Machine learning-based fusion model for predicting HER2 expression in breast cancer by Sonazoid-enhanced ultrasound - Original research study

Aug 5, 2026

Upload by Rowan

Machine learning model development focuses on predicting human epidermal growth factor receptor 2 (HER2) expression in breast cancer using Sonazoid-enhanced ultrasound. Prospective multi-center data from 17 Chinese hospitals (Aug 2020–Feb 2021) were used without neoadjuvant chemotherapy, with HER2 status determined by immunohistochemistry or FISH. Features from clinical factors, B-mode, CEUS, and time-intensity curves were selected via LASSO, and multiple classifiers were trained and validated across separate hospital groups. A fusion approach combining logistic regression, SVM, and XGBoost delivered strong validation discrimination, supporting robust HER2 stratification.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages11

Reading Time28 min

AccessFree to download

Purpose

Target and objective

Materials and methods

Study design and cohorts

HER2 reference assessment

Feature selection

Prediction models

Results

Selected features

Model performance

Validation

Fusion model performance metrics

OPEN ACCESS EDITED BY Juan Wang, The Second Affiliated Hospital of Xi’an Jiaotong University, China REVIEWED BY Qiao Hu, The People’s Hospital of Guangxi Zhuang Autonomous Region, China Xuejun Ni, Affiliated Hospital of Nantong University, China *CORRESPONDENCE Jie Yu [jiemi301@163.com](jiemi301@163.com) †These authors have contributed equally to this work and share first authorship RECEIVED 01 March 2025 ACCEPTED 30 April 2025 PUBLISHED 21 May 2025 CITATION Zhang H, Lang M, Shen H, Li H, Yang N, Chen B, Chen Y, Ding H, Yang W, Ji X, Zhou P, Cui L, Wang J, Xu W, Ye X, Liu Z, Yang Y, Wei T, Wang H, Yan Y, Wu C, Wu Y, Shi J, Wang Y, Fang X, Li R, Liang P and Yu J (2025) Machine learning-based fusion model for predicting HER2 expression in breast cancer by Sonazoid-enhanced ultrasound: amulticenter study. Front. Med. 12:1585823 . doi: 10.3389/fmed.2025.1585823 COPYRIGHT © 2025 Zhang, Lang, Shen, Li, Yang, Chen, Chen, Ding, Yang, Ji, Zhou, Cui, Wang, Xu, Ye, Liu, Yang, Wei, Wang, Yan, Wu, Wu, Shi, Wang, Fang, Li, Liang and Yu. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. TYPE Original Research PUBLISHED 21 May 2025 DOI 10.3389/fmed.2025.1585823 Machine learning-based fusion model for predicting HER2 expression in breast cancer by Sonazoid-enhanced ultrasound: amulticenter study Huiting Zhang 1†, Manlin Lang 1†, Huiming Shen 2, Hang Li3, Ning Yang4, Bo Chen 5, Yixu Chen 6, Hong Ding7, Weiping Yang 8, Xiaohui Ji 9, Ping Zhou 10, Ligang Cui 11, Jiandong Wang 12, Wentong Xu 12, Xiuqin Ye 13, Zhixing Liu 14, Yu Yang 15, Tianci Wei 16, Hui Wang 17, Yuanyuan Yan 18, Changjun Wu 19, Yiyun Wu 20, Jingwen Shi 21, Yaxi Wang 22, Xiuxia Fang 22, Ran Li 23, Ping Liang 1 and Jie Yu 1* 1 Department of Interventional Ultrasound, PLA Medical College & Chinese PLA General Hospital, Beijing, China, 2 Department of Ultrasound, Zhongda Hospital, Nanjing, China, 3 Department of Breast Surgery, Affiliated Hospital of Putian University, Putian, China, 4 Department of Ultrasound, Xingcheng People’s Hospital, Xingcheng, China, 5 Department of Ultrasound Medicine, Lu’ an People’s Hospital of Anhui Province, Liuan, China, 6 Department of Ultrasound, The Fifth People's Hospital of Chengdu, Chengdu, China, 7 Department of Ultrasound, Huashan Hospital, Shanghai, China, 8 Department of Ultrasound, Guangxi Medical University Cancer Hospital, Nanning, China, 9 Department of Ultrasound, The Fourth Hospital of Hebei Medical University, Shijiazhuang, China, 10 Department of Ultrasound, The Third Xiangya Hospital, Changsha, China, 11 Department of Ultrasound, Peking University Third Hospital, Beijing, China, 12General Surgery, Chinese PLA General Hospital, Beijing, China, 13 Department of Ultrasound, First Affiliated Hospital of Southern University of Science and Technology, Second Clinical College of Jinan University, Shenzhen Medical Ultrasound Engineering Center, Shenzhen People’s Hospital, Shenzhen, China, 14 Department of Ultrasound Medicine, The First Affiliated Hospital of Nanchang University, Nanchang, China, 15 Department of Ultrasound, Beijing Friendship Hospital, Beijing, China, 16 Department of Ultrasound, The 2nd Affiliated Hospital of Harbin, Harbin, China, 17 Department of Ultrasound, China-Japan Union Hospital of Jilin University, Changchun, China, 18 Department of Ultrasound, Zhengzhou Central Hospital, Zhengzhou, China, 19 Department of Ultrasound, The First Affiliated Hospital of Harbin Medical University, Harbin, China, 20 Department of Ultrasound, Affiliated Hospital of Nanjing University of C

FAQs

What does the study aim to predict in breast cancer patients?

It aims to predict HER2 expression status using Sonazoid-enhanced ultrasound within a machine learning-based fusion model.

How was HER2 expression determined as the reference standard?

HER2 expression was assessed using immunohistochemistry or fluorescence in situ hybridization (FISH).

Which modeling approach performed best in validation?

A fusion model combining logistic regression, SVM, and XGBoost achieved the highest validation ROC area under the curve for both HER2 3+ and HER2 2+/1+.

Aug 5, 2026

Upload by Rowan