Related Experiment Video
Updated: May 28, 2026
03:05
Establishment and Evaluation of a Risk Prediction Model for Pathological Escalation of Gastric Low-Grade Intraepithelial Neoplasia
Published on: February 16, 2024
Machine Learning-Based Survival Prediction Models for Young Patients With Gastric Cancer: Model Development and
Ha Ye Jin Kang1, Wooyeong Jang2, Minsam Ko3
1Department of Computer Science, Semyung University, Jecheon-si, Chungcheongbuk-do, Republic of Korea.
JMIR Cancer
|May 26, 2026
Summary
Machine learning models accurately predict mortality in young gastric cancer (GC) patients. These tools can identify high-risk individuals for tailored treatment, improving outcomes for this growing patient group.
Area of Science:
- Oncology
- Machine Learning
- Biostatistics
Background:
- Gastric cancer (GC) incidence is declining globally, yet cases in younger individuals are rising.
- Developing accurate mortality prediction models for young GC patients is crucial.
Purpose of the Study:
- To develop and evaluate machine learning-based survival models for predicting 3- and 5-year mortality in young GC patients (≤50 years).
Main Methods:
- Utilized data from 813 young GC patients from the Gastric Cancer Public Staging Database (2013-2015).
- Applied Random Survival Forest, Gradient Boosting, Extra Survival Tree, and Cox proportional hazards models.
- Assessed model performance using the concordance index (C-index).
Main Results:
- Random Survival Forest achieved a 95.89% C-index for 3-year and 91.82% for 5-year mortality prediction.
- Extra Survival Tree model showed high performance with a 94.60% C-index for 5-year mortality.
- Tumor stage and size were key predictors, with other variables showing varied contributions.
Conclusions:
- Machine learning models demonstrate high accuracy in predicting mortality for young GC patients.
- These models can aid in identifying high-risk individuals for more aggressive treatment strategies.
