Related Experiment Video
Updated: May 9, 2026
Establishing a Competing Risk Regression Nomogram Model for Survival Data
Published on: October 23, 2020
Late distant recurrence prediction model in premenopausal women with ER-positive/HER 2-negative breast cancer: A
Dong Seung Shin1, Janghee Lee2, Eunhye Kang3
1Division of Breast Surgery, Department of Surgery, Samsung Medical Center, Sungkyunkwan University of Medicine, Seoul, Republic of Korea.
Background:
Late distant recurrence (DR) remains a significant challenge in estrogen receptor (ER)-positive/Human Epidermal Growth Factor Receptor 2 (HER2)-negative breast cancer, especially in premenopausal patients. This study aimed to develop a machine-learning model predicting late DR risk in premenopausal patients and to assess the clinical benefit of extended endocrine therapy (ET) according to risk stratification.
Methods:
This retrospective multicenter study included patients aged ≤45 years with ER-positive/HER2-negative breast cancer who underwent surgery between January 2000 and December 2011. This study was designed as a landmark analysis, with the effective baseline set at 5 years after surgery. Eligible patients had five to 10 years of follow-up and received adjuvant ET for at least two years. The primary outcome was late DR, defined as distant metastasis occurring between five and 10 years after surgery.
Results:
Among 1701 included patients (median age, 41 years), late DR occurred in 108 patients (6.3%). A machine-learning model using eight clinicopathologic variables demonstrated strong predictive performance (AUC = 0.78). Patients classified as high-risk by the model exhibited significantly worse late DMFS compared to low-risk patients (HR,7.36; 95% CI,4.43-12.20; P<0.001). Among high-risk patients, those who received extended ET had significantly improved late DMFS compared to those who did not (HR,0.32; 95% CI,0.18-0.55; P < 0.001). In low-risk patients, extended ET was not associated with a statistically significant benefit (HR,0.45; 95% CI,0.16-1.22; P = 0.081).
Conclusion:
The machine-learning model effectively stratified patients into distinct DR risk groups and highlighted the benefit of extended ET in high-risk patients. This model supports tailored decision-making regarding extended ET in premenopausal patients with ER-positive/HER2-negative breast cancer.
