JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Prediction of axillary lymph node metastasis using a magnetic resonance imaging radiomics model of invasive breast cancer primary tumor

  • 0Division of Life Sciences and Medicine, School of Biomedical Engineering (Suzhou), University of Science and Technology of China, Suzhou, Jiangsu, 215163, China.
Cancer Imaging : the Official Publication of the International Cancer Imaging Society +

|

September 13, 2024

|

September 13, 2024

View abstract on PubMed

Summary

This summary is machine-generated.

Breast MRI radiomics effectively predicts axillary lymph node metastasis (ALNM). Combining dynamic contrast-enhanced (DCE), T2-weighted, and diffusion-weighted imaging sequences offers the best noninvasive prediction tool for ALNM in breast cancer patients.

Area Of Science

  • Radiology
  • Oncology
  • Medical Imaging Analysis

Background

  • Investigating the clinical utility of breast MRI radiomics for predicting axillary lymph node metastasis (ALNM).
  • Comparing the diagnostic performance of various MRI sequence combinations for ALNM prediction.

Purpose Of The Study

  • To evaluate the effectiveness of breast MRI radiomics in predicting ALNM.
  • To determine the optimal combination of MRI sequences for ALNM prediction.

Main Methods

  • Radiomics features were extracted from 141 invasive breast cancer patients' preoperative MRI scans.
  • Single- and multisequence radiomics models were built using logistic regression.
  • Model performance was assessed using AUC, accuracy, sensitivity, specificity, and precision, and compared to radiologists' diagnoses.

Main Results

  • The best single-sequence ALNM classifier used DCE postcontrast phase 1 (AUC=0.891 for test set 1, AUC=0.619 for test set 2).
  • The optimal multisequence classifiers combined DCE postcontrast phase 1, T2-weighted, and diffusion-weighted imaging (AUC=0.910 for test set 1, AUC=0.717 for test set 2).
  • These radiomics models outperformed both junior and senior radiologists in diagnostic accuracy.

Conclusions

  • A combination of DCE postcontrast phase 1, T2-weighted, and diffusion-weighted imaging radiomics features demonstrates superior performance in predicting ALNM.
  • This study highlights a promising noninvasive tool for ALNM prediction in breast cancer management.

Keywords: