Radiomics models to predict axillary lymph node metastasis in breast cancer and analysis of the biological significance of radiomic features

ObjectivesTo explore the effectiveness of radiomics in predicting axillary lymph node metastasis (ALNM) and the relationship between radiomics features and genes.MethodThe 379 patients with breast cancer (186 ALNM-positive and 193 ALNM-negative) recruited from three hospitals were divided into the t...

Full description

Saved in:
Bibliographic Details
Main Authors: Xinhua Li, Minping Hong, Zhendong Lu, Zilin Liu, Lifu Lin, Hongfa Xu
Format: Article
Language:English
Published: Frontiers Media S.A. 2025-06-01
Series:Frontiers in Oncology
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/fonc.2025.1546229/full
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:ObjectivesTo explore the effectiveness of radiomics in predicting axillary lymph node metastasis (ALNM) and the relationship between radiomics features and genes.MethodThe 379 patients with breast cancer (186 ALNM-positive and 193 ALNM-negative) recruited from three hospitals were divided into the training (n=224), testing (n=96), and validation (n=59) cohorts. The Cancer Imaging Archive-The Cancer Genome Atlas (TCIA-TCGA) group included 107 patients with breast cancer. A total of 1888 intratumoral and peritumoral radiomics features were extracted from DCE-MRI sequences. Radiomics models were established using a multivariate regression algorithm for each region and their combinations. Clinical and combined nomogram models integrating the Radscore with clinical risk factors were constructed. The biological significance of the radiomic features was analyzed by combining the TCIA database.ResultsThe area under the ROC curve (AUC) of radiomics model in the external validation was 0.760 (95% confidence interval [CI]: 0.626-0.874). The performance of the nomogram combined model (AUC: 0.818; 95% CI:0.702-0.916) surpassed those of both the radiomics and clinical models (AUC: 0.753; 95% CI: 0.630-0.869). Additionally, the DCA results demonstrated the usefulness of the radiomics and nomogram model.ConclusionMRI-based radiomics has the potential to predict the ALNM status in patients with invasive breast cancer. Additionally, radiogenomic analysis demonstrated a correlation between radiomic features and the immune microenvironment.
ISSN:2234-943X