Coronary Computed Tomographic Angiography-Derived Fractional Flow Reserve Based on Machine Learning for Risk Stratification of Non-Culprit Coronary Narrowings in Patients with Acute Coronary Syndrome.

This study investigated the prognostic value of coronary computed tomography angiography (cCTA)-derived fractional flow reserve (CT-FFR) in patients with acute coronary syndrome (ACS) and multivessel disease to gauge significance and guide management of non-culprit lesions. We retrospectively analyzed data of 48 patients (56 ± 10 years, 60% men) who were admitted for symptoms suggestive of ACS and underwent dual-source cCTA followed by invasive coronary angiography with culprit lesion intervention. Culprit lesions were retrospectively identified on cCTA using images obtained during invasive coronary angiography. Non-culprit lesions with ≥25% luminal stenosis and deferred intervention were evaluated using a machine learning CT-FFR algorithm to determine lesion-specific ischemia (CT-FFR ≤0.80). Follow-up was performed. CT-FFR identified lesion-specific ischemia in 23 of 81 non-culprit lesions. After a median follow-up of 19.5 months, 14 patients (29%) had major adverse cardiac events (MACE). Univariate Cox regression analysis revealed that CT-FFR ≤0.80 (hazard ratio [HR] 3.77 [95% confidence interval 1.16 to 12.29], p = 0.027), Framingham risk score (FRS) (HR 2.96 [1.01 to 7.63], p = 0.038), and a CAD-RADS classification ≥3 (HR 3.12 [1.03 to 10.17], p = 0.051) were predictors of MACE. In a risk-adjusted model controlling for FRS and CAD-RADS ≥3, CT-FFR ≤0.80 remained a predictor of MACE (1.56 [1.01 to 2.83], p = 0.048). Receiver operating characteristics analysis including FRS, CAD-RADS ≥ 3, and CT-FFR ≤0.80 (area under the curve 0.78) showed incremental discriminatory power over FRS alone (area under the curve 0.66, p = 0.032). CT-FFR of non-culprit lesions in patients with ACS and multivessel disease adds prognostic value to identify risk of future MACE.