Meta-Analysis of Machine Learning Models for Aortic Dissection Risk

Yijun Mao's systematic review synthesizes performance metrics from 40 studies on machine learning models for predicting adverse outcomes in aortic dissection. The meta-analysis reports pooled C-statistics, such as 0.891 for early mortality prediction, and assesses study quality using PROBAST and TRIPOD guidelines.

Use Cases

• Analyze the distribution of pooled C-statistics (AUC) across different adverse outcomes like early mortality, long-term mortality, and acute kidney injury.
• Assess methodological quality by examining the PROBAST risk-of-bias assessment results reported for the 40 included studies.
• Compare the performance of specific ML algorithms, such as random forest, SVM, and XGBoost, as documented in the review.
• Evaluate reporting completeness using the average TRIPOD adherence score of 78.7% and its breakdown across reporting domains.

Strengths

• Synthesizes data from 40 distinct studies on a critical medical topic.
• Reports specific performance metrics, including a pooled C-statistic of 0.891 for early mortality prediction.
• Follows established PRISMA, CHARMS, and TRIPOD guidelines for systematic review methodology.
• Provides detailed subgroup analysis results, such as heterogeneity measured by I² values ranging from 61.3% to 78.8%.

Limitations

• The underlying primary studies have significant methodological limitations, with 100% assessed at high or unclear risk of bias.
• Report is a 19.5 KB document summarizing findings, not the raw dataset of primary studies.
• Findings are based on a systematic review; the original model data, features, and patient-level records are not included.

Provenance

Source

Yijun Mao via figshare

Collection Method

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

Time Range

Studies from database inception to September 30, 2025.

Freshness

null

Geography

null