August 2, 2025

Figure 4 from Diagnostic Risk Prediction Models for Upper Gastrointestinal Cancers: A Systematic Review

Key Points

MAIN FINDING: Various diagnostic risk prediction models for upper gastrointestinal cancers showed differing AUCs in external validations.
KEY EVIDENCE: AUC scores varied significantly across different model development methods, including conventional statistics and machine learning.
APPROACH: Data was aggregated from multiple studies, visually represented in dot plots to differentiate cancer sites and model types.
SIGNIFICANCE: Understanding AUC variability helps improve diagnostic accuracy for gastrointestinal cancers, which is essential for patient outcomes.

Abstract

AUCs reported in external validation studies. Dot plot showing the AUC reported by each external validation study, separated by study setting. Each data point represents a single model. Data points are colored according to UGI cancer site (red, gallbladder; green, gastric; blue, esophageal; purple, pancreatic). Data point shaped corresponds to the model development method (circle, conventional statistics; triangle, ML).

Figure 4 from Diagnostic Risk Prediction Models for Upper Gastrointestinal Cancers: A Systematic Review

Key Points

Abstract

Cite This Study

Also Consider

Also Consider