Coronary CT angiography-derived quantitative markers for predicting in-stent restenosis

Abstract

ObjectiveTo evaluate quantitative markers derived from coronary CT angiography (coronary CTA) performed prior to percutaneous coronary intervention (PCI) with stent placement for predicting in-stent restenosis (ISR) as defined by quantitative coronary angiography (QCA). Materials and methodsWe retrospectively analyzed the data of 74 patients (60 ± 12 years, 72% male) who had undergone dual-source coronary CTA within 3 months prior to a PCI procedure that included stent placement. Quantitative markers of the target vessel were derived from coronary CTA: Total plaque volume (TPV), calcified and non-calcified plaque volumes (CPV and NCPV), plaque burden (PB in %), remodeling index (RI), and lesion length (LL). Marker performance for predicting ISR, as defined by QCA at follow-up, was assessed. ResultsTwenty-one of 74 stented lesions showed ISR on follow-up (mean 616 ± 447 days). When comparing stent length and LL in patients with ISR, a trend towards less complete stent coverage of the target lesion was observed in cases with ISR (17/21 vs. 4/53 cases, p = 0.07). In multivariate analysis (corrected for dyslipidemia), the following markers showed predictive value for ISR (odds ratio [OR]): NCPV (OR 1.08, p = 0.045), LL (OR 1.38, p = 0.0024), and RI (OR 1.13, p = 0.0019). Sensitivity and specificity for ISR were: NCPV 65% and 80%, LL 74% and 74%, and RI 71% and 78%. At receiver-operating characteristics analysis, NCPV (0.72, p = 0.001), LL (0.77, p < 0.0001), and RI (0.79, p < 0.0001) showed discriminatory power for predicting ISR. A combination of these markers showed incremental predictive value (AUC 0.89, p < 0.0001) with sensitivity and specificity of 90% and 84%, respectively. ConclusionCoronary CTA-derived NCPV, LL, and RI portend predictive value for ISR with incremental predictive value when combining these parameters.