Related Experiment Video
Updated: May 20, 2026
07:13
Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
Published on: April 18, 2025
Development of novel breast cancer recurrence prediction model using support vector machine
Woojae Kim1, Ku Sang Kim, Jeong Eon Lee
1Department of Biomedical Informatics, Ajou University School of Medicine, Suwon, Korea.
Journal of Breast Cancer
|July 19, 2012
Summary
This study developed a new breast cancer recurrence prediction model using support vector machine (SVM) that outperforms existing methods. The model, BCRSVM, offers high accuracy for predicting recurrence within 5 years post-surgery.
Area of Science:
- Oncology
- Medical Informatics
- Biostatistics
Background:
- Accurate prediction of breast cancer recurrence is vital for treatment planning.
- Existing prognostic models have limitations in predictive accuracy.
Purpose of the Study:
- To develop a novel prognostic model using Support Vector Machine (SVM) for predicting 5-year breast cancer recurrence in Korean patients.
- To compare the performance of the SVM model against established prediction models.
Main Methods:
- Retrospective analysis of 679 breast cancer patients' data (1994-2002).
- Key variables included histological grade, tumor size, lymph node metastasis, ER status, and invasion.
- Developed and compared SVM, artificial neural network, and Cox-proportional hazard models.
Main Results:
- The SVM-based model (BCRSVM) achieved superior predictive performance (AUC=0.85) compared to Adjuvant! Online (0.71) and Nottingham Prognostic Index (NPI) (0.70).
- BCRSVM demonstrated high sensitivity (0.89), specificity (0.73), positive predictive value (0.75), and negative predictive value (0.89).
Conclusions:
- The proposed BCRSVM model, utilizing easily obtainable clinical factors, shows significant potential for predicting breast cancer recurrence.
- The model is accessible online for clinical application.
