Meta-Analysis of Deep Learning Models for Aortic Dissection Segmentation and Diagnosis

48 studies on deep learning models for aortic dissection segmentation or diagnostic tasks were included in this systematic review and meta-analysis. The document, authored by Yichen Zhao and last updated in April 2026, reports performance metrics such as a mean Dice coefficient of 89.2% for false lumen segmentation and pooled sensitivity of 0.94 for CT-based diagnosis.

Use Cases

• Benchmarking new segmentation models against reported mean Dice coefficients for false lumen, true lumen, and entire aorta.
• Evaluating diagnostic AI performance using the pooled sensitivity and specificity metrics for CT, CTA, and ECG-based models.
• Comparing AI diagnostic performance to clinicians using the reported pooled sensitivity (0.79) and specificity (0.95).

Strengths

• Includes results from 48 studies, with 28 focused on segmentation and 20 on diagnostic tasks.
• Reports specific performance metrics, including mean Dice coefficients with standard deviations and pooled sensitivity/specificity with 95% confidence intervals.
• Registered as a systematic review (CRD42024619403) and follows a defined methodology using QUADAS-2 for risk evaluation.

Limitations

• The dataset is a 28.9 KB DOCX file containing a review article, not the underlying image or model data.
• Row count and column-level documentation for any underlying data are absent.
• The description does not specify the geographic or institutional origin of the included studies, which may reflect source bias.

Provenance

Source

figshare

Collection Method

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

Time Range

Studies identified up to November 3, 2024.

Freshness

Last updated 2026-04-14 04:21:49; includes studies up to November 3, 2024.