Angiographic Quantitative Flow Ratio-Guided Treatment of Patients With Physiologically Intermediate Coronary Lesions.
Mar 21, 2025DOI: 10.1161/jaha.124.035756Journal ArticleAHA
Written by: Jiannan DaiChangdong GuanXueming XuJingbo HouHaibo JiaHuai YuZening JinGuosheng FuXiaofan WuLiansheng WangRongchong HuangZhujun ShenYanyan ZhaoYuanzhe JinLei SongShengxian TuShubin QiaoBo YuBo XuGregg W StoneFAVOR III China study group
The data supporting this subgroup analysis belong to the FAVOR III China study and will not be shared because of informed consent limitations. However, the study investigators are open to collaborating on summary‐level analyses. Requests should be directed to the principal investigator.The study protocol and principal results from the FAVOR III China trial have been previously reported.12, 13, 21 In brief, the FAVOR III China trial was an investigator‐initiated, multicenter, sham‐controlled, blinded randomized trial that was performed at 26 hospitals in China (ClinicalTrials.gov identifier: NCT 03656848). In brief, adult patients with stable or unstable angina or ≥72 hours post–myocardial infarction (MI) in whom PCI was planned according to standard angiography were eligible for enrollment. Angiographic inclusion criteria required at least 1 lesion with a percentage diameter stenosis of 50% to 90% in a coronary artery with at least a 2.5‐mm reference vessel diameter by visual assessment. Complete inclusion and exclusion criteria have been previously reported.12 This study adhered to the principles of the Declaration of Helsinki and received approval from the ethics committees at each participating site. All participating patients provided written informed consent.Before randomization, the operator documented all angiography‐based target vessels intended for PCI. Eligible patients were randomly assigned to QFR guidance or angiography guidance in a 1:1 ratio, stratified by diabetes, multivessel disease, the presence of any vessel with diameter stenosis >90%, Thrombolysis in Myocardial Infarction flow grade <3, and study center. Patients, their families, and all postcatheterization laboratory physicians and researchers were blinded to randomization. The blinding procedures have been previously described.12In the QFR‐guided group, online QFR was measured in all target vessels with 2 angiographic imaging runs with minimum 25° separation in projection angle, and the data were transmitted to the AngioPlus system (Pulse Medical, Shanghai, China) for real‐time calculation. PCI was performed for lesions with an online QFR ≤0.80 or with a diameter stenosis >90%; otherwise, PCI was deferred (not performed). In the angiography‐guided group, PCI was performed on the basis of visual angiographic assessment according to local standard practice. In both groups, staged PCI procedures within 60 days were allowed for multivessel disease, and optimal medical therapy was prescribed during follow‐up. Patients were followed up by telephone or clinic visits at 1 month, 6 months, 1 year, 2 years, and up to 5 years after randomization. At present, all patients have completed 2‐year follow‐up.
The angiograms from both groups were analyzed at an independent core laboratory (Core Medical Co, Ltd, Beijing, China). In addition to standard quantitative coronary angiographic analyses, offline QFR measurements were performed in all patients, including the sham control group. For the present prespecified substudy, patients were categorized as the uncertainty‐zone subgroup if offline QFR was 0.75 to 0.85 in all coronary arteries with a lesion with diameter stenosis ≥50%. Otherwise, if patients had at least 1 vessel with offline QFR <0.75, they were categorized as the QFR <0.75 subgroup, and the remaining patients were categorized as the QFR >0.85 subgroup. Additionally, a sensitivity analysis was conducted to further divide patients into 2 subgroups: patients with a QFR of 0.75 to 0.80 and patients with a QFR of 0.80 to 0.85.
The primary end point of the FAVOR III China trial was the 1‐year rate of major adverse cardiac events (MACEs; the composite of all‐cause death, MI, or ischemia‐driven revascularization). The major secondary end point was the 1‐year rate of MACEs, excluding procedural MI (PMI). PMIs were those related to the initial index or planned staged procedures, but not from unplanned procedures during follow‐up. Revascularization‐related MIs occurring during unplanned follow‐up were counted as non‐PMIs. An independent committee, blinded to randomized allocation, adjudicated all end point events. The detailed end point definitions have been previously reported12 and are provided in Table S1. The present prespecified subanalysis reports the 2‐year rates of MACEs, MACEs excluding PMI, and the individual components of MACEs.
All end points were analyzed in the intention‐to‐treat population. Continuous variables were presented as mean±SD or median (interquartile range) and were compared by 2‐sample Student t‐tests or the Wilcoxon rank‐sum test as appropriate. Categorical variables were presented as number (percentage) and compared by means of the likelihood ratio χ2 test or Fisher exact test. Time‐to‐first event rates were estimated using Kaplan‐Meier methods and were compared by the log‐rank test. Between‐group differences were estimated by hazard ratios (HRs) with 95% CIs using a Cox proportional hazards model. Landmark analyses were performed to examine outcomes before and after 1 year, testing the interaction between treatment and time. A sensitivity analysis focusing on Pearson correlation was used to measure the correlation between online QFR and offline QFR, and the agreement between them was evaluated using Bland‐Altman plots. The relative hazards between subgroups were examined for consistency with interaction testing. Statistical analyses were performed using SAS software version 9.4 (SAS Institute, Cary, NC), adopting a 2‐tailed P<0.05 for statistical significance.
RESULTS
Between December 25, 2018, and January 19, 2020, 3825 patients were randomized (1913 in the QFR‐guided group and 1912 in the angiography‐guided group), 3806 of whom had angiograms sufficient for offline QFR analysis by the core laboratory. On the basis of prespecified subgroup definitions, this offline QFR assessment categorized 873 (22.9%) patients as the uncertainty‐zone subgroup (448 in the QFR‐guided group and 425 in the angiography‐guided group), representing the present study population (Figure 1).
Of the 3825 patients included in the intention‐to‐treat population, 3806 underwent offline quantitative flow ratio (QFR) assessment in the core laboratory. Among them, 873 (22.9%) patients were categorized as the uncertainty‐zone subgroup, in whom all interrogated vessels had QFR values between 0.75 and 0.85, including 448 assigned to the QFR guidance and 425 assigned to conventional angiography guidance.
The baseline clinical and angiographic characteristics were well balanced between the randomized treatment groups in the uncertainty‐zone (Table 1). The mean patient age was 63.2 years, 67.8% were men, 31.6% had diabetes, and 64.6% presented with an acute coronary syndrome (unstable angina or post‐MI). The mean±SD anatomic Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery (SYNTAX) score was 7.5±5.1.
Baseline Clinical and Angiographic Characteristics
Data are number (percentage), mean±SD, median (interquartile range), or number/toatl (percentage). MDRD indicates Modification of Diet in Renal Disease; QFR, quantitative flow ratio; SYNTAX, Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery; and TIMI, Thrombolysis in Myocardial Infarction.
The anatomic SYNTAX score is a scoring system that quantifies angiographic lesion extent and complexity; it was calculated by the angiographic core laboratory.
The functional SYNTAX score was calculated by summing the segmental anatomic SYNTAX scores only in vessels with functional ischemia, as defined by offline QFR ≤0.80, as established by the angiographic core laboratory.
The procedural characteristics for patients in the uncertainty‐zone subgroup are summarized in Table 2. The prerandomization vessel revascularization plan was changed in 76 (17.0%) of 448 patients in the QFR‐guided group compared with 13 (3.1%) of 425 patients in the angiography‐guided group (P<0.001), driven by the deferral of at least 1 vessel originally intended for PCI (69 [15.4%] versus 10 [2.4%]; P<0.001) as well as treatment of at least 1 vessel not originally intended for PCI (12 [2.7%] versus 4 [0.9%]; P=0.05). Consequently, PCI was performed in fewer patients guided by QFR than those guided by angiography (402 [89.7%] versus 421 [99.1%]; P<0.001). In addition, patients in the QFR‐guided group also required fewer stents, required less contrast volume, and had shorter fluoroscopy and procedure times. Compared with angiography guidance, QFR‐guided decision‐making resulted in a higher rate of patients achieving complete functional revascularization (residual functional SYNTAX score=0 based on the offline QFR assessment) (421 [94.2%] versus 361 [86.0%]; P<0.001). Medication use at discharge and during follow‐up between the randomized treatment groups was similar, except fewer antiplatelet agents in the QFR‐guided group (Table S2), reflecting the less frequent performance of PCI.
Procedural Characteristics
Data are number (percentage), mean±SD, or number/total (percentage). PCI indicates percutaneous coronary intervention; QFR, quantitative flow ratio; and SYNTAX, Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery.
Intravascular ultrasound or optical coherence tomography.
Procedure time was defined as the time between the first and last angiogram, including diagnostic coronary angiography, the randomization process, QFR calculation time or 10‐minute sham control time, and PCI time; in the adjusted procedure time, the mandated 10‐minute delay for the real or sham QFR calculation was subtracted.
Defined as residual stenosis of <30% for patients treated with stents or <50% for patients treated with balloon angioplasty by visual estimation, with Thrombolysis in Myocardial Infarction flow grade 3 in the treated vessel.
Defined as lesion success in all treated lesions without in‐hospital major adverse cardiac events (up to a maximum of 7 days).
Among the uncertainty‐zone cohort, 2‐year MACEs occurred in 31 (7.0%) patients in the QFR‐guided group and 35 (8.3%) in the angiography‐guided group (HR, 0.85 [95% CI, 0.52–1.37]) (Table 3 and Figure 2). For the major secondary outcome, the 2‐year MACE rates excluding PMI were 4.8% with QFR guidance versus 7.1% (HR, 0.67 [95% CI, 0.38–1.16]) (Table 3 and Figure 2). Although the risk of MACEs and MACEs excluding PMI was significantly reduced by QFR guidance compared with angiography guidance in the physically ischemic or nonischemic population, the P value for interaction between uncertainty‐zone and non–uncertainty‐zone patients was not statistically significant (Figure S1). The sensitivity analysis indicated a consistent trend, showing that QFR guidance resulted in similar outcomes compared with angiography guidance for 2‐year MACEs and MACEs excluding PMI events in patients with QFR between 0.75 and 0.80, as well as those with QFR between 0.80 and 0.85 (Figures S2 and S3).
Two‐Year Clinical Outcomes
Data are number (Kaplan‐Meier estimated percentage) unless otherwise stated. MACE indicates major adverse cardiac event; PMI, procedural myocardial infarction; and QFR, quantitative flow ratio.
Revascularization of vessels that were actually treated with percutaneous coronary intervention after randomization.
Revascularization of vessels in which percutaneous coronary intervention was not previously performed.
Kaplan‐Meier cumulative event curves for 2‐year major adverse cardiac events (MACEs; A); 2‐year MACEs excluding procedural myocardial infarction (PMI; B); 2‐year MACEs with landmark analysis at 1 year (C); and 2‐year MACEs excluding PMI with landmark analysis at 1 year (D). HR indicates hazard ratio; and QFR, quantitative flow ratio.
As shown in Figure 2, the landmark analysis revealed a delayed reduction in the risk of MACEs with QFR guidance compared with angiography guidance in uncertainty‐zone patients after 1 year (0–1 year: 4.7% versus 3.8%; HR, 1.25 [95% CI, 0.65–2.40]; 1–2 years: 2.3% versus 5.5%; HR, 0.41 [95% CI, 0.20–0.87]; Pinteraction=0.03). Similarly, among uncertainty‐zone patients, the QFR‐guided strategy provided marginally greater relative risk reductions in the major secondary outcome of MACEs excluding PMI after 1 year (2.3% versus 5.5%; HR, 0.41 [95% CI, 0.20–0.87]) compared with before 1 year (2.5% versus 2.1%; HR, 1.16 [95% CI, 0.48–2.80]) (Pinteraction=0.07) (Figure 2). The sensitivity analysis showed that the reduced risk for MACEs after 1 year with QFR guidance was more pronounced in patients with a QFR between 0.80 and 0.85 (Figures S2 and S3).
There were no significant differences in the rates of all‐cause death and MI at 2 years between patients who were treated with QFR guidance versus angiography guidance (Table 3). Notably, however, the QFR‐guided strategy was associated with a lower risk of non‐PMI (0.5% versus 2.4%; HR, 0.19 [95% CI, 0.04–0.86]), and the event curves continued to spread throughout the 2‐year follow‐up, with similar benefit favoring QFR‐guided treatment before and after 1 year (Pinteraction=1.00) (Figure S4). Conversely, the rate of PMI, irrespective of alternative definitions, was comparable between the randomized treatment groups (Table 3 and Table S3). Ischemia‐driven revascularization was performed in 18 (4.1%) patients in the QFR‐guided group and in 19 (4.5%) patients in the angiography‐guided group (HR, 0.91 [95% CI, 0.48–1.73]), and there was a borderline significant interaction between the relative treatment effect on ischemia‐driven revascularization before (2.2% versus 1.2%; HR, 1.90 [95% CI, 0.65–5.57]) and after (1.8% versus 3.6%; HR, 0.51 [95% CI, 0.22–1.20]) 1 year (Pinteraction=0.06) (Figure S5). The rates of MACEs, MACEs excluding PMI, and the individual components arising from treated or deferred vessels were comparable between patients who underwent QFR‐guided versus angiography‐guided treatment (Table S4).
Among the “uncertainty‐zone” subgroup population, offline and online QFR computation showed a correlation of r=0.490 (P<0.0001) and precision with a mean difference of 0.032±0.067, whereas the overall agreement between offline and online QFR ≤0.80 or >0.80 was 92.4% (Figure 3). Moreover, in the QFR‐guided group, 4 patients had 4 vessels with offline QFR ≤0.80 in whom the operators deferred treatment, and 39 patients had 41 vessels with offline QFR >0.80 that were treated with PCI (Table S5).
A, The correlation between offline QFR and online QFR: r=0.490; mean difference=0.032. B, Bland‐Altman plot; dashed line illustrates the mean difference±2 SDs. C, Overall agreement between offline and online QFR ≤0.80 or >0.80 was 92.4%.
DISCUSSION
The present prespecified substudy from the FAVOR III China trial focused on the 2‐year clinical outcomes of the online QFR‐guided versus angiography‐guided strategy in patients with all vessels within the offline QFR uncertainty‐zone subgroup. QFR‐guided lesion selection for PCI in such patients resulted in more changes to the treatment plan (in particular less stenting) compared with angiography‐guided lesion selection, which reduced resource consumption. The observed 2‐year risk of MACEs between the randomized treatment groups did not reach statistical significance in the modest‐sized uncertainty‐zone subgroup. There was a trend toward a lower risk for non‐PMI with QFR guidance. In addition, landmark analysis showed a significant interaction for the relative treatment effects of QFR guidance compared with angiography guidance on MACEs before and after 1 year in the uncertainty‐zone subgroup.
Recent studies have examined the relative merits of revascularization versus deferral for lesions with FFR values between 0.75 and 0.80 (the so‐called gray zone),17, 18, 22, 23, 24, 25 yielding conflicting results. A systematic review has also demonstrated that there is little difference in clinical events attributed to treatment strategies in this group, whether guided by FFR or angiography.26 As an alternative to pressure wire–based FFR, the accuracy of angiography‐based QFR in evaluating functionally significant coronary stenoses has been demonstrated.9, 10 Disagreements between the 2 methods are generally of small magnitude and concentrated in the uncertainty‐zone near their cutoffs (0.75–0.85), which accounts for approximately one‐third of vessels interrogated.15, 16 In the present study, 22.9% of patients with coronary artery disease had all assessed vessels with QFR uncertainty‐zone values. Since the introduction of QFR, its role in guiding treatment decisions in lesions with narrow uncertainty‐zone values has been debated, with some proposing a hybrid QFR‐FFR diagnostic strategy in which decisions based on QFR are limited to those vessels >0.85 and <0.75, a range in which its agreement with FFR is high (90%–95%), otherwise evaluating vessels in the uncertainty‐zone with FFR.11, 14 However, this approach remains theoretical, and its impact on clinical outcomes has not been prospectively evaluated. To our knowledge, the present study is the first to examine the clinical impact of computational QFR lesion selection for revascularization in the uncertainty‐zone subgroup. The treatment effects of the QFR‐guided strategy in the uncertainty‐zone subgroup of patients were consistent with the findings from the QFR <0.75 and QFR >0.85 subgroups. However, the difference in the incidences of MACEs and MACEs excluding PMI in uncertainty‐zone patients randomized to QFR guidance versus angiography guidance (HRs of 0.85 and 0.67, respectively) did not reach statistical significance, likely attributable to reduced power owing to the modest sample size of this subgroup. In addition, the SYNTAX score of 7.5 was low, suggesting the subgroup with a potentially low event rate, making it challenging to detect differences in event rates.
Notably, among uncertainty‐zone patients, the Kaplan‐Meier curves for MACEs and MACEs excluding PMI in the 2 groups crossed at ≈1 year, and landmark analysis demonstrated that QFR‐guided lesion selection had HRs for significantly reduced MACEs and MACEs excluding PMI after 1 year of 0.41 for each. The observed late benefits provided by QFR‐guided strategy may be explained by the sustained risk reduction in non‐PMI (not diluted by PMI in the early period) and a trend for a greater relative treatment effect of ischemia‐driven revascularization after 1 year. Importantly, considering all outcomes through 2‐year follow‐up, QFR guidance was associated with an 81% reduction in the hazard of non‐PMI compared with angiography guidance in uncertainty‐zone lesions (0.5% versus 2.4%; HR, 0.19 [95% CI, 0.04–0.86]). A higher proportion of patients achieving complete functional revascularization (residual functional SYNTAX score=0) and the avoidance of unnecessary treatment of lesions that are not hemodynamically obstructive may explain the lower risk of spontaneous MI in the QFR‐guided group compared with the angiography‐guided group. Meanwhile, early risks for adverse events might be related to the rupture of unstable plaques, which might not be sensitive to functional assessment. This indicates the potential need for intracoronary imaging evaluation for those uncertain‐zone lesions during clinical practice. Additionally, the detailed sensitivity analysis showed that the delayed benefit was more pronounced in patients within the functionally nonischemic range (0.80<QFR≤0.85). These findings suggest that for lesions in the uncertain zone, a safe deferral might offer greater benefits for patients, although as a subgroup analysis, the present results should be considered hypothesis generating because of the small sample size and potential bias. A long‐term follow‐up and further prospective studies are needed to determine whether QFR guidance could provide lasting benefits for this subgroup of patients. Finally, the present data revealed a reduction in stent use with QFR guidance, with similar midterm clinical outcomes compared with angiography guidance. Additional benefits to the QFR strategy are less contrast use and radiation exposure. This partially demonstrates a potential benefit of QFR in identifying lesions that have an excellent prognosis with optimal medical therapy, rather than stenting, even in the uncertainty‐zone population. Whether a QFR‐guided lesion selection strategy provides similar outcomes to an FFR‐guided lesion selection strategy in uncertainty‐zone (and all other) patients is being studied in the ongoing FAVOR III European‐Japan randomized trial in 2000 randomized patients (ClinicalTrials.gov Identifier: NCT03729739).27 Additionally, considering there is no physiological uncertainty‐zone with nonhyperemic pressure ratios (like instantaneous wave‐free ratio), whether guidance with these tools could provide clinical benefits among this population needs further study.
Although the treatment effects of QFR guidance versus angiography guidance were consistent in uncertainty‐zone and non–uncertainty‐zone patients, the uncertainty‐zone subgroup was insufficiently powered to allow for a conclusive interpretation; these results should thus be considered hypothesis generating. Second, the accuracy and reproducibility of QFR measurements are dependent on the technique and quality of the angiography acquisition. In this subgroup analysis, a moderate correlation of QFR values between site and core laboratory determination were demonstrated in the uncertainty‐zone patients, and the diagnostic consistency between online QFR and offline QFR using 0.80 as the cutoff value of 92.4% was acceptable. However, online QFR values were somewhat lower than offline QFR values, leading to a greater propensity for revascularization, increasing the risk of PMI. Nevertheless, the QFR‐guided strategy did not result in a higher risk of MACEs compared with the angiography‐guided strategy. Third, the early‐generation QFR software used in this study required 2 projections for each vessel. The next‐generation method, termed Murray Law–based QFR (μQFR) assessment, enables QFR calculation from a single angiographic view, uses artificial intelligence to delineate major coronary arteries and their side branches, and offers higher diagnostic accuracy in identifying hemodynamically significant coronary stenoses, with excellent reproducibility and convenience.28 Fourth, the PCI operators were aware of the group assignments, potentially leading to procedural bias. Nevertheless, the operators declared their intended treatment strategy before randomization, and the sham‐controlled design facilitated robust blinding of patients and clinical assessors during follow‐up, thereby minimizing placebo and Hawthorne effects, and ascertainment bias. Fifth, the limited use of intracoronary imaging in the FAVOR III China trial precludes insight as to how intravascular imaging may have affected the relative outcomes of a QFR‐guided lesion selection strategy. Sixth, the outcomes of the current study may not be generalizable to excluded patient cohorts, such as those with moderate or severe chronic kidney disease or severe vessel tortuosity. Finally, follow‐up to date is only complete through 2 years; the ongoing 5‐year follow‐up will provide a long‐term perspective on the delayed outcomes of the QFR‐guided lesion selection strategy in all patients as well as those in the uncertainty‐zone subgroup.
Conclusions
The present subgroup analysis of the FAVOR III China trial, which focused on uncertainty‐zone patients with QFR values between 0.75 and 0.85 in all interrogated vessels, demonstrates that lesion selection for PCI based on QFR guidance did not significantly improve 2‐year clinical outcomes compared with angiography guidance, while reducing the rate of PCI and resource consumption.
Sources of Funding
The study was supported by research grants from the National High Level Hospital Clinical Research Funding (2022‐GSP‐GG‐20, 2023‐GSP‐GG‐3), the Capital's Funds for Health Improvement and Research (2022‐2‐4033), the Chinese Academy of Medical Sciences (2022‐I2M‐C&T‐A‐009), and the National Natural Science Foundation of China (82322036; 82072091).
Disclosures
Dr Tu is the cofounder of Pulse Medical, and reports grants and consultancy from Pulse Medical. Dr Stone has received speaker honoraria from Medtronic, Pulnovo, Infraredx, Abiomed, Abbott, Amgen, and Boehringer Ingelheim; has served as a consultant to Daiichi Sankyo, Ablative Solutions, CorFlow, Apollo Therapeutics, Cardiomech, Gore, Robocath, Miracor, Vectorious, Abiomed, Valfix, TherOx, HeartFlow, Neovasc, Ancora, Elucid Bio, Occlutech, Impulse Dynamics, Adona Medical, Millennia Biopharma, and Oxitope; and has equity/options from Ancora, Cagent, Applied Therapeutics, Biostar family of funds, SpectraWave, Orchestra Biomed, Aria, Cardiac Success, Valfix, and Xenter. Dr Stone's employer, Mount Sinai Hospital, receives research support from Abbott, Abiomed, Bioventrix, Cardiovascular Systems Inc, Phillips, Biosense‐Webster, Shockwave, Vascular Dynamics, Pulnovo, and V‐wave. Family disclosure: Dr Stone's daughter is an employee at IQVIA. The remaining authors have no disclosures to report.
Supporting information
Tables S1–S5
Figures S1–S5
GS
Dr. Gregg Stone
Cardiology · Interventional CardiologyIcahn School of Medicine at Mount Sinaigregg
