Related Experiment Video
Updated: Dec 1, 2025
07:41
Performing Data Mining And Integrative Analysis Of Biomarker in Breast Cancer Using Multiple Publicly Accessible Databases
Published on: May 17, 2019
9.3K
Multidimensional Machine Learning Personalized Prognostic Model in an Early Invasive Breast Cancer Population-Based
Xiaorong Zhong1, Ting Luo1, Ling Deng2
1Department of Head, Neck and Mammary Gland Oncology, Cancer Center, West China Hospital, Sichuan University, Chengdu, China.
JMIR Medical Informatics
|November 9, 2020
Summary
A new breast cancer prognosis model developed in China shows high accuracy in predicting disease progression and mortality. This advanced model offers improved calibration compared to existing tools, aiding clinical decisions for Chinese patients.
Area of Science:
- Oncology
- Biostatistics
- Medical Informatics
Background:
- Existing breast cancer prognostic models (e.g., Adjuvant! Online, PREDICT) are population-specific, primarily validated in Western cohorts.
- Suboptimal prediction accuracy has been observed when applying these models in non-European populations.
Purpose of the Study:
- To develop an advanced breast cancer prognosis model utilizing tumor, demographic, and treatment characteristics.
- To predict disease progression, cancer-specific mortality, and all-cause mortality in a large Chinese breast cancer cohort.
Main Methods:
- Utilized extreme gradient boosting to develop the predictive model.
- Collected data from 5293 women with stage I-III invasive breast cancer (2000-2013).
- Assessed model performance using Area Under the Receiver Operating Characteristic Curve (AUROC) and calibration analysis, comparing with the PREDICT model.
Main Results:
- The model demonstrated strong predictive performance across training, testing, and validation sets, with AUROC values ranging from 0.76 to 0.88.
- Calibration analysis confirmed good agreement between predicted and observed outcomes within a 5-year period.
- The developed model showed similar AUROC but improved calibration compared to the PREDICT model.
Conclusions:
- The developed prognostic model exhibits high discrimination and good calibration for breast cancer outcomes.
- This model has the potential to enhance prognosis prediction and support clinical decision-making for breast cancer patients in China.
