Clinical implication of quantitative flow ratio to predict clinical events after drug‐coated balloon angioplasty in patients with in‐stent restenosis

Abstract Background The association between the quantitative flow ratio (QFR) and adverse events after drug‐coated balloon (DCB) angioplasty for in‐stent restenosis (ISR) lesions has not been investigated. Hypothesis Post‐procedural QFR is related to adverse events in patients undergoing DCB angioplasty for ISR lesions. Methods This retrospective study included data from patients undergoing DCB angioplasty for drug‐eluting stent (DES) ISR between January 2016 and February 2019. The QFR was measured at baseline and after DCB angioplasty. The endpoint was the vessel‐oriented composite endpoint (VOCE), defined as a composite of cardiac death, vessel‐related myocardial infarction, and ischemia‐driven target vessel revascularization. Results Overall, 177 patients with 185 DES‐ISR lesions were included. During 1‐year follow‐up, 27 VOCEs occurred in 26 patients. The area under curve (AUC) of the post‐procedural QFR was statistically greater than that of the in‐stent percent diameter stenosis (0.77, 95% confidence interval [CI] 0.67–0.87 vs. 0.64, 95% CI 0.53–0.75; p = .032). Final QFR cutoff of 0.94 has the best predictive accuracy for VOCE. A QFR > 0.94 was associated with a lower risk of VOCE compared to a QFR ≤ 0.94 (log‐rank test, p < .0001). Survival analysis using the multivariable Cox model showed that a post‐procedural QFR ≤ 0.94 was an independent predictor of 1‐year VOCE (hazard ratio 6.53, 95% CI 2.70–15.8, p < .001). Conclusions A lower QFR value was associated with worse clinical outcomes at 1 year after DCB angioplasty for DES‐ISR.


| INTRODUCTION
Although drug-eluting stents (DES) effectively inhibit neointimal proliferation and markedly reduce the incidence of in-stent restenosis (ISR), 1,2 recurrent ISR requiring repeat revascularization still occurs after DES implantation, 3 and treatment of DES-ISR remains a major challenge in the percutaneous coronary intervention (PCI) field. Drugcoated balloons (DCB) are semi-compliant balloons covered with antiproliferative drugs such as lipophilic paclitaxel and have been proposed as an alternative to DES. 4,5 During balloon inflation, the lipophilic paclitaxel is delivered to the vessel wall surface providing an antiproliferative effect and preventing neointimal hyperplasia without additional metallic layers. In this respect, several clinical trials support the efficacy of DCB in the treatment of ISR. [6][7][8] The fractional flow reserve (FFR) is the gold standard to assess the physiological severity of coronary stenosis. 9 Several trials have reported an inverse relationship between the post-interventional FFR and the risk of subsequent adverse events. 10,11 Additionally, the FFR has been identified as a reference standard to ascertain functional ISR severity. 12 Quantitative flow ratio (QFR), a novel technique for the rapid computation of FFR from coronary angiography without the use of pressure wires or hyperemic agents, has good correlation with FFR and proven clinical value in guiding pre-and post-PCI management. [13][14][15] Although the diagnostic performance of the QFR in assessing ISR lesions using FFR as a reference standard has been recently investigated, 16 its utility in the context of DCB angioplasty remains unknown. In the present study, we aimed to perform the first validation of the QFR as a tool to predict events in patients with ISR treated with DCB.

| Study design and population
This was a retrospective analysis of consecutive patients with at least one ISR lesion who underwent DCB angioplasty between January 2016 and February 2019 from two centers in Shanghai, China. Exclusion criteria were a TIMI flow grade <3 at baseline or after DCB angioplasty, ST-segment elevation myocardial infarction (MI), major procedural complications requiring DES implantation, and significant left main ISR lesion or inadequate angiographic image quality limiting the QFR computation. The study protocol complied with the Declaration of Helsinki and was approved by the ethics committee at each participating institution.

| PCI procedures and angiographic characteristics
ISR lesions were treated using paclitaxel-coated balloons (Sequent Please, B Braun Melungeon, Germany). All patients were pretreated with aspirin and a P2Y12 inhibitor (i.e., clopidogrel or ticagrelor). The DCB size was selected based on the length of the target lesion and the diameter of the previously used stents. The details of the PCI strategy and intravascular imaging utilization were left entirely at the operators' discretion. The ISR was classified as focal (type I, <10 mm in length), diffuse (type II, >10 mm in length) and proliferative (type III, >10 mm in length and extending outside the stent). 17 Quantitative coronary angiography analyses were performed before and after DCB angioplasty using an offline computerized quantitative coronary angiographic system (CAAS system; Pie Medical Instruments, Maastricht, The Netherlands). The following parameters were measured: reference vessel diameter (RVD), minimal lumen diameter (MLD), and percent diameter stenosis (%DS). Measurements of the stented area were obtained shoulder-to-shoulder (in-stent) and including the total stented area plus 5 mm proximally and distally (insegment). The acute lumen gain was calculated as the difference between post-and pre-procedural MLD.

| Off-line QFR assessment
Off-line QFR was performed by experienced analysts certified for the use of the QFR system software (AngioPlus, Pulse Medical Imaging Technology, Shanghai Co. Ltd., Shanghai, China). For QFR computations, two angiographic projections, at least 25 apart, were transferred to the QFR system, and three-dimensional reconstruction of the interrogated vessel without its side branches was performed as described elsewhere. 13,18,19 The lumen contour was estimated automatically. Manual correction was performed in cases of suboptimal angiographic image quality based on a standard operation procedure. A contrast flow model incorporating contrast flow velocity based on the frame count method was used for QFR computations. Threedimensional quantitative coronary analysis data were readily available.
Subsequently, the QFR was computed. Case example is provided in

| Follow-up and outcome definition
In the present study, we investigated the relationship between the post-procedural QFR and clinical outcomes at the vessel level. Clinical follow-up data derived from clinical visits, telephone interviews, and from hospital records of any readmissions.
The primary endpoint of this study was the vessel-oriented composite endpoint (VOCE) at 1 year, defined as the composite of cardiac death, vessel-related MI, or ischemia-driven target vessel revascularization (TVR). The secondary endpoints were the individual components of the VOCE.
Any death of unknown cause was assumed as due to cardiac cause. The diagnosis of MI was based on the fourth universal definition of MI, which requires a combination of symptoms, electrocardiographic changes, and significant increase in troponin. 20 When the identification of the culprit vessel was not possible, the endpoint was evaluated considering each vessel treated with DCB. Ischemia-driven TVR was defined as any repeated revascularization of the target vessel by either PCI or coronary artery bypass grafting (CABG) in the presence of a lesion with a %DS >50%, and with at least one of the following: (i) recurrence of angina, (ii) positive non-invasive test, and (iii) positive invasive physiologic test. All angiograms of patients who underwent TVR were reviewed to identify the target lesion revascularization.

| Statistical analysis
Continuous variables are presented as the median with interquartile range and compared between groups using the Mann-Whitney U test. Categorical variables are summarized as frequencies and proportions and compared using Pearson's chisquare or Fisher's exact test. Intraclass correlation coefficient F I G U R E 1 Case example of reconstructed 3-D QCA and measured QFR. QFR calculation was based on the 3D-QCA reconstructed from two angiographic projections with angles ≥25 apart and 3D reconstruction of the interrogated vessel without its side branches was performed. (A) Pre-procedural angiographic image shows a ISR lesion, and QFR was 0.70 (B) Final angiography showed minimal residual stenosis after DCB treatment, and QFR was 0.96. Red arrows indicate the target ISR lesion. DCB, drug-coated balloon; ISR, in-stent restenosis; QCA, quantitative coronary angiography; QFR, quantitative flow ratio

| Patient characteristics
The flow chart of this study was shown in Figure S1. During the study period, a total of 177 patients with 185 DES-ISR lesions undergoing The distribution of individual QFRs before and after DCB angioplasty is reported in Figure S2A. While there was no significant difference between the two groups with respect to the pre-procedural QFR

| Definition of a potential cutoff value
The occurrence of the VOCE stratified according to the post-procedural QFR is depicted in Figure S2B. ROC analysis showed that the optimal post-procedural QFR cut-off value for predicting the VOCE was 0.94, with a sensitivity of 74% and specificity of 75%. The ROC curves for the in-stent %DS and post-procedural QFR are shown in Figure 2 Figure 2(B).

| Clinical outcomes
Lower post-procedural QFR values were significantly associated with a higher incidence of the VOCE (p < .0001).
Variables predicting VOCE obtained after penalization using LASSO method included following three variables: post-procedural QFR ≤0.94; diabetes mellitus; post-procedural in-stent %DS.

| DISCUSSION
The main findings of this study are as follows: (1) a low QFR after DCB angioplasty in DES-ISR lesions was an independent predictor of adverse clinical events during a 1-year follow up; (2) compared with the in-stent %DS, the QFR has a better ability to predict vesselrelated clinical outcomes after DCB angioplasty.
To the best of our knowledge, this is the first study to describe the utility of the QFR to predict clinical events after DCB angioplasty for ISR lesions. Our results showed that the QFR measured immediately after the procedure had an inverse relationship with future clinical events. Notably, the pre-procedure QFR had no association with outcomes. Conceptually, interventions destined to increase the postprocedural QFR may be able to improve long-term outcomes.
Although the choice of the PCI strategy has routinely been based on angiographic findings, these have limited efficacy to predict immediate physiological results or clinical outcomes of coronary stenting. 23 As a result, the concept of functional optimization of PCI results has been explored for a long time. In this regard, a previous study found a graded relationship between post-PCI FFR and major adverse cardiovascular events. 24  angiographic residual %DS has been commonly used in clinical practice, but its limitation is well known. In this study, we compared the prognostic ability of QFR and in-stent stenosis severity after DCB angioplasty on clinical events, and we found that the former had a better performance.
The optimal treatment strategy for DES-ISR remains undefined.
Based on clinical trials that support its efficacy, angioplasty with DCB is recommended for the treatment of ISR in the European clinical practice guidelines (Class I, Level of Evidence: A). 26 Prospective studies have confirmed the utility of the FFR to guide clinical decision making in ISR treatment and suggest that revascularization can be safely deferred in patients with an FFR > 0.75. 27 Recent studies found that FFR-guided DCB treatment of de novo lesions appeared feasible and safe in stable patients. 28 Whether the use of post-procedural physiological assessments could be expanded to angioplasty with DCB for ISR lesions had not been sufficiently studied to date.
Nevertheless, while evidence supporting the utility of FFR has been mounting, FFR is still largely underutilized in clinical practice. 29 Reasons for this may include the high equipment and drug costs, and the risk of related complications. Progress in angiography-derived FFR such as the QFR can reduce these limitations by calculation of functional parameters in a simpler and rapid way. A recent study demonstrated substantial applicability of the QFR in functional assessment of ISR lesions, using FFR as reference standard. 16 Furthermore, Li et al. 30 investigated the functional results following DCB or DES treatment in small-vessel disease and demonstrated that assessing the QFR in small coronary arteries was feasible. Therefore, these studies suggest that the QFR is not only a promising tool in assessing clinical results of stenting but can also be applied to assess the efficacy of DCB in the treatment of ISR and de novo lesions.
Unlike pre-procedural functional assessment, where parameters to determine an ischemia-causing stenosis are clearly defined, postprocedural cutoff values vary widely due to multiple factors, including the population studied, lesion and procedural characteristics, the presence of multivessel disease and the incidence of clinical events. 11,12,14 Recently, in the HAWKEYE study, Biscaglia et al. 14 reported that post-PCI QFR lower than 0.90 was associated with a higher rate of VOCE. Kogame et al. 15 demonstrated that the cutoff value of post-PCI QFR was 0.91 in relatively high-risk patients with de novo 3-vessel disease. Our previous study found that post-PCI QFR ≥0.91 was associated with a lower rate of VOCE. 31 In our present study, we were able to identify a threshold for post-procedural QFR (0.94) that could be used to discriminate ISR lesions treated by DCB angioplasty at a higher risk of clinical events. In addition to the utility of this dichotomous approach, results of Cox proportional hazards regression analysis revealed that progressive decreases in the QFR were also related to a higher risk of adverse clinical outcomes at 1-year follow-up.
There are a number of limitations in the current study. First, the small sample size precluded subgroup analysis. Second, the rigorous inclusion and exclusion criteria have theoretically introduced a selection bias; thus, our conclusions cannot be extrapolated to the excluded patients. Third, the determination of the threshold value was based solely on our clinical data; the optimal values may vary by different populations. Additional studies are needed to validate the cutoff value derived in our study. Finally, although a lower QFR was associated with a higher rate of adverse outcomes, this study was not able to evaluate the clinical impact of interventions to improve suboptimal post-procedural functional results. Large randomized controlled trials are necessary to address this question.

| CONCLUSION
A low post-procedural QFR was associated with poor clinical outcomes within 1 year after DCB angioplasty for DES-ISR. Compared with the post-procedural in-stent %DS, the QFR may be superior to predict vessel-related clinical outcomes. Future studies should be conducted to confirm our results and evaluate the utility of the QFR to guide therapeutic decisions in ISR lesions.