This study aims to develop and validate a Random Survival Forest (RSF) model for predicting long-term survival in elderly patients following curative resection for gastric cancer. The study is a retrospective multi-center analysis involving patients aged 75 and above who underwent gastric resection from January 2009 to December 2018 at nine top-tier hospitals in China. An online prognostic tool is introduced to assist clinicians in predicting patient prognosis and customizing treatment and follow-up strategies.

This retrospective multi-center study focuses on the development and validation of a predictive model for elderly gastric cancer patients. Data were collected from 16,344 gastric cancer patients, with 1,202 elderly patients ultimately included after applying exclusion criteria. Patients were randomly divided into training and testing cohorts in a 7:3 ratio. The study was approved by the institutional review boards with a waiver of informed consent due to the use of anonymized secondary data.

The analysis employs the Random Survival Forest (RSF) method, incorporating variable importance and minimal depth techniques to select key variables for predicting overall survival (OS) and disease-free survival (DFS). The study also implements rigorous data handling procedures, including multiple imputations for missing data.

The development of an online prognostic tool based on the RSF model is part of the project, designed to provide real-time survival predictions through a user-friendly interface for clinical application.

Inclusion Criteria:

• Patients aged 75 years or older.
• Histologically confirmed diagnosis of gastric cancer.
• Patients who have undergone curative gastrectomy.

Exclusion Criteria:

• Patients with non-primary gastric malignancies.
• Pathological confirmation of metastatic disease (M1).
• Incomplete follow-up data or loss to follow-up.

The study population comprises elderly patients, aged 75 and above, diagnosed with gastric cancer and who underwent curative gastrectomy within a ten-year period from January 2009 to December 2018. The participants were selected from nine major tertiary care centers across China, ensuring a comprehensive representation of the demographic. The selection criteria were strictly adhered to, ensuring a uniform and high-quality dataset for subsequent analysis in developing a robust prognostic model. The study's focus on the elderly demographic addresses a significant gap in prognostic tools for this age group, which often presents with distinct clinical and therapeutic challenges.