AI-Enabled Fractional Flow Reserve Derived From Coronary CT Angiography: Real-World Impact on Downstream Care and Clinical Outcomes

Purpose: To evaluate utilization patterns and clinical outcomes associated with AI–enabled, machine learning driven fractional flow reserve derived from coronary CT angiography (FFR-CT) compared with coronary computed tomography angiography (CCTA) alone.

Methods/Materials: We performed a retrospective multicenter cohort study using the TriNetX federated electronic health record network (112 U.S. HCOs). Adults undergoing CCTA were stratified by receipt of adjunctive FFR-CT (CPT 75580 within 7 days) versus CCTA alone, excluding patients with recent myocardial infarction (MI), angina, or coronary revascularization. Rigorous 1:1 propensity score matching balanced demographics and clinical covariates. Analyses were stratified by baseline renal function (eGFR ≥60 vs ≤59 mL/min/1.73 m²). Outcomes included invasive coronary angiography, revascularization, MI, acute kidney injury (AKI), bleeding (downstream clinically significant hemorrhage identified by diagnosis codes), and mortality.

Results: Among 403,874 eligible patients, stratification and matching yielded 13,559 patients per cohort with normal-eGFR. FFR-CT was associated with higher rates of invasive coronary angiography at 30 days (RD 10.6%, 95% CI 9.8 to 11.4; p< 0.001) and 90 days (RD 13.5%; p< 0.001) compared to CCTA alone. Revascularization remained selective, with higher 90-day rates for FFR-CT (RD 6.6%; p< 0.001). At 30 days, rates of MI, AKI, and bleeding did not differ in the normal-eGFR stratum (all p>0.05). In contrast, at 30 days, among patients with reduced-eGFR (5,645 per cohort), FFR-CT was associated with lower rates of bleeding (RD −0.5%; p< 0.001) and AKI (RD −1.0%; p< 0.001). 1-year mortality was lower for FFR-CT in both normal-eGFR (RD −0.6%, 95% CI −0.9 to −0.3; p< 0.001) and reduced-eGFR patients (RD −1.0%, 95% CI −1.6 to −0.4; p=0.001).

Conclusions: Integration of AI-enabled FFR-CT into CCTA workflows was associated with altered downstream care pathways, characterized by increased invasive angiography with selective revascularization. Importantly, this was associated with improved clinical outcomes across renal function strata, supporting FFR-CT as an effective physiologic gatekeeper that enhances decision-making in real-world practice.