Vessel fractional flow reserve-based non-culprit lesion reclassification in patients with ST-segment elevation myocardial infarction: Impact on treatment strategy and clinical outcome (FAST STEMI I study)

Background: Complete revascularization in patients with ST-segment elevation myocardial (STEMI) improves clinical outcome. Vessel fractional flow reserve (vFFR) has been validated as a non-invasive physiological technology to evaluate hemodynamic lesion significance without need for a dedicated pressure wire or hyper- emic agent. This study aimed to assess discordance between vFFR reclassification and treatment strategy in intermediate non-culprit lesions of STEMI patients and to assess the clinical impact of this discordance. Methods: This was a single-center, retrospective cohort study. From January 2018 to December 2019, consecutive eligible STEMI patients were screened based on the presence of a non-culprit vessel with an intermediate lesion (30 – 80% angiographic stenosis) feasible for offline vFFR analysis. The primary outcome was the percentage of non-culprit vessels with discordance between vFFR and actual treatment strategy. The secondary outcome was two-year vessel-oriented composite endpoint (VOCE), a composite of vessel-related cardiovascular death, vessel-related myocardial infarction, and target vessel revascularization. Results: A total of 441 patients (598 non-culprit vessels) met the inclusion criteria. Median vFFR was 0.85 (0.73 – 0.91). Revascularization was performed in 34.4% of vessels. Discordance between vFFR and actual treatment strategy occurred in 126 (21.1%) vessels. Freedom from VOCE was higher for concordant vessels (97.5%) as compared to discordant vessels (90.6%)( p = 0.003), particularly due to higher adverse event rates in discordant vessels with a vFFR ≤ 0.80 but deferred revascularization. Conclusions: In STEMI patients with multivessel disease, discordance between vFFR reclassification and treatment strategy was observed in 21.1% of non-culprit vessels with an intermediate lesion and was associated with increased vessel-related adverse events.


Introduction
Approximately 50% of patients with ST-segment myocardial infarction (STEMI) have concomitant multivessel disease. [1] Previous studies demonstrated improved clinical outcome with complete revascularization. [2,3] Several trials challenged the additional value of fractional flow reserve (FFR) for intermediate non-culprit lesions. [4][5][6] Whereas no conclusive outcome data is available on the superiority of FFR-vs. angiography-guided complete revascularization in patients presenting with STEMI, the relevance of the topic is illustrated by the fact that FFR for intermediate non-culprit lesions with at least 50% angiographic diameter stenosis appeared negative in 30-50%, questioning the need for percutaneous coronary intervention (PCI). [4][5][6] As FFR relies on the use of a dedicated pressure wire and hyperemic agent, simplified physiological tools could enhance the adoption of physiological lesion assessment in the acute setting. Vessel fractional flow reserve (vFFR), which is based on three-dimensional quantitative coronary angiography (3D-QCA), is a novel non-invasive physiological technology which showed a good diagnostic agreement with invasively measured FFR. [7,8] Given its concept, the technology has the potential to better guide non-culprit lesion treatment, both in an acute or offline Heart Team setting, without the need for a dedicated pressure wire and hyperemic agent. [9] The aim of this study was to assess discordance between vFFR reclassification and actual treatment strategy in intermediate nonculprit lesions of STEMI patients and to assess the clinical impact of this discordance.

Study design and patient population
This was a single-center, retrospective cohort study. Consecutive patients presenting with STEMI between January 1st, 2018, to December 31st, 2019, and admitted to the catheterization laboratory for primary PCI were screened for eligibility. Patients were eligible for enrollment if at least one intermediate lesion (30-80% angiographic diameter stenosis and a reference diameter ≥2.00 mm by visual estimation or offline 3D-QCA) was present in a non-culprit vessel. Patients with prior coronary artery bypass graft surgery (CABG), prior heart transplantation, or presentation with cardiac arrest or cardiogenic shock were excluded. Coronary angiograms were subsequently screened for criteria precluding the feasibility of vFFR computation (presence of aorta-ostial lesions, insufficient angiographic projections (two angiograms with a rotation/angulation of <30 degrees), severe overlap or foreshortening of the target lesion, and table movement during projection acquisition).
PCI was performed according to routine clinical practice and in correspondence with current guidelines, including the use of peri-and post-procedural antithrombotic therapy. [10] The use of FFR or nonhyperemic pressure ratios for the physiological assessment of nonculprit lesions in the acute setting was at the discretion of the operator.
The Ethical Committee of the Erasmus University Medical Center approved the study protocol and waived the need for informed consent, since the study was not subject to the Dutch Research on Humans Subject Act.

vFFR analysis
vFFR computations were performed offline by the Erasmus University Medical Center CoreLab. Two angiographic projections of the nonculprit vessel were exported to the CAAS workstation 8.5.1 (Pie Medical Imaging, Maastricht, the Netherlands). Temporal alignment of the two coronary angiograms was performed automatically by electrocardiogram triggering and optimal end diastolic frames were semiautomatically identified. Subsequently, semi-automatic contouring of the vessel was achieved for both angiographic projections, by selecting at least two points (at the ostium of the vessel and distal to the stenosis). Manual contour correction was allowed if deemed necessary. Results on interobserver variability of the methodology have been published previously. [7,8,11] The vessel contouring resulted in a 3D-QCA vessel model, providing the following parameters: lesion obstruction length (mm), lesion position (mm), minimal lumen diameter (mm), diameter stenosis (%) and reference diameter (mm). Based on this 3D-QCA model, the vFFR value was calculated automatically after entering the systolic and diastolic aortic root pressure.

Study outcomes
The primary and clinical secondary outcomes were assessed at vessel level.
The primary outcome was the percentage of vessels with discordance between offline vFFR and actual treatment strategy. Two treatment strategies were distinguished: 1) Subsequent revascularization, defined as revascularization at the time of primary PCI or in a staged setting (within 3 months); 2) Deferral from treatment. If treatment options were first discussed within a multidisciplinary Heart Team, the recommended treatment strategy of the (ad-hoc) Heart Team was used for this study. Concordance was defined as a vFFR ≤0.80 with subsequent revascularization or a vFFR >0.80 with deferred revascularization. Discordance was defined as a vFFR ≤0.80 with deferred revascularization or a vFFR >0.80 with subsequent revascularization.
Secondary outcomes were 1) vessel-oriented composite endpoint (VOCE) at two years, including vessel-related cardiovascular death, vessel-related myocardial infarction and target vessel revascularization (TVR); 2) the diagnostic performance of offline vFFR with acute-setting FFR as the reference standard (cutoff value ≤0.80).
Events were designated as vessel related or non-vessel related. [12,13] Cardiovascular death was defined as any death without a clear non-cardiovascular cause. If cardiovascular death was not clearly related to a specific coronary artery, vessel-related cardiovascular death was assumed. Likewise, vessel-related myocardial infarction was considered if no clear culprit vessel could be identified. Consequently, vessel-related cardiovascular death and myocardial infarction could be assigned to multiple non-culprit vessels per patient. [12,13] TVR was defined as any revascularization of the non-culprit vessel.

Patient data and follow-up
Baseline and procedural data were extracted from the hospital's electronic medical record system and stored in a dedicated database. Follow-up data was collected by screening hospital's electronic medical records, telephone surveys and the use of a dedicated local online platform for automated collection of patient reported outcome measurements (CathSuite).

Statistical analysis
The Shapiro-Wilk test was used to evaluate whether continuous variables followed normal distribution. Normally distributed variables were presented as mean with standard deviation (SD), while nonnormally distributed continuous variables were presented as median with 25th-75th percentiles. Categorical variables were reported as counts with percentages.
Continuous patient-level variables were compared using the independent-samples t-test or Mann-Whitney U test, while categorical patient-level variables were compared using the Pearson's χ2 test or Fisher's exact test (as appropriate). Non-culprit lesion characteristics were compared using generalized linear mixed models to adjust for clustering of multiple non-culprit vessels per patient.
Cumulative freedom from event percentages were derived from the Kaplan-Meier function for discordant and concordant groups. Censoring was performed at the time of event, non-vessel-related cardiovascular death, last contact or after two years of follow-up. Univariate Cox regression models with robust standard errors to take into account clustering of multiple non-culprit vessels per patient were used to estimate hazard ratios (HRs) including 95% confidence intervals (CIs) for VOCE and its individual components. To test robustness of results, a sensitivity analysis including only a single non-culprit vessel per patient was performed. In patients with multiple vessels, the non-culprit vessel with the lowest vFFR value was used for this analysis. For this sensitivity analysis, Kaplan-Meier curves were compared with the log-rank test.
The correlation between offline vFFR and acute-setting FFR was displayed in a scatter plot and numerically expressed with the Pearson's correlation coefficient (r). The diagnostic performance of vFFR, including sensitivity, specificity, diagnostic accuracy, positive predictive value (PPV) and negative predictive value (NPV), was determined for a cutoff value of ≤0.80 for both FFR and vFFR. Additionally, receiver operating characteristic (ROC) curve analysis was performed to evaluate the discriminative ability of vFFR. If a patient was included with multiple non-culprit vessels, only the vessel with the lowest FFR value was used for this subgroup analysis.
Statistical analysis was performed with SPSS version 28 and R (R Core Team 2021; version 4.1.0, packages: glmm, survival). A 2-sided p < 0.05 was considered statistically significant.

Study population
Of the 923 consecutive STEMI patients screened, 425 did not meet the clinical entry criteria due to cardiac arrest or cardiogenic shock at presentation (n = 158), prior coronary artery bypass graft surgery (n = 34), prior heart transplant (n = 1), and absence of an intermediate lesion in a non-infarct-related artery (n = 232) (supplemental Fig. A). Subsequently, 57 patients with aorta-ostial lesions (n = 10) or insufficient angiographic projections precluding the feasibility of vFFR computation (n = 47) were excluded. Finally, 441 patients (598 vessels) were included in the present study.

Patient, procedural and culprit lesion characteristics
The baseline characteristics of the included patients are presented in Table 1. Mean age was 65.8 (SD 11.6) years and 73.7% of patients were male. Diabetes was present in 15.2% of patients and 14.7% had undergone prior PCI.
Primary PCI of the culprit vessel was performed in 99.5% with final thrombolysis in myocardial infarction (TIMI) flow grade 3 in 93.7% (supplemental table A).
Baseline, culprit lesion and procedural characteristics did not differ significantly for patients with discordant and concordant nonculprit vessels, except for a higher rate of prior cerebrovascular accidents in discordant patients (10.7% vs. 5.3%, p = 0.047) (supplemental table B and C). Comparing non-culprit lesion characteristics between discordant and concordant vessels revealed that the overall lesion severity as assessed by percentage diameter stenosis, minimal lumen diameter and (v)FFR was significantly worse in discordant vessels as compared to concordant vessels (supplemental table D). More specifically, revascularized vessels with a vFFR ≤0.80 (concordant) had a greater diameter stenosis (60.0% vs. 54.5%, p = 0.002) and obstruction length (24.6 mm vs. 21.3 mm, p = 0.0498), with a lower median vFFR Values are mean (standard deviation) or n (%). CVA = cerebrovascular accident, MI = myocardial infarction, PCI = percutaneous coronary intervention, TIA = transient ischemic attack.   Values are n (%). CABG = coronary artery bypass graft surgery, PCI = percutaneous coronary intervention, vFFR = vessel fractional flow reserve.
(0.66 vs. 0.73, p = 0.008), as compared to vessels with a vFFR ≤0.80 that were not revascularized (discordant) (supplemental table E). The distribution of discordance and concordance within the different nonculprit vessels did not differ significantly (p = 0.08, Fisher's exact test) (supplemental Fig. B).

Clinical outcome
Two-year follow-up was available in 88.1% (vessel based). Freedom from VOCE was significantly higher for concordant vessels (97.5%) as compared to discordant vessels (90.6%) (HR 3.8; 95% CI 1.6-9.0; p = 0.003), particularly due to higher event rates in vessels with a vFFR ≤0.80 but deferred revascularization ( Fig. 1 and supplemental Fig. C). More specifically, as compared to an event-free survival of 98.0% in concordant vessels with a vFFR >0.80 and deferred revascularization, event-free survival was 96.6% in concordant vessels with a vFFR ≤0.80 and subsequent revascularization (p = 0.34), 97.9% in discordant vessels with a vFFR >0.80 and subsequent revascularization (p = 0.96), and 85.8% in discordant vessels with a vFFR ≤0.80 but deferred revascularization (p < 0.001).
In the sensitivity analysis, including one non-culprit vessel per patient, freedom from VOCE was higher for patients with a concordant vessel (97.4%) as compared to patients with a discordant vessel (89.9%) (p = 0.002) (supplemental Fig. D).

Diagnostic performance of vFFR with acute-setting FFR as the reference standard
Acute-setting FFR measurement was performed in 42 patients. The correlation between offline vFFR and acute-setting FFR was moderate (r = 0.61, p < 0.001) (supplemental Fig. E). Diagnostic performance of offline vFFR with acute-setting FFR as the reference standard was characterized by a sensitivity of 70%, specificity of 87.5%, diagnostic accuracy of 83.3%, PPV of 63.3%, and NPV of 90.3%. The ROC curve analysis revealed a good discriminative ability of vFFR to predict FFR ≤0.80, with an area under the curve of 0.86 (p = 0.001).

Discussion
The results of this study can be summarized as follows: 1) In STEMI patients with multivessel disease, discordance between physiological lesion classification based on vFFR and actual treatment strategy was present in 21.1% of non-culprit vessels; 2) Freedom from VOCE at two years was significantly higher for concordant non-culprit vessels as compared to discordant non-culprit vessels, particularly due to higher adverse event rates in vessels with a vFFR ≤0.80 but deferred revascularization; 3) offline vFFR showed good diagnostic performance with acute-setting FFR as the reference standard.
The present study demonstrates that vFFR has a potential role in guiding revascularization of intermediate non-culprit lesions in patients presenting with STEMI. Discordance between offline vFFR and actual treatment strategy was observed in over one fifth of the non-culprit vessels. More specifically, 33.1% of vessels underwent no subsequent revascularization despite a vFFR of ≤0.80 whereas 13.3% of vessels with a vFFR >0.80 underwent subsequent PCI or CABG which likely could have been avoided. Event rates were significantly higher in the discordant group, particularly driven by events related to non-culprit lesions with a vFFR ≤0.80 but deferred revascularization. The latter follows the accepted concept of impaired clinical outcome related to incomplete revascularization and was also found in a small post-hoc analysis (n = 110) of the EXAMINATION trial, demonstrating that deferring treatment of intermediate lesions in non-culprit vessels with a QFR ≤ 0.80 was associated with higher patient-oriented cardiac events at 5 years (HR 2.3; 95% CI 1.2-4.5; p = 0.01). [2,3,14] Finally, specifically addressing the potential role of vFFR in this subset of patients, also a small retrospective study (n = 156) found a significant number of non-culprit lesions with a positive vFFR but deferred revascularization in STEMI patients. Likely due to a lack of power, this did not result in higher adverse event rates. [15] Previous studies in patients presenting with STEMI and multivessel disease demonstrated the limitations of visual lesion assessment to adequately interpret physiological lesion significance and reported negative FFR values for intermediate non-culprit lesions with at least 50% angiographic diameter stenosis in 30-50%. [4][5][6] With a more liberal definition in the present study (30-80% angiographic diameter stenosis), we found negative vFFR values in 60.5% of the cases. Whereas prospective randomized data on the superiority of physiology-guided non-culprit vessel PCI in patients presenting with STEMI is lacking, data derived from important registries showed low event rates in vessels with a negative FFR and no subsequent revascularization in patients presenting with stable or unstable angina, supporting a conservative approach. [16,17] A meta-analysis of large national FFR registries extrapolated these findings to patients presenting with acute coronary syndrome (ACS) by demonstrating that FFR-based deferral to medical treatment was as safe as in patients with non-ACS (major cardiovascular event, 8.0% vs. 8.5%, p = 0.83; revascularization, 3.8% vs. 5.9%, p = 0.24; and freedom from angina, 93.6% vs. 90.2%, p = 0.35). [18] Also in the present study, event rates related to lesions with a vFFR >0.80 were low irrespective of subsequent revascularization. The ongoing prospective FRAME-AMI trial (NCT02715518) will provide more evidence on the topic.
Thus far, the uptake of physiology in the primary PCI setting remains limited and is often restricted to staged settings. [19] The latter supports the development of faster and easier means of physiological lesion assessment. With a short analysis time (3.4 to 5.0 min on average), and no need for dedicated pressure wires, microcatheters and/or hyperemic agents, angiography-based FFR offers a unique opportunity for both acute-setting as well as offline physiological lesion assessment guiding complete revascularization or subsequent Heart Team discussion. [9,20] Out of the 498 patients meeting clinical entry criteria in our study, vFFR computation appeared not feasible in only 57 patients (11.4%), illustrating that vFFR is a suitable technique for physiological lesion assessment in the majority of patients, even in a study population with no specific focus on proper image acquisition for the purpose of angiography-based FFR.
Finally, offline vFFR showed good diagnostic performance with acute-setting FFR as the reference standard. However, the use of angiography-based technologies in the acute setting needs further validation, especially since acute-setting FFR slightly underestimates the hemodynamic significance of non-culprit lesions due to microvascular vasoconstriction and a blunted hyperemic response. [21,22] Conversely, angiography-based physiological tools that do not include TIMI frame counting in their computational model, such as vFFR, are likely not affected by changes in the microvasculature or an insufficient hyperemic response. Our subgroup analysis showed promising results with a diagnostic accuracy of 83.3%. Using competitive technologies, small studies demonstrated a good correlation between acute-setting QFR and acute-setting FFR, as well as between offline QFR and staged FFR. [14,23] Nevertheless, a dedicated prospective validation study is needed to further investigate the use of angiography-based technologies in STEMI patients.

Limitations
This study has several limitations that should be acknowledged. First, this was a single-center, retrospective cohort study. All types of bias related to its single-center design and retrospective nature should thus be considered. In addition, multivariable Cox regression analysis to adjust for any potential confounding was not performed due to the low number of events. Second, angiography-based FFR technologies, including ischemic cutoff values, have largely been validated in patients with stable coronary artery disease and non-ST-segment elevation acute coronary syndrome. Despite promising results in patients presenting with STEMI, a prospective validation study investigating the diagnostic performance of non-invasive physiological tools with acute-setting and/ or staged FFR as the reference standard, is needed. Third, two-year follow-up was available in 88.1% (vessel based), indicating that events could have been missed.

Conclusions
In STEMI patients with multivessel disease, discordance between vFFR reclassification and actual treatment strategy was observed in 21.1% of non-culprit vessels with an intermediate lesion and was associated with increased vessel-related adverse events, particularly driven by deferred revascularization in vessels with a vFFR ≤0.80. Offline vFFR showed good diagnostic performance with acute-setting FFR as the reference standard. NPV = negative predictive value, PPV = positive predictive value.