Key result
Machine learning-based CT-FFR demonstrated a per-patient diagnostic accuracy of 93% and a negative predictive value of 97% for identifying significant coronary artery disease compared to invasive coronary angiography.
Why the study?
Pre-TAVI CAD assessment relies on ICA, which carries peri-procedural risks, making noninvasive exclusion of significant CAD using ML-based CT-FFR desirable to potentially avoid ICA.
Does machine learning-based CT-FFR accurately identify significant CAD compared to invasive coronary angiography in patients undergoing evaluation for TAVI?
Population
41 TAVI candidates with both TAVI-planning CT and ICA
Comparison
ML-based CT-FFR vs ICA
Design
Single-center retrospective study
Authors
Loading...
May support ML CT-FFR to exclude CAD pre-TAVI; hypothesis-generating, should not yet change practice.
Observational (n=41)
No
Does machine learning-based CT-FFR accurately identify significant CAD compared to invasive coronary angiography in patients undergoing evaluation for TAVI?
Integration of ML-based CT-FFR into routine TAVI-planning CT protocols provides accurate screening for significant CAD, potentially avoiding invasive angiography in up to 80% of patients.
Leung et al. (2025) conducted an observational in Severe symptomatic aortic stenosis (n=41). Machine learning-based computed tomography-derived fractional flow reserve (ML-based CT-FFR) vs. Invasive coronary angiography (ICA) was evaluated on Per-patient diagnostic accuracy for significant coronary artery disease (95% CI 77-97). Machine learning-based CT-FFR demonstrated a per-patient diagnostic accuracy of 93% and a negative predictive value of 97% for identifying significant coronary artery disease compared to invasive coronary angiography.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: