Development and validation of machine-learning-based survival prediction model for young patients with gastric cancer

Abstract

Background Despite the global decline in the incidence of gastric cancer, the number of young individuals diagnosed with it continues to rise. Several studies have been conducted to predict the mortality of patients with gastric cancer; however, they employ traditional methodologies and have limitations. Therefore, we propose short-, medium-, and long-term mortality prediction models for young patients with gastric cancer based on a survival machine learning model. Methods Data of 1,200 young (< 50 years) patients diagnosed with gastric cancer between 2013–2015 were obtained from the Gastric Cancer Public Staging Database. Data of 840 and 360 patients were used for training and testing, respectively. We employed the random survival forest (RSF), gradient boosting survival analysis (GBSA), and extra survival tree (EST) prediction models for 1-, 3-, and 5-year survival prediction, and the concordance index (C-index) metric to objectively assess the models. This study also examined the key determinants of mortality based on the prediction time points. Results The results indicate that the EST model (1-year mortality: 97.08 ± 0.01, 3-year mortality: 96.19 ± 0.01, 5-year mortality: 93.68 ± 1) exhibited a slightly better performance than the GBSA (1-year mortality: 96.91 ± 0.01, 3-year mortality: 94.91 ± 0.01, 5-year mortality: 93.57 ± 0.01) and RSF (1-year mortality: 96.67 ± 0.01, 3-year mortality: 95.65 ± 0.01, 5-year mortality: 92.82 ± 0.01) models. Tumour stage and size were the primary variables employed for training the models to predict mortality at different time points. The other variables exhibited varying degrees of consistency for each time point. Conclusions The findings are expected to facilitate the identification of high-risk young patients with gastric cancer who may benefit from aggressive treatment by predicting their risk of death at various time points.