Post‐PCI quantitative flow ratio predicts 3‐year outcome after rotational atherectomy in patients with heavily calcified lesions

Abstract Background The study sought to investigate the clinical predictive value of quantitative flow ratio (QFR) for the long‐term outcome in patients with heavily calcified lesions who underwent percutaneous coronary intervention (PCI) following rotational atherectomy (RA). Methods In this retrospective study, 393 consecutive patients from 2009 to 2017 were enrolled. The QFR of the entire target vessel (QFRv) and the QFR of the stent plus 5 mm proximally and distally (in‐segment) (QFRi) were measured. The primary endpoint was target lesion failure (TLF), including target lesion‐cardiac death (TL‐CD), target lesion‐myocardial infarction (TL‐MI), and clinically driven‐target lesion revascularization (CD‐TLR). Results A total of 224 patients with 224 calcified lesions completed the clinical follow‐up, and 52 patients had TLF. There was no significant difference in QFRv post‐PCI between non‐TLF and TLF groups (p > .05). However, QFRi post PCI was significantly higher in the non‐TLF group than in the TLF group. Multivariate Cox regression showed that QFRi post‐PCI was an excellent predictor of TLF after a 3‐year follow‐up (HR 1.7E−8 [5.3E−11–5.6E‐6]; p < .01). Furthermore, receiver‐operating characteristic curve analysis demonstrated that the optimal cutoff value of QFRi for predicting the long‐term TLF was 0.94 (area under the curve: 0.826, 95% confidence interval: 0.756–0.895; sensitivity: 89.5%, specificity: 69.2%; p < .01). The QFRi ≤ 0.94 post‐PCI was negatively associated with TLF, including TL‐CD, TL‐MI, and CD‐TLR (p < .01). Conclusions QFRi post‐PCI showed a high predictive value for TLF for during a 3‐year follow‐up in patients who underwent PCI following RA; specifically, lower QFRi values post‐PCI were associated with worse TLF.


| INTRODUCTION
Rotational atherectomy (RA) technique was first introduced more than 30 years ago and used to reduce plaque burden by the debulking idea before the stent era and by the modification idea in the current stage.
However, both strategies failed to be proven superior to only balloon angioplasty, bare-metal stent (BMS) implantation, or drug-eluting stent (DES) implantation for coronary calcified lesions in a number of previous studies. [1][2][3][4][5] At present, the significant indications of RA in daily practice are limited to heavily calcified lesions detected by coronary angiography (CAG) or intravascular imaging (IVI) as a bailout strategy. 6 A previous study has shown that the final minimal luminal diameter (MLD) post-BMS implantation following RA was the only significant independent predictor of event-free survival. 7 Even with DES used in the current stage for complex coronary lesions with/without the use of RA, nearly one-third of patients experienced major adverse cardiac events (MACE, defined as the composite of death, myocardial infarction, and target vessel revascularization [TVR]) at 2-year follow-up. 8 A question arises as to why percutaneous coronary intervention (PCI) of heavily calcified lesions can cause poor outcomes, we hypothesize that not only because of RA itself (which can cause thermal injury and additional vessel trauma and decrease the efficacy of DES in reducing neointimal growth) but unsatisfactory lesion preparation could also influence the final stent result. A physiological index such as fractional flow reserve (FFR) post-PCI without RA has a good predictive value for the late outcome of target vessel failure (TVF, defined as the composite of cardiac death, target vessel-related myocardial infarction, and clinically driven TVR). 9 However, previous studies have not focused on the predictive value of physiological indexes post-PCI on clinical outcome after the PCI procedure using RA due to complex manipulation of pressure wire.
Quantitative flow ratio (QFR), a novel physiological index derived from three-dimensional (3-D) angiographic analysis positively correlates with traditional invasive FFR, as shown in many studies. [10][11][12][13] The main advantage is that QFR measurement does not require hyperemia and a pressure wire. The online or offline analysis permits physiological guiding of PCI and retrospective physiological functional studies if the angiographic quality is satisfied. [10][11][12][13] Therefore, we designed a retrospective study to determine the predictive value of QFR post-PCI for the target lesion failure (TLF) in the long-term follow-up. Currently, the QFR research has mainly focused on the vessel QFR, and there have been no studies on the stent QFR. [10][11][12][13][14] Generally, the definition of the stent segment includes the stent segment and the 5-mm area from its proximal and distal ends. 15 In this study, we aimed to compare the predictive value of TLF between the vessel QFR and the stent QFR post-PCI in patients with heavily calcified lesions using the RA technique.

| Study design
This was a retrospective cohort study. The study protocol was approved by the ethics committees of the four participating centers

| Study population
We included patients treated with RA for target lesions with the indications of heavily calcified lesions (detected by angiography or IVI), or uncrossable or undilated calcified lesions. The exclusion criteria were as follows: severe complications of RA, such as perforation, post-PCI slow flow or no flow (defined as coronary thrombolysis in myocardial infarction flow grade <3 or =0); culprit lesions that were not de novo lesions, such as in-stent restenosis (ISR); PCI with drug-coated balloon(s) or bioabsorbable scaffold implantation post-RA; previous target vessel PCI; and a life expectancy < 12 months.

| PCI procedure
All of the patients were administered 300 mg clopidogrel as a loading dose before CAG. Unfractionated heparin (100-120 μ/kg) was administered by bolus injection via sheath to maintain activated clotting time (ACT) > 300 s during the whole procedure. Standard selective CAG was performed via a radial approach with 6-French catheters without a side hole in accordance with the routine practice.

| Intravascular ultrasound (IVUS) procedure
After intracoronary injection of nitroglycerin (100-200 mg), an IVUS catheter was pushed at least 10 mm distal to the lesion or stent edge. IVUS images were obtained through the automatic pullback (0.5 mm/s) by a commercially available imaging system with a 40-MHz mechanical transducer (Boston Scientific) for measuring on-site. All of the IVUS images were stored on a DVD for offline measurements. Minimal stent diameter, maximal stent diameter, and minimal stent area were measured, and the stent eccentricity index was calculated as minimal stent diameter/maximal stent diameter.

| Quantitative coronary angiography (QCA) and QFR measurements
Two high-quality coronary angiographic projections at least 25°apart post-PCI are required to satisfy the requirements of QCA and QFR analyses, which were performed offline by two experienced technicians (with a high interrater agreement in all cases [к > 0.90]) at Nanjing Heart Center (core lab) using AngioPlus software (Pulse Medical Imaging Technology) as previously described. 10,14 QCA data included pre-and post-PCI proximal and distal reference vessel diameter (RVD), total stent length (TSL), and in-stent MLD immediately post-PCI. In the present study, we considered traditional QFR of the entire target vessel starting from the most proximal available segment until the distal part where the vessel diameter was ≥ 1.5 mm (vessel QFR [QFRv]), and the QFR in a segment was measured from 5 mm proximal to 5 mm distal to the stent edge (stent QFR [QFRi]) for subsequent analysis. 11,14,15 Figure S1 shows how QFRv and QFRi post-PCI were measured on a representative case. The figure was generated by AngioPlus QFR 1.0 software (Pulse Medical Imaging Technology, Shanghai Co., Ltd.) as previously described. 14  noninvasive functional stress test. 17,18 All of the events were judged by an independent clinical event committee that was blinded to the PCI procedure and QFR and QCA data.

| Statistical analysis
According to the ROTAXUS trial, 8 the MACE rate was 29.4% in patients with heavily calcified lesions after the RA procedure at a 2-year follow-up. The predictor had a standard deviation of 0.05, and the hazard ratio (HR) was set at 1.0. We tested the hypothesis using a 5% significance level with a two-sided Wald test. As a result, the sample size of 224 was calculated with 1.00 power by PASS11.0 software (NCSS, LLC).
Categorical variables were expressed as counts with percentages, whereas continuous variables were expressed as mean with standard deviation (SD) or median with IQRs. Categorical variables were compared using the χ 2 test. The Kolmogorov-Smirnov test was used to assess the distributions of continuous variables. Continuous variables were expressed as mean ± SD for normally distributed data The study flowchart. PCI, percutaneous coronary intervention; QFRv, vessel quantitative flow ratio; QFRi, quantitative flow ratio in a segment; RA, rotational atherectomy; TLF, target lesion failure; TL-CD, target lesion-cardiac death; TL-MI, target lesion-myocardial infarction; TLR, target lesion revascularization and were compared using the Student's t test. Data that were not normally distributed were expressed as medians and were compared using the Mann-Whitney U test. The Kaplan-Meier method was used to derive the event rates at the follow-up and to plot time-toevent curves, which were then compared by the log-rank test. To study TLF predictors, a univariate Cox regression was performed.
Variables that were found to be significant were entered into a multivariate model. Their outputs included HR, 95% confidence interval (CI), and p value. The receiver operating characteristic curve was used to compare the variables' predictive ability of the rates of TLF. All of the statistical tests were two-tailed. Statistical significance was set at .05. For the statistical analysis, the SPSS version 24.0 (SPSS Institute Inc.) was used.  (Table 1).
However, the target vessel location in the TLF group was significantly different from that in the non-TLF group (p < .05). Post-dilated pressure and post-PCI DS of the TLF group were notably higher than those in the non-TLF group, and post-PCI MLD and QFRi post-PCI of the TLF group were markedly lower (p < .05 or p < .01) ( Table 1). Furthermore, we analyzed the minimal stent diameter, maximal stent diameter, minimal stent area, and stent eccentricity index in patients with heavily calcified lesions immediately after PCI. The minimal stent diameter, minimal stent area, and stent eccentricity index in the non-TLF group were larger than those in the TLF group (p < .05). However, there was no significant difference in maximal stent diameter between the two groups (p > .05) (Table S1). At 1-year follow-up, there were no differences in proximal RVD and distal RVD between the TLF group and the non-TLF group (p > .05). However, MLD in the TLF group was smaller than that in the non-TLF group, and DS in the TLF group was larger (p < .05) (Table S2).
These results indicated that TLF occurrence was closely related to post-dilated pressure, MLD, DS, and QFRi post-PCI, and MLD and DS 1 year after PCI in patients with heavily calcified lesions after RA.
Moreover, lower minimal lumen diameter, minimal stent area, and eccentricity index were associated with a failure rate of the target lesion.

| Cox regression analysis of factors associated with TLF and their predictive value analyzed by ROC curve in patients with heavily calcified lesions after RA at 3-year follow-up
To further analyze factors associated with TLF in patients with heavily

| Clinical outcome in patients with heavily calcified lesions after RA at the 3-year follow-up
According to the QFRi post-PCI cutoff value of 0.94, we divided these patients into high-and low-QFR groups. The incidence rate of TLF, TL-CD, TL-MI, and TLR in the high-QFR group were significantly lower than those in the low-QFR group (p < .05 or p < .01) ( Table 3).
TLF and its compositions analyzed using the Kaplan-Meier curves are shown in Figure 3. The TLF ratio in the low-QFRi group was significantly higher than in the high-QFRi group (   PCI is a predictor of in-stent restenosis and stent thrombosis. 25 Therefore, in the present study, DS post PCI in the TLF group was notably high compared with that in the non-TLF group, and the burr-to-vessel ratio was likely to decrease in the TLF group. Additionally, the postdilated pressure of the balloon in the TLF group was higher than that in the non-TLF group, suggesting that the lesion preparation was insufficient in the

| Limitations
This study had some limitations. First, our study was a retrospective study in which the imaging use ratio was about 50%, and <60% of the cases had clinical follow-up, and 50% of the cases had the angiographic follow-up. The reasons for this situation were the increased financial burden and unconventional clinical and CAG follow-up. Second, this study did not provide post-PCI FFR values.
Finally, there was no established methodology for determining the QFR values of the stented segment.

| CONCLUSIONS
QFRi post-PCI showed a high predictive value for the long-term clinical outcome in patients who underwent RA during the PCI procedure. Besides, the lower QFRi post-PCI was associated with higher TLF. QFRi could be applied for evaluating the coronary stenting outcome in patients who underwent RA during the complex PCI.