Integrated diagnostic and prognostic value of CT-FFR for coronary in-stent restenosis: a morphology-function analysis

Background and purpose: In-stent restenosis (ISR) remains an important limitation of percutaneous coronary intervention. Computed tomography-derived fractional flow reserve (CT-FFR) combines anatomical and functional assessment, but its diagnostic and prognostic value in patients with suspected ISR is not well established. We evaluated the diagnostic performance of a computational fluid dynamics (CFD)-based CT-FFR method for detecting ISR and its incremental value, beyond coronary computed tomography angiography (CCTA), for predicting major adverse cardiovascular events (MACE). Methods: This single-center retrospective study included 2 analytical cohorts. Cohort A comprised 105 patients (122 stents) with suspected ISR who underwent CCTA and invasive coronary angiography (ICA). Cohort B comprised 86 patients with complete 18-month follow-up for MACE. Diagnostic performance was evaluated using receiver operating characteristic (ROC) analysis. Cox proportional hazards regression was used to identify factors associated with MACE, and prognostic discrimination was assessed using time-dependent ROC and Kaplan-Meier analyses. Results: In cohort A, CT-FFR ≤ 0.80 identified angiographic ISR with an area under the ROC curve (AUC) of 0.909. The combined CCTA and CT-FFR model had greater diagnostic discrimination than CCTA alone (AUC: 0.951 vs. 0.830; P < 0.001). In cohort B, 24 patients (27.9%) experienced MACE. CT-FFR ≤ 0.80 was independently associated with MACE (hazard ratio: 6.24, 95% confidence interval: 1.83–21.26; P = 0.003). At 18 months, the combined model had an AUC of 0.882 and outperformed CCTA alone (P = 0.003) and CT-FFR alone (P = 0.026). MACE-free survival differed significantly between the model-defined risk groups (log-rank P < 0.001). Conclusions: CFD-based CT-FFR showed good diagnostic discrimination for angiographic ISR and was associated with 18-month clinical outcomes. Combining CT-FFR with CCTA improved diagnostic and prognostic discrimination; however, these findings require external validation.