Related Experiment Video
Updated: Oct 6, 2025
10:37
A Multimodal Imaging Framework to Advance Phenotyping of Living Label-free Breast Cancer Cells
Published on: August 22, 2025
502
Characterizing breast masses using an integrative framework of machine learning and CEUS-based radiomics
Bino A Varghese1, Sandy Lee2, Steven Cen2
1Keck School of Medicine, University of Southern California, 1441 Eastlake Avenue, Ground Floor, G360, Los Angeles, CA, 90033, USA. bino.varghese@med.usc.edu.
Journal of Ultrasound
|January 18, 2022
Summary
Radiomics analysis of contrast-enhanced ultrasound (CEUS) effectively differentiates benign from malignant breast masses. This approach using texture metrics can help reduce unnecessary breast biopsies.
Area of Science:
- Medical Imaging
- Radiology
- Oncology
Background:
- Distinguishing benign from malignant breast masses is crucial for appropriate patient management.
- Ultrasound-guided biopsy is often performed for suspicious masses, but can lead to unnecessary procedures for benign cases.
Purpose of the Study:
- To evaluate the performance of radiomics analysis of contrast-enhanced ultrasound (CEUS) in differentiating benign from malignant breast masses.
- To assess the potential of CEUS-based radiomics to reduce unnecessary breast biopsies.
Main Methods:
- Retrospective analysis of 131 women with suspicious breast masses (BI-RADS 4a-4c) who underwent CEUS and biopsy.
- Extraction of 112 radiomic features from CEUS images across four phases (precontrast, early, peak, delay).
- Application of machine learning models (AdaBoost, Random Forest) with tenfold cross-validation to predict mass malignancy.
Main Results:
- Univariate analysis identified 35 radiomic variables (p < 0.05) differentiating benign from malignant masses.
- AdaBoost model achieved an Area Under the Curve (AUC) of 0.72 (95% CI: 0.56, 0.89).
- Random Forest model achieved an AUC of 0.71 (95% CI: 0.56, 0.87).
Conclusions:
- CEUS-based texture metrics demonstrate capability in distinguishing benign from malignant breast masses.
- Radiomics analysis of CEUS shows promise as a tool to potentially decrease the rate of unnecessary breast biopsies.
