EUVIMEDEUVIMED entdecken
Crossref

Establishing a Multimodal Model for Predicting Lymphovascular Invasion in Breast Cancer Using Deep Transfer Learning Based on Ultrasound and Dynamic Contrast-Enhanced Magnetic Resonance Imaging

Miaomiao He, Feifei Zhu, Feng Jiang, Yinhua Liu, Yuxin Guo, Yuan Yang, Hao Jiang, Fei Xiong, Weibo Cheng, Zhengyuan Xu, Biaohu Liu

Oncology Research and Treatment · 10.07.2026 · 1-18

DOI: 10.1159/000553438 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Oncology Research and Treatment, online erschienen am 10.07.2026. Verfasst von Miaomiao He und 10 weiteren Autor:innen. Beteiligte Einrichtung: Department of Ultrasound Medicine, First Affiliated Hospital of Wannan Medical University, Wuhu, China. Hauptthemen laut MEDLINE: Breast Neoplasms, Ultrasonography, Mammary, Deep Learning. Die Arbeit nennt 37 Literaturangaben.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction

Develop a multimodal model based on ultrasound (US) and dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) data for predicting lymphovascular invasion (LVI) in breast cancer.

Method

A total of 326 patients with breast cancer were enrolled in this retrospective study. All participants underwent preoperative US and DCE-MRI. Data from Center 1 (n = 167) were divided into a training set and an internal validation set, while data from Center 2 (n = 159) were used as the external validation set for analysis. A fusion feature set was constructed from intratumoral regions of US and DCE-MRI using deep transfer learning (DTL) features. Among three machine learning models – support vector machine, k-nearest neighbors, and multi-layer perceptron (MLP) – the MLP classifier trained with a feature subset selected via LASSO achieved optimal performance across multiple feature selection methods and was thus designated as the final model. The diagnostic performance and clinical utility of the multimodal model were validated and assessed through the area under the receiver operating characteristic curve and decision curve analysis. Model interpretability was assessed using Shapley Additive Explanations (SHAP).

Result

Results indicate that the multimodal model, utilizing DTL features from US and DCE-MRI within the tumor region, demonstrates outstanding performance in classifying LVI in breast cancer. The area under the curve (AUC) values of the model were 0.987 in the training set, 0.967 in the internal validation set, and 0.942 in the external validation set. Its performance was significantly better than the models using DTL features derived from single-region US (AUC: 0.966, 0.852, 0.789) or DCE-MRI (AUC: 0.936, 0.716, 0.746), indicating great value for clinical application.

Conclusion

Multimodal models based on DTL features can accurately distinguish LVI status in breast cancer patients. This model demonstrates promising application prospects in distinguishing LVI in breast cancer, with its value lying in effectively enhancing diagnostic accuracy and robustness.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Breast NeoplasmsUltrasonography, MammaryDeep LearningHumansFemaleDynamic Contrast Enhanced Magnetic Resonance ImagingRetrospective StudiesMiddle AgedNeoplasm InvasivenessLymphatic MetastasisPredictive Learning ModelsAdultMultilayer PerceptronsMultimodal ImagingAgedMagnetic Resonance ImagingContrast Media

Stichwörter der Autor:innen

Breast cancerDeep transfer learningLymphovascular invasionMagnetic resonance imagingUltrasound

Substanzen

Contrast Media

Autor:innen und Institutionen

  • Department of Ultrasound Medicine, First Affiliated Hospital of Wannan Medical University, Wuhu, China.
  • Department of Pathology, First Affiliated Hospital of Wannan Medical University, Wuhu, China.
  • Department of Medical Imaging, First Affiliated Hospital of Wannan Medical University, Wuhu, China.
  • Department of Ultrasound Medicine, Second Affiliated Hospital of Wannan Medical University, Wuhu, China.
  • School of Medical Imaging, Wannan Medical University, Wuhu, China, xuzy@wnmc.edu.cn.
  • Department of Ultrasound Medicine, First Affiliated Hospital of Wannan Medical University, Wuhu, China, 13515538949@163.com.

Zitieren

APA

He, M., Zhu, F., Jiang, F., Liu, Y., Guo, Y., Yang, Y., Jiang, H., Xiong, F., Cheng, W., Xu, Z., & Liu, B. (2026). Establishing a Multimodal Model for Predicting Lymphovascular Invasion in Breast Cancer Using Deep Transfer Learning Based on Ultrasound and Dynamic Contrast-Enhanced Magnetic Resonance Imaging. Oncology Research and Treatment, 1-18. https://doi.org/10.1159/000553438

Vancouver

He M, Zhu F, Jiang F, Liu Y, Guo Y, Yang Y, et al. Establishing a Multimodal Model for Predicting Lymphovascular Invasion in Breast Cancer Using Deep Transfer Learning Based on Ultrasound and Dynamic Contrast-Enhanced Magnetic Resonance Imaging. Oncology Research and Treatment. 2026:1-18. doi:10.1159/000553438

BibTeX
@article{He2026,
  author = {He, Miaomiao and Zhu, Feifei and Jiang, Feng and Liu, Yinhua and Guo, Yuxin and Yang, Yuan and Jiang, Hao and Xiong, Fei and Cheng, Weibo and Xu, Zhengyuan and Liu, Biaohu},
  title = {Establishing a Multimodal Model for Predicting Lymphovascular Invasion in Breast Cancer Using Deep Transfer Learning Based on Ultrasound and Dynamic Contrast-Enhanced Magnetic Resonance Imaging},
  journal = {Oncology Research and Treatment},
  year = {2026},
  pages = {1--18},
  doi = {10.1159/000553438}
}
Kostenlos starten →