Random Survival Forest Predicts Immunochemotherapy Efficacy
Researchers at The First Affiliated Hospital of Nanjing Medical University retrospectively developed and temporally validated machine learning models (Jan 2018–Oct 2023 training/validation; Nov 2023–Sep 2024 temporal cohort) to predict first-line immunochemotherapy outcomes in 316 advanced gastric cancer patients. The random survival forest (RSF) model outperformed LASSO‑Cox, XGBoost, and SVM, identifying age, histology, CD19+ B cells, CD16+CD56+ NK cells, and liver metastasis as key predictors, enabling risk stratification.
Key Points
- 1Demonstrates RSF model outperforms LASSO‑Cox, XGBoost, and SVM predicting progression‑free survival with higher AUC
- 2Highlights key predictors: age, histological subtype, CD19+ B cells, CD16+CD56+ NK cells, and liver metastasis
- 3Enables clinicians to stratify AGC patients into risk groups to guide personalized first‑line immunochemotherapy
Scoring Rationale
Provides a validated, clinically actionable ML prognostic model; limited novelty and single-center design reduce broader generalizability
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
