Related Experiment Video
Updated: May 24, 2026
07:15
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
A Simple and Robust MRI Radiomics Feature for Predicting Pathological Complete Response: A Proof-of-Concept Study in
Hongen Li1, Li Zhang1, Yihui Zeng1
1Department of Radiology, Guangdong Women and Children Hospital/Women and Children's Hospital, Southern University of Science and Technology, Guangzhou, China.
Cancer Reports (Hoboken, N.J.)
|May 22, 2026
Summary
This study developed a radiomics model using mid-treatment MRI scans to predict pathological complete response (pCR) to neoadjuvant chemotherapy (NAC) in breast cancer patients, showing moderate predictive capability.
Area of Science:
- Radiology
- Oncology
- Medical Imaging Analysis
Background:
- Predicting pathological complete response (pCR) to neoadjuvant chemotherapy (NAC) in breast cancer is crucial but challenging.
- Early prediction allows for timely treatment adjustments.
Purpose of the Study:
- To evaluate a radiomics model using mid-treatment dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) for early pCR prediction in breast cancer patients receiving NAC.
Main Methods:
- Retrospective analysis of 119 breast cancer patients treated with NAC.
- Extraction of 851 radiomic features from DCE-MRI, followed by dimensionality reduction to 88 key features.
- Development of a support vector machine (SVM) classifier using three optimal wavelet-based texture features, internally validated.
Main Results:
- The radiomics model achieved an AUC of 0.667 (training) and 0.647 (validation) with accuracies of 66.27% and 73.49%, respectively.
- The model demonstrated moderate and stable predictive performance for pCR.
- Exploratory analysis suggested potential clinical utility for decision-making.
Conclusions:
- A minimalist radiomics model based on mid-treatment MRI shows moderate predictive capability for pCR after NAC in breast cancer.
- The model has potential for aiding clinical decision-making, but requires external validation.
- Future research should focus on multicenter validation and multimodal feature integration.
