NETL
NSTL
万方数据