Associations of AI-Derived Coronary CT Features with CFR and FFR-CT in 251 Patients

A retrospective study of 251 patients (753 vessels) with suspected or known coronary artery disease, conducted by Jing Ni and published on figshare in April 2026. The dataset likely contains associations between AI-derived coronary CT angiography features and two ischemia definitions: coronary flow reserve and CT-derived fractional flow reserve.

Use Cases

• Comparing the predictive power of coronary features for CFR-defined versus FFR-CT-defined ischemia based on the reported AUC values.
• Investigating associations between specific plaque features (e.g., perivascular fat attenuation index, calcified plaque burden) and microvascular dysfunction.
• Analyzing distinct plaque feature patterns among discordant CFR/FFR-CT statuses as suggested by the subgroup analysis.
• Training models to predict FFR-CT-defined ischemia using the set of coronary features that achieved an adjusted AUC of 0.892.

Strengths

• Includes data from 251 patients and 753 vessels, providing a multi-vessel perspective.
• Analyzes associations with two distinct physiological measures of ischemia (CFR and FFR-CT).
• Uses generalized estimating equations (GEE) for statistical analysis, accounting for clustered data.

Limitations

• The dataset is very small at 46.0 KB, indicating limited scope.
• Column-level documentation is absent; field semantics must be inferred after download.
• The study is retrospective and, as noted in the conclusions, findings are hypothesis-generating without invasive physiological confirmation.

Provenance

Source

figshare

Collection Method

Retrospective study of patients who underwent CZT-SPECT and concurrent coronary CT angiography.

Freshness

Last updated 2026-04-21 04:20:12; freshness should be verified.