JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Ultrasound-based radiomic nomogram for predicting the invasive status of breast cancer: a multicenter study

  • 0Department of Ultrasound, The Second Affiliated Hospital, Guangzhou University of Chinese Medicine, Guangzhou, Guangdong, China.
European Journal of Medical Research +

|

July 2, 2025

|

July 2, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

This study developed a combined model using conventional ultrasound (CUS) radiomics, clinical, and sonographic features to accurately distinguish invasive breast cancer (IBC) from non-invasive breast cancer (non-IBC). The new model significantly improved diagnostic accuracy, showing potential for clinical use.

Area Of Science

  • Medical Imaging
  • Oncology
  • Radiology

Background

  • Distinguishing invasive breast cancer (IBC) from non-invasive breast cancer (non-IBC) is crucial for treatment planning.
  • Conventional ultrasound (CUS) and radiomic features offer potential for non-invasive cancer assessment.
  • Developing accurate predictive models can improve diagnostic workflows.

Purpose Of The Study

  • To develop and validate a nomogram combining CUS-based radiomic features, clinical, and sonographic data to differentiate IBC from non-IBC.
  • To assess the prognostic value of CUS-based radiomic signatures in predicting breast cancer invasiveness.

Main Methods

  • A retrospective study included 403 IBCs and 221 non-IBCs across multiple institutions.
  • 1125 radiomic features were extracted from CUS images, and Radiomics Scores (Rad-scores) were computed.
  • Logistic regression was used to construct and compare clinical-radiomics, CUS-clinical, and combined CUS-clinical-radiomics nomogram models.

Main Results

  • The combined CUS-clinical-radiomics model demonstrated superior diagnostic performance with AUCs of 0.91 (training), 0.94 (internal test), and 0.90 (external test).
  • This model significantly outperformed other models in distinguishing IBC from non-IBC.
  • The combined model showed a significant increase in sensitivity (91.7%) compared to the Rad-score alone (80.0%).

Conclusions

  • Radiomic features significantly enhance diagnostic accuracy in differentiating invasive from non-invasive breast cancer.
  • The developed CUS-clinical-radiomics nomogram exhibits robust performance and clinical utility for predicting breast cancer invasiveness.
  • This combined model holds promise for clinical translation in breast cancer diagnosis.

Keywords: