Aortic Dissection Risk Prediction ML Models Meta-Analysis

A systematic review and meta-analysis of 40 studies evaluating machine learning models for predicting adverse outcomes in aortic dissection. The analysis, conducted by Yijun Mao, synthesized performance metrics like C-statistics (AUC) for outcomes including early mortality, long-term mortality, and acute kidney injury.

Use Cases

• Analyze the pooled C-statistic performance for early mortality (0.891) and long-term mortality (0.847) across 40 included studies.
• Compare the discriminative performance of ML algorithms like random forest, SVM, and XGBoost for predicting acute kidney injury (AUC 0.825).
• Assess the reported adherence to TRIPOD standards (average 78.7%) and the prevalence of high/unclear PROBAST risk of bias (100% of studies).
• Evaluate the frequency of external validation use (only 5 of 40 studies) and calibration assessment omission (27.5% of models).

Strengths

• Synthesizes data from 40 distinct studies on a critical medical topic.
• Reports pooled C-statistics with confidence intervals for key outcomes like early mortality (0.891, 95% CI: 0.854–0.927).
• Follows established PRISMA, CHARMS, and TRIPOD guidelines and was registered with PROSPERO.
• Includes detailed methodological assessment using PROBAST and reports on TRIPOD adherence (average 78.7%).

Limitations

• The underlying studies show significant heterogeneity (I² = 61.3–78.8%) and 100% were at high or unclear risk of bias.
• Critical reporting shortcomings exist in the original studies, such as unreported predictor definitions (62.5%) and sample size justifications (82.5%).
• The dataset is a 13.8 KB DOCX file containing textual analysis, not a structured tabular dataset for direct ML training.

Provenance

Source

figshare, authored by Yijun Mao.

Collection Method

Systematic review and meta-analysis of six major databases (PubMed, Web of Science, Cochrane Library, Embase, CNKI, Wanfang).

Time Range

Literature search from database inception to September 30, —.

Freshness

Last updated March 26, 2026, with literature search up to September 30, —.

Geography

null