Related Experiment Video
Updated: Jun 17, 2026
07:15
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Creation and Verification of a Machine Learning-Based Model for Predicting Breast Cancer Risk in Patients with
Jiahua Zhao1, Bo Liang1, Jiaxin Yan1
1Department of Medical Ultrasound, Affiliated Hospital 2 of Nantong University, Nantong, China.
Current Medical Imaging
|June 16, 2026
Summary
This study developed a stable logistic regression (LR) model to predict breast cancer (BC) risk in BI-RADS 4 lesions. The LR model aids clinicians in early BC identification and treatment for improved patient outcomes.
Area of Science:
- Medical Imaging
- Machine Learning
- Oncology
Background:
- Breast Imaging Reporting and Data System (BI-RADS) 4 lesions require accurate risk assessment for breast cancer (BC).
- Machine learning (ML) offers potential for developing predictive models in this area.
Purpose of the Study:
- To develop and validate an ML-based model for predicting BC occurrence in BI-RADS 4 patients.
- To identify independent risk factors for BC in this cohort.
Main Methods:
- Retrospective analysis of 216 breast lesions from 212 patients.
- Logistic and LASSO regressions for risk factor identification.
- Construction and comparison of eight ML models, selecting the most stable one (LR).
Main Results:
- Six independent risk factors for BC were identified.
- The logistic regression (LR) model demonstrated superior stability compared to other ML algorithms.
- A predictive nomogram was constructed based on the selected LR model.
Conclusions:
- The LR model effectively enhances the capacity to identify BI-RADS 4 lesions as malignant.
- This model can assist clinicians in early BC detection and treatment planning.
- Limitations include the single-center nature and relatively small sample size.
