Machine Learning Model Performance Review for Aortic Dissection

Forty studies on aortic dissection risk prediction were systematically reviewed and meta-analyzed. The analysis synthesized model performance metrics, reporting a pooled C-statistic of 0.891 for early mortality prediction. Yijun Mao authored this review, which was registered with PROSPERO and followed PRISMA guidelines.

Use Cases

• Analyze the distribution of pooled C-statistics for outcomes like early mortality, long-term mortality, and acute kidney injury across 40 studies.
• Assess the reported risk of bias using PROBAST criteria and adherence to TRIPOD reporting standards (average 78.7%).
• Examine the frequency of specific ML algorithm usage, such as random forest, SVM, and XGBoost, within the included research.
• Investigate the prevalence of methodological limitations like inadequate events-per-variable ratios and lack of external validation (only 5 of 40 studies).

Strengths

• Includes 40 studies sourced from six major databases, providing a broad evidence base.
• Reports specific performance metrics, such as a pooled C-statistic of 0.891 for early mortality prediction.
• Conducted a detailed risk of bias assessment using PROBAST, finding 100% of studies at high or unclear risk.
• Adhered to systematic review standards (PRISMA, CHARMS) and was registered with PROSPERO.

Limitations

• Significant heterogeneity was observed in the meta-analysis (I² = 61.3–78.8%), limiting the certainty of pooled estimates.
• Critical reporting shortcomings exist, with 82.5% of studies not justifying sample size and 72.5% not providing full model specifications.
• The dataset is a 13.6 KB document summarizing findings, not the raw study data, limiting direct analytical utility.

Provenance

Source

figshare

Collection Method

Systematic review and meta-analysis of studies from PubMed, Web of Science, Cochrane Library, Embase, CNKI, and Wanfang.

Time Range

Studies from database inception to September 30, 2025.

Freshness

Last updated March 26, 2026.

Geography

null