Objective:

To assess the efficacy of artificial intelligence models utilizing routinely gathered electronic health record (EHR) data for forecasting 5-year overall survival (5-OS) in individuals receiving surgical intervention for gastric cancer.

Approach:

Key Findings:

• Ten retrospective investigations involving 15,643 patients were analyzed.
• Machine learning models showed a statistically significant improvement in predictive efficacy over traditional methods with a pooled mean AUC improvement of 0.04.
• Boosting techniques outperformed bagging strategies with an AUC improvement of 0.02.
• Clinical input data, particularly blood-derived biomarkers, influenced algorithm performance.
• Common prognostic factors included age, T stage, tumor size, serum albumin or prealbumin levels, and metastatic-to-examined lymph node ratio.

Interpretation:

AI-driven prognostic models utilizing routinely accessible clinical information enhance the accuracy of 5-year survival forecasts following gastric cancer surgery.

Limitations:

• Studies predominantly utilizing non-standard multimodal data were excluded.
• Heterogeneity among studies may affect the generalizability of findings.

Conclusion:

The choice of the most suitable AI algorithm should be informed by the organization and nature of the input data to optimize predictive efficacy and operational utility in clinical decision support frameworks.

Sources:

• Systematic Review Registration
Attribution Notice

This content is an AI-generated, fully rewritten summary based on a published scholarly article. It does not reproduce the original text and is not a substitute for the original publication. Readers are encouraged to consult the source for full context, data, and methodology.