常规超声联合影像组学的分层可视化模型预测乳腺癌Luminal分子分型及Ki-67表达

doi:10.3969/j.issn.1006-5725.2026.18.003

实用医学杂志 ›› 2026, Vol. 42 ›› Issue (18): 3256-3262.doi: 10.3969/j.issn.1006-5725.2026.18.003

• 肿瘤诊治与预后专栏 • 上一篇

常规超声联合影像组学的分层可视化模型预测乳腺癌Luminal分子分型及Ki-67表达

牛若晨¹, 刘灿², 孙芳¹, 徐星宇¹, 高岩冰¹()

^1. 山东医药大学附属医院超声医学科（山东滨州 256603 ）
^2. 山东医药大学烟台附属医院超声医学科 ;（山东烟台 264100 ）

收稿日期:2026-07-16 修回日期:2026-08-17 接受日期:2026-08-20 出版日期:2026-09-25 发布日期:2026-10-19
通讯作者: 高岩冰
基金资助:
山东省自然科学基金项目(ZR2023MH348); 山东省自然科学基金项目(ZR2023QH231); 山东省医药卫生科技重点项目(202509020999); 山东省医药卫生科技重点项目(202409021312)

A stratified visual model based on conventional ultrasound combined with radiomics for predicting Luminal molecular subtype and Ki-67 expression status in breast cancer

Ruochen Niu¹, Can Liu², Fang Sun¹, Xingyu Xu¹, Yanbing Gao¹()

^1. Department of Ultrasound Medicine，Shandong Medical and Pharmaceutical University Hospital，Binzhou 256603，Shandong，China
^2. Department of Ultrasound Medicine，Yantai Affiliated Hospital of Shandong Medical and Pharmaceutical University，Yantai 264100，Shandong，China

Received:2026-07-16 Revised:2026-08-17 Accepted:2026-08-20 Online:2026-09-25 Published:2026-10-19
Contact: Yanbing Gao
About author:
Niu Ruochen performed the experiments and wrote the article. Xu Xingyu performed the experiments. Sun Fang and Gao Yanbing revised the article. Liu Can designed the study and reviewed the article. All authors read and approved the final manuscript as submitted.

摘要/Abstract

摘要：

目的探讨常规超声联合影像组学构建的联合模型在术前预测乳腺癌Luminal分子分型及Luminal型乳腺癌Ki-67表达状态中的价值。方法回顾性分析375例乳腺癌患者的超声图像，首先通过单因素及多因素分析筛选独立预测因素，同时对影像组学特征进行筛选，随后采用LightGBM算法基于两者构建联合预测模型，用于鉴别乳腺癌Luminal分子分型并预测Luminal型乳腺癌Ki-67表达状态。采用ROC曲线及DCA曲线评价模型性能，并应用SHAP方法分析特征对模型预测结果的贡献。结果构建的Luminal分子分型鉴别模型在训练集和验证集中的AUC分别为0.836和0.776；基于Luminal型乳腺癌患者中建立Ki-67表达状态预测模型，其AUC分别为0.854和0.822，两个模型均表现出良好的预测效能。DCA结果表明两模型在较宽阈值范围内具有较高净收益。SHAP分析显示，original_glcm_InverseVariance及wavelet.LH_firstorder_Median分别为两模型中贡献度最高的特征，对预测结果具有重要影响。结论基于常规超声联合影像组学的分层可视化模型鉴别乳腺癌Luminal分子分型及Luminal型乳腺癌Ki-67表达预测的研究，具有潜在的临床应用价值。

关键词: Luminal乳腺癌, Ki-67, 影像组学, Shapley加性解释方法

Abstract:

Objective This study aims to investigate the value of a combined model integrating conventional ultrasound and radiomics for the preoperative prediction of the Luminal molecular subtype and Ki-67 expression status in patients with Luminal breast cancer. Methods A retrospective analysis was performed on ultrasound imaging data obtained from 375 patients diagnosed with breast cancer. Univariate and multivariate regression analyses were implemented first to identify independent predictive factors， and radiomic features were screened independently. Next， an integrated prediction model based on LightGBM was constructed by combining clinical ultrasound characteristics and radiomic features， to differentiate the Luminal molecular subtype of breast cancer and predict the Ki-67 expression status in patients with Luminal breast cancer. The performance of the established model was evaluated via receiver operating characteristic （ROC） curves and decision curve analysis （DCA）， and the SHapley Additive exPlanations （SHAP） method was applied to interpret the contribution of individual features to the model's prediction output. Results The combined model for Luminal molecular subtype discrimination achieved areas under the receiver operating characteristic curve （AUCs） of 0.836 and 0.776 in the training set and validation set， respectively. For the Ki-67 expression status prediction model constructed for patients with Luminal breast cancer， the AUC values reached 0.854 and 0.822 in the training cohort and validation cohort， respectively. Both models exhibited satisfactory predictive performance. Decision curve analysis demonstrated that the two models delivered substantial net clinical benefits across a broad range of threshold probabilities. SHAP （SHapley Additive exPlanation） analysis revealed that original_glcm_InverseVariance and wavelet.LH_firstorder_Median were the most impactful features for the two models respectively， confirming their critical contributions to the predictive performance of the models. Conclusion The stratified visualization model based on conventional ultrasound combined with radiomics shows potential clinical value for differentiating the Luminal molecular subtype and predicting Ki-67 expression status in breast cancer.

Key words: Luminal breast cancer, Ki-67, radiomics, shapley additive explanations

中图分类号:

R737.9

牛若晨, 刘灿, 孙芳, 徐星宇, 高岩冰. 常规超声联合影像组学的分层可视化模型预测乳腺癌Luminal分子分型及Ki-67表达[J]. 实用医学杂志, 2026, 42(18): 3256-3262.

Ruochen Niu, Can Liu, Fang Sun, Xingyu Xu, Yanbing Gao. A stratified visual model based on conventional ultrasound combined with radiomics for predicting Luminal molecular subtype and Ki-67 expression status in breast cancer[J]. The Journal of Practical Medicine, 2026, 42(18): 3256-3262.

图/表 6

表1

图1

表2

图2

图3

图4

参考文献 22

[1]	Xia Changfa， Dong Xuesi， Li He， et al. Cancer statistics in China and United States， 2022： Profiles， trends， and determinants［J］. Chin Med J， 2022， 135（5）： 584-590. doi：10.1097/CM9.0000000000002108 .
[2]	Shen Jie， Liu Yajing， Liu Aihong， et al. Artificial intelligence-assisted ultrasound screening for breast cancer in China： A prospective， clustered， controlled， population-based study［J］. Breast Cancer Res， 2025， 27（1）： 173. doi：10.1186/s13058-025-02128-0 .
[3]	周楠，汪咏苗，周延鑫，等. 磁共振弹性成像对Luminal型和非Luminal型乳腺癌的鉴别价值［J］. 临床放射学杂志， 2026， 45（5）： 800-806. doi：10.3969/j.issn.1001-9324.2026.05.007 .
[4]	罗为尧，范誉铧，黎一夫，等. 基于深度学习超声模型对乳腺癌分子分型与新辅助治疗早期最佳疗效时机的预测［J］. 实用医学杂志， 2026， 42（9）： 1501-1510. doi：10.3969/j.issn.1006-5725.2026.09.003 .
[5]	Torlakovic E E， Baniak N， Barnes P J， et al. Fit-for-purpose ki-67 immunohistochemistry assays for breast cancer［J］. Lab Investig， 2024， 104（7）： 102076. doi：10.1016/j.labinv.2024.102076 .
[6]	Tabnak P， Hajiesmailpoor Z， Baradaran B， et al. MRI-based radiomics methods for predicting ki-67 expression in breast cancer： A systematic review and meta-analysis［J］. Acad Radiol， 2024， 31（3）： 763-787. doi：10.1016/j.acra.2023.10.010 .
[7]	杨雁雯，周伟，李伟伟，等. 常规超声联合超声造影构建列线图预测小乳腺癌Ki-67表达水平［J］. 中国超声医学杂志， 2025， 41（S1）： 1-5. doi：10.3969/j.issn.1002-0101.2025.z1.001 .
[8]	张欣然，沈燕，胡姣姣，等. 联合临床及超声多参数构建列线图预测人表皮生长因子受体2阳性乳腺癌的应用价值［J］. 实用医学杂志， 2025， 41（18）： 2812-2819. doi：10.3969/j.issn. 1006-5725.2025.18.005 .
[9]	秦浩，张曦，胡雪超，等. 基于全视野数字乳腺X线摄影影像组学与随机森林算法预测乳腺非特殊型浸润性导管癌的HER-2表达的应用［J］. 实用放射学杂志， 2026， 42（5）： 746-749， 780. doi：10.3969/j.issn.1002-1671.2026.05.014 .
[10]	宋宏萍，舒瑞. ACR第6版BI-RADS超声部分更新内容解读［J］. 中华超声影像学杂志， 2025， 34（5）： 369-376. doi：10.3760/cma.j.cn131148-20250217-00080 .
[11]	Yuan Quan， Ye Rongjie， Qian Yao， et al. Machine learning and SHAP value interpretation for predicting the response to neoadjuvant chemotherapy and long-term clinical outcomes in Chinese female breast cancer［J］. Ann Med， 2025， 57（1）： 2541316. doi：10.1080/07853890.2025.2541316 .
[12]	Burstein H J， Curigliano G， Gnant M， et al. Tailoring treatment to cancer risk and patient preference： The 2025 St Gallen International Breast Cancer Consensus Statement on individualizing therapy for patients with early breast cancer［J］. Ann Oncol， 2025， 36（12）： 1433-1446. doi：10.1016/j.annonc.2025.09.007 .
[13]	Hamann M， Christgen M， Gluz O， et al. Endocrine response assessment in HR-positive HER2-negative early breast cancer： Concordance of local versus central Ki67 measurements in the WSG ADAPTcycle trial （n = 5292）［J］. ESMO Open， 2025， 10（8）： 105552. doi：10.1016/j.esmoop.2025.105552 .
[14]	Mo Sijie， Luo Hui， Wang Mengyun， et al. Machine learning radiomics based on intra and peri tumor PA/US images distinguish between luminal and non-luminal tumors in breast cancers［J］. Photoacoustics， 2024， 40： 100653. doi：10.1016/j.pacs.2024. 100653 .
[15]	Li Fang， Zhu Tongwei， Lin Miao， et al. Enhancing ki-67 prediction in breast cancer： Integrating intratumoral and peritumoral radiomics from automated breast ultrasound via machine learning［J］. Acad Radiol， 2024， 31（7）： 2663-2673. doi：10.1016/j.acra.2023.12.036 .
[16]	Huang Yini， Yao Zhao， Li Lingling， et al. Deep learning radiopathomics based on preoperative US images and biopsy whole slide images can distinguish between luminal and non-luminal tumors in early-stage breast cancers［J］. EBioMedicine， 2023， 94： 104706. doi：10.1016/j.ebiom.2023.104706 .
[17]	Wu Jiangfeng， Ge Lifang， Jin Yun， et al. Development and validation of an ultrasound-based radiomics nomogram for predicting the luminal from non-luminal type in patients with breast carcinoma［J］. Front Oncol， 2022， 12： 993466. doi：10.3389/fonc. 2022.993466 .
[18]	Zhou Jianan， Wu Jialing， Zhu Changfu， et al. Ultrasound radiomics in the assessment of breast cancer molecular subtypes： A systematic review and meta-analysis［J］. Med Ultrason， 2025， 27（3）： 315. doi：10.11152/mu-4449 .
[19]	Jiang Chun， Zhu Tong， Zong Yunfei， et al. Defining the optimal Ki67 cutoff values for survival prediction in neoadjuvant chemotherapy-treated patients with breast cancer［J］. Front Surg， 2025， 12： 1697963. doi：10.3389/fsurg.2025.1697963 .
[20]	Song S E， Cho K R， Cho Y， et al. Machine learning with multiparametric breast MRI for prediction of Ki-67 and histologic grade in early-stage luminal breast cancer［J］. Eur Radiol， 2022， 32（2）： 853-863. doi：10.1007/s00330-021-08127-x .
[21]	Crombé A， Kataoka M. Breast cancer molecular subtype prediction： Improving interpretability of complex machine-learning models based on multiparametric-MRI features using SHapley Additive exPlanations （SHAP） methodology［J］. Diagn Interv Imaging， 2024， 105（5）： 161-162. doi：10.1016/j.diii.2024.01.008 .
[22]	Ejiyi C J， Qin Zhen， Monday H， et al. Breast cancer diagnosis and management guided by data augmentation， utilizing an integrated framework of SHAP and random augmentation［J］. Biofactors， 2024， 50（1）： 114-134. doi：10.1002/biof.1995 .

项目	影响因素	β	SE	Z值	Wald χ ²	P值	OR值	95%CI
Luminal分子分型	方位	1.094	0.229	4.772	22.769	< 0.001	2.985	1.905 ~ 4.678
Luminal分子分型	边缘	-0.088	0.089	-0.987	0.973	0.324	0.916	0.768 ~ 1.091
Ki-67表达预测	位置	0.128	0.131	0.981	0.962	0.327	1.137	0.880 ~ 1.470
	形状	1.201	0.432	2.778	7.716	0.005	3.323	1.424 ~ 7.755
	方位	0.954	0.323	2.949	8.699	0.003	2.595	1.377 ~ 4.890
	边缘	0.088	0.110	0.799	0.639	0.424	1.092	0.880 ~ 1.355
	内部回声	-0.494	0.191	-2.587	6.693	0.010	0.610	0.420 ~ 0.887

模型	组别	AUC	95%CI	准确度/%	灵敏度/%	特异度/%
Luminal分子分型联合模型	训练集	0.836	0.789 ~ 0.884	76.3	72.4	83.0
Luminal分子分型联合模型	验证集	0.776	0.680 ~ 0.872	71.3	65.7	82.4
Ki-67表达预测联合模型	训练集	0.854	0.799 ~ 0.910	80.4	82.5	78.0
Ki-67表达预测联合模型	验证集	0.822	0.714 ~ 0.929	80.6	75.0	88.5

常规超声联合影像组学的分层可视化模型预测乳腺癌Luminal分子分型及Ki-67表达

A stratified visual model based on conventional ultrasound combined with radiomics for predicting Luminal molecular subtype and Ki-67 expression status in breast cancer

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 6

参考文献 22

相关文章 14

编辑推荐

Metrics

本文评价

[1]	尹晶,张平洋,汪珺莉,尹薇薇,储小爱,赵文燕. 基于超声影像组学、O-RADS分类与临床因素的卵巢肿瘤诊断模型[J]. 实用医学杂志, 2026, 42(7): 1192-1200.
[2]	朱欣,梁杨标,肖雄,谢佳霖,朱俊燕,陈彦. 基于超声影像组学与深度学习的类风湿关节炎疾病活动度评估模型构建[J]. 实用医学杂志, 2026, 42(15): 2786-2797.
[3]	赖丹惠,江燕辉,叶思婷,庄淑莲,杨爽,薛文,张建兴. 颈动脉斑块超声影像组学特征对颈动脉支架置入术后再狭窄的预测能力[J]. 实用医学杂志, 2025, 41(5): 742-750.
[4]	文国良,方行,张卫. 常规影像和CT影像组学特征鉴别腮腺腺淋巴瘤与恶性肿瘤的价值[J]. 实用医学杂志, 2025, 41(21): 3442-3448.
[5]	杜阳春,郑红雨,陈海宁,郭文文,姚金秀,蓝通柳,肖艳菊. 深度学习超声影像组学列线图模型鉴别Ⅰ和Ⅱ型上皮性卵巢癌[J]. 实用医学杂志, 2025, 41(18): 2920-2927.
[6]	王兆阳,张楠. 基于CT影像组学与中医舌象特征构建肝细胞癌根治术后早期复发的Nomogram预测模型[J]. 实用医学杂志, 2025, 41(16): 2590-2596.
[7]	蔡惠亮,张乾营,黄莹,彭伟生,王成立,杨翠婷,邓娜,章思竹,徐妮娜,韩晓兵. 增强MRI瘤内瘤周影像组学联合临床影像学特征评估肝细胞癌ki-67的表达[J]. 实用医学杂志, 2025, 41(15): 2311-2319.
[8]	何尚鹏,黄炜贤,江燕辉,彭雄强,孟令萃,张建兴. 超声特征与超声影像组学对Bethesda Ⅲ类甲状腺结节良恶性再评价的价值[J]. 实用医学杂志, 2025, 41(12): 1892-1898.
[9]	陈丽萍,罗菊玉,彭章艳,吴秀兰,杨宇宏,石连炎,李笑云,王灵. 肿瘤病灶体积与子宫体积比值和组织中Ki-67、p16蛋白表达与子宫内膜癌病理特征及复发的关联[J]. 实用医学杂志, 2024, 40(23): 3367-3372.
[10]	周汇恩陈婉明王梦蝶罗斯佳陈佳琳何堃过新民. 基于临床-超声影像组学列线图模型预测孕晚期早产的价值 [J]. 实用医学杂志, 2023, 39(14): 1835-1841.
[11]	廖淑婷于向荣 . 能谱CT和人工智能在甲状腺癌诊断中的应用 [J]. 实用医学杂志, 2022, 38(2): 129-1133.
[12]	黄忠江姜增誉李健丁张智星陈文青 . 基于增强CT影像组学联合机器学习鉴别均质性肾透明细胞癌与肾乏脂肪血管平滑肌脂肪瘤 [J]. 实用医学杂志, 2021, 37(17): 2266-2270.
[13]	暴珞宁, 王瑛, 陈东, 刘再毅, . 超声影像组学标签预测乳腺癌前哨淋巴结转移的价值 [J]. 实用医学杂志, 2021, 37(15): 2007-2011.
[14]	李羚, 胡大涛, 夏春华, 李红霞 . 基于CT影像组学诺模图预测头颈部恶性肿瘤淋巴结转移 [J]. 实用医学杂志, 2021, 37(14): 1872-1878.