Abstract
Abstract Background Coronary microvascular disease (CMD) is responsible for nearly half of angina in the general population. The index of microvascular resistance (IMR) is the gold standard for the diagnosis of CMD. Abnormal IMR is independently associated with adverse outcomes and persistent angina after percutaneous coronary intervention (PCI) regardless of the achieved fractional flow reserve (FFR). However, the implementation of IMR in clinical practice is low due to cost and the need for hyperemia. Thus, multiple software and formulas have been developed to calculate wireless IMR via angiography analysis. However, these angiography-derived IMRs have never been compared head-to-head by an independent core laboratory. Aim In this study, we aim to investigate the diagnostic accuracies of five angiography-derived IMR software/formulas by an independent core lab in a retrospective cohort undergoing follow-up angiograms after successful PCI. Methods We analyzed 43 patients from the FUNCOMBO trial and 34 patients from the RETROFI II trial. All patients have received successful PCI for STEMI, and wire-based IMR was measured at six months (for FUNCOMBO) and three years follow-up (for RETROFI II). Angiography-derived IMR was calculated with the QFR®2.1 software, the Flashangio CaIMR® software, and the AngioPlus software. In addition, using QFR-derived data, we calculated the angio-IMR using the formula of Escaned et al. [angio-IMR = (Pa–[0.1*Pa])*QFR*e-Tmn (where e-Tmn is an estimation of hyperaemic mean transit time)] and Scarsini et al. [NH-IMRangio=Pa(resting)x QFR x N frame(resting)/frame per second]. The threshold of CMD diagnosis is 25 for all five angiography-derived IMR methods. The correlation and agreement of the five methods were evaluated against wire-based IMR measurement. Results The wire-based IMR was significantly correlated with Medis IMR (r= 0.396, p=0.001) and NH-IMRangio (r=0.378, p=0.001). The agreements between IMR and the angiography-derived IMRs are demonstrated in figure 1. The sensitivity, specificity, positive predicted value, negative predicted value, and accuracy were demonstrated in figure 2, which showed that using the published criteria, the accuracy of diagnosis was 52%, 68%, 67%, 68%, and 49% for Medis IMR, CaIMR, AngioPlus IMR, angio-IMR, and NH-IMRangio, respectively. Comparison of all angiography-derived IMR with wire-based IMR yielded the AUC’s of 0.648, 0.561, 0.606, 0.674, 0.688 for Medis IMR, CaIMR, AngioPlus IMR, angio-IMR, and NH-IMRangio respectively. (Figure 2.) Conclusion In this small validation cohort, we found that all five angiography-derived IMR calculation methods had either poor or very poor diagnostic accuracy. Among the five methods, the Medis IMR, Angio-IMR, and NH-IMRangio may be viable options for wireless CMD diagnosis.Figure 1Figure 2
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have