Gastric cancer survival prediction using artificial intelligence models based on electronic health records: a systematic review and meta-analysis

Category	Detail
Condition	Gastric Cancer
Key Mechanisms	Artificial intelligence models utilizing electronic health record data for survival prediction.
Target Population	Adults aged 18 and older with histologically verified gastric cancer undergoing surgical intervention.
Care Setting	Oncology

AI-driven models show a statistically significant improvement in 5-year overall survival prediction compared to traditional methods.
Pooled mean AUC improvement of 0.04 (95% CI 0.02–0.07; p = 0.001) for machine learning models.
Boosting techniques outperform bagging strategies in predictive efficacy.
Key prognostic factors include age, T stage, tumor size, serum albumin/prealbumin levels, and metastatic-to-examined lymph node ratio.
The nature of clinical input data affects algorithm performance.

Utilize AI models for predicting 5-year overall survival in gastric cancer patients.

Incorporate routinely accessible clinical information from EHRs for prognostic assessment.

Evaluate the performance of AI-driven models regularly to ensure accuracy in predictions.

Adults with histologically confirmed gastric cancer undergoing curative surgical procedures.

AI models can enhance the accuracy of survival forecasts, aiding in personalized treatment decisions.

Select AI algorithms based on the organization and nature of input data.
Focus future studies on identifying data characteristics with significant predictive value.

Clinical Scorecard: Predicting Survival in Gastric Cancer Patients Using AI Models Derived from Electronic Health Records: A Systematic Review and Meta-Analysis