Integrated CT-derived Fractional Flow Reserve and Perivascular Fat Attenuation Index: A Multimodal Approach to Predict In-Stent Restenosis

Wentao Zhao¹Ling Huang¹Mingyuan Zhu¹Xingchao Li¹Xiaojing Liu¹Wanming Zhang²Chunchun Shao³Yixin Li^4*

• ¹Department of Radiology, The Second Hospital of Shandong University, Jinan, China
• ²Catheterization Laboratory, The Second Hospital of Shandong University, Jinan, China
• ³Department of Medical Administration, The Second Hospital of Shandong University, Jinan, China
• ⁴Health Management Center, The Second Hospital of Shandong University, Jinan, China

Background: A significant proportion of patients develop in-stent restenosis (ISR) following percutaneous coronary intervention (PCI) with stent implantation, adversely impacting long-term outcomes. Pericoronary fat attenuation index (FAI) and coronary computed tomography angiography-derived fractional flow reserve (CT-FFR) represent emerging non-invasive imaging biomarkers potentially contributing to ISR pathogenesis through inflammatory and hemodynamic mechanisms. Objective: To construct and evaluate an integrated multimodal predictive model integrating non-invasive computed tomography angiography-derived fractional flow reserve (CT-FFR) and pericoronary fat attenuation index (FAI) for assessing drug-eluting stent (DES)-associated ISR. Methods: This retrospective cohort study enrolled 144 patients (225 coronary lesions) undergoing coronary CT angiography (CCTA) between 2020 and 2024, followed by PCI within one month and subsequent angiographic follow-up (either invasive coronary angiography or CCTA). Computed tomography angiography-derived fractional flow reserve (CT-FFR) values for target lesions were reconstructed using deep learning algorithms. Pericoronary FAI was quantified at the stented coronary segment. The primary endpoint was angiographically confirmed in-stent restenosis (ISR, defined as luminal stenosis ≥50%) during follow-up. A hybrid-effects logistic regression model generated the integrated CT-FAI-FFR score. Predictive efficacy was evaluated using receiver operating characteristic (ROC) curve analysis and decision curve analysis. Results: During a median follow-up duration of 25.5 months, ISR occurred in 55 lesions (24.4%). The study cohort was 67.0% male. the multimodal CT-FAI-FFR model achieved an area under the ROC curve (AUC) of 0.793 for predicting ISR. This predictive performance significantly surpassed that of the baseline alone (AUC 0.656, P =0.009). Conclusion: The CT-FAI-FFR score optimizes ISR risk stratification by integrating complementary information reflecting hemodynamic impairment and local vascular inflammation. This finding suggests a potential pathophysiological interplay between perivascular inflammation and hemodynamic impairment in ISR development.