J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 3 , 2 0 1 9

ª 2 0 1 9 B Y T H E AM E R I C A N C O L L E G E O F C A R D I O L O G Y F O UN DA T I O N

P U B L I S H E D B Y E L S E V I E R

Angiography-Derived Fractional FlowReserve in the SYNTAX II TrialFeasibility, Diagnostic Performance of Quantitative Flow Ratio,and Clinical Prognostic Value of Functional SYNTAX ScoreDerived From Quantitative Flow Ratio inPatients With 3-Vessel Disease

Taku Asano, MD,a,b Yuki Katagiri, MD,a Chun Chin Chang, MD,c Norihiro Kogame, MD,a Ply Chichareon, MD,a

Kuniaki Takahashi, MD,a Rodrigo Modolo, MD,a Erhan Tenekecioglu, MD,c Carlos Collet, MD,a,d Hans Jonker, BSC,e

Clare Appleby, MD,f Azfar Zaman, MD,g Nicolas van Mieghem, MD, PHD,c Neal Uren, MD,h Javier Zueco, MD,i

Jan J. Piek, MD, PHD,a Johan H.C. Reiber, MD, PHD,j Vasim Farooq, MD, PHD,k Javier Escaned, MD, PHD,l

Adrian P. Banning, MD,m Patrick W. Serruys, MD, PHD,c Yoshinobu Onuma, MD, PHDc,f

ABSTRACT

ISS

FrobD

the

Ne

Ca

firmjDe

Ro

Un

Ox

OBJECTIVES The aims of the present study were to investigate the applicability of quantitative flow ratio (QFR) in

patients with 3-vessel disease and to demonstrate the impact of functional SYNTAX (Synergy Between Percutaneous

Coronary Intervention With Taxus and Cardiac Surgery) score derived from QFR (fSSQFR) on clinical outcomes.

BACKGROUND The applicability of QFR in patients with 3-vessel disease and the feasibility of fSSQFR have not yet been

investigated.

METHODS All lesions interrogated using instantaneous wave-free ratio and/or fractional flow reserve in the SYNTAX II

trial were retrospectively screened and analyzed for QFR. The diagnostic performance of QFR was investigated using

hybrid wire-derived pressure assessment (instantaneous wave-free ratio and fractional flow reserve), used in the trial as a

reference. Patients with analyzable QFR in 3 vessels were stratified according to fSSQFR to evaluate its clinical prognostic

value on the basis of 2-year patient-oriented composite endpoint.

RESULTS QFRs were analyzable in 71.0% of lesions (836 lesions). The diagnostic performance of QFR to predict binary

wire-based ischemia was substantial (area under the curve 0.81, accuracy 73.8%), with a positive predictive value of

85.9%. Independent predictors of diagnostic discordance were lesions in side branches, involvement of bifurcation or

trifurcation, and small vessel. According to the 2-year patient-oriented composite endpoint, fSSQFR reclassified 26.1% of

the patients (36 of 138) in the high- to intermediate-risk group into the low-risk group appropriately (net reclassification

improvement 0.32; p < 0.001). The area under the curve for fSSQFR to predict the 2-year patient-oriented composite

endpoint was higher than that of the classic anatomic SYNTAX score (0.68 vs. 0.56; p ¼ 0.002).

CONCLUSIONS QFRdemonstrated substantial applicability in patientswith 3-vessel disease. The fSSQFR has the potential

to further refine prognostic risk estimation compared with the classic anatomic SYNTAX score. (J Am Coll Cardiol Intv

2019;12:259–70) © 2019 by the American College of Cardiology Foundation.

N 1936-8798/$36.00 https://doi.org/10.1016/j.jcin.2018.09.023

m the aDepartment of Cardiology, Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands;

epartment of Cardiology, St. Luke’s International Hospital, Tokyo, Japan; cThoraxCenter, Erasmus Medical Center, Rotterdam,

Netherlands; dDepartment of Cardiology, Universitair Ziekenhuis Brussels, Brussels, Belgium; eCardialysis, Rotterdam, the

therlands; fDepartment of Cardiology, Liverpool Heart and Chest Hospital, Liverpool, United Kingdom; gDepartment of

rdiology, Freeman Hospital Newcastle, Newcastle upon Tyne, United Kingdom; hDepartment of Cardiology, The Royal In-

ary of Edinburgh, Edinburgh, United Kingdom; iDepartment of Cardiology, Hospital Universitario Valdecilla, Cantabria, Spain;

partment of Radiology, Leiden University Medical Center, Leiden, the Netherlands; kManchester Heart Centre, Manchester

yal Infirmary, Central Manchester University Hospitals, Manchester, United Kingdom; lHospital Clinico San Carlos IDISSC and

iversidad Complutense de Madrid, Madrid, Spain; and the mDepartment of Cardiology, John Radcliffe Hospital, Cardiology,

ford, United Kingdom. Dr. Reiber is the CEO of Medis Medical Imaging Systems and has a part-time appointment at Leiden

ABBR EV I A T I ON S

AND ACRONYMS

3VD = 3-vessel disease

AUC = area under the curve

CI = confidence interval

cSS = classic anatomic SYNTAX

score(s)

FFR = fractional flow reserve

fSS = functional SYNTAX

score(s)

fSSiFR/FFR = functional

SYNTAX score derived from

hybrid instantaneous wave-free

ratio and fractional flow

reserve assessment

fSSQFR = functional SYNTAX

score derived from quantitative

flow ratio

iFR = instantaneous wave-free

ratio

IQR = interquartile range

NPV = negative predictive

value

PCI = percutaneous coronary

intervention

POCE = patient-oriented

composite endpoint

PPV = positive predictive value

QFR = quantitative flow ratio

ROC = receiver-operating

characteristic

RVD = reference vessel

diameter

University

Dr. Bannin

relationship

Manuscript

Asano et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 3 , 2 0 1 9

The SYNTAX II QFR Substudy F E B R U A R Y 1 1 , 2 0 1 9 : 2 5 9 – 7 0

260

I schemia-driven percutaneous coronaryintervention (PCI) on the basis of thepressure-derived index of fractional

flow reserve (FFR) has been associatedwith improved outcomes in several largeclinical trials (1–3). PCI with the guidanceof instantaneous wave-free ratio (iFR)demonstrated noninferiority to FFR-guidedPCI with respect to the rate of major adversecardiac events at 12 months in 2 large clin-ical trials (4,5). In the current European So-ciety of Cardiology and American Collegeof Cardiology/American Heart Associationguidelines, pressure-derived physiologicalassessment of stenotic lesions is recommen-ded (Class Ia) for decision making in revas-cularization (6,7).

SEE PAGE 271

The efficacy of PCI with the guidance ofwire-derived pressure index for patients with3-vessel disease (3VD) was demonstrated inthe SYNTAX (Synergy Between PCI WithTaxus and Cardiac Surgery II trial (8). TheSYNTAX II trial was designed to comparethe clinical efficacy of physiology-guidedcontemporary PCI in patients with 3VD withthe state-of-art PCI including the current-generation drug-eluting stent (Synergy, Bos-ton Scientific, Natick, Massachusetts) andintravascular ultrasound with the conven-tional PCI arm of the SYNTAX I trial (8,9).

State-of-the-art PCI guided by hybrid physiologicalassessment comprising iFR and FFR reduced thenumber of lesions to treat and demonstrated improvedclinical results at 12 months (composite of all-causedeath, cerebrovascular event, any myocardial infarc-tion, and any revascularization: 10.6% for SYNTAX IIvs. 17.4% for SYNTAX I; hazard ratio: 0.58; 95%confidence interval [CI]: 0.39 to 0.85; p ¼ 0.006) (8).

On the basis of 3-dimensional angiography,angiography-derived FFR (quantitative flow ratio[QFR]) is computed and results in virtual color-codedpull-backs of FFR of the angiographically imaged ar-teries with stenosis, without the use of a pressurewire or hyperemia (10,11). Several studies reportedsubstantial correlation between QFR and wire-derived FFR in patients with coronary artery disease

Medical Center as professor of medical imaging. Dr. Escaned is a

g receives lecture fees and grant support from Boston Scientific

s relevant to the contents of this paper to disclose.

received June 11, 2018; revised manuscript received September

(10,12,13). Previous studies validating QFR includednoncomplex anatomic lesions and patients (10,12,13).The applicability of QFR in patients with anatomicallycomplex coronary artery disease, such as those with3VD, has not yet been investigated.

It was reported that the functional SYNTAX score(fSS) derived from FFR, which was obtained byscoring only ischemia-provoking lesions on the basisof invasive FFR, has better discriminant ability for therisk for adverse events than the anatomic SYNTAXscore alone (14).

The aims of the present study were to investigatethe diagnostic performance of QFR in patients with3VD using the wire-based physiological assessment asa reference and to demonstrate the impact of fSSderived from QFR (fSSQFR) on prognosis and clinicaloutcomes.

METHODS

The present study was a post hoc substudy of theSYNTAX II trial to investigate the feasibility of phys-iological assessment with QFR in patients with 3VDand the impact of fSSQFR on prognosis and clinicaloutcomes. In the present study, all lesions interro-gated with iFR and/or FFR in the SYNTAX II trial werescreened and analyzed for QFR. In the patients withanalyzable QFR of angiographic stenosis in the 3vessels, fSS was calculated on the basis of QFR, andpatients were stratified according to fSS to evaluateits clinical impact on the basis of 2-year clinicalevents.

STUDY DESIGN AND HYBRID WIRE-BASED

PHYSIOLOGICAL ASSESSMENT OF THE SYNTAX II

TRIAL. The SYNTAX II trial is a multicenter,all-comers, open-label, single-arm study, enrollingpatients with de novo 3VD without left main steminvolvement (NCT02015832) (8). The study design hasbeen described previously (15). The trial enrolled454 patients with 1,559 anatomic target lesions at 22interventional cardiology centers from 4 Europeancountries, following the local heart team consensuswith an equipoise recommendation between coronaryartery bypass grafting and PCI on the basis ofthe SYNTAX score II, which is based on a combinationof the anatomic SYNTAX score, clinical characteris-tics, and comorbidities (16). The trial was approvedby the local ethics committee at all participating sites.

consultant for Philips/Volcano and Boston Scientific.

. All other authors have reported that they have no

3, 2018, accepted September 17, 2018.

FIGURE 1 Study Flowchart

DICOM ¼ Digital Imaging and Communications in Medicine; FFR ¼ fractional flow reserve; iFR ¼ instantaneous wave-free ratio; L ¼ number of

lesions; N ¼ number of patients; QFR ¼ quantitative flow ratio; TIMI ¼ Thrombolysis In Myocardial Infarction.

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 3 , 2 0 1 9 Asano et al.F E B R U A R Y 1 1 , 2 0 1 9 : 2 5 9 – 7 0 The SYNTAX II QFR Substudy

261

The enrolled patients were treated with contem-porary PCI on the basis of the specific hybrid physi-ological assessment with iFR and FFR. The details ofthe specific physiological assessment are describedin the Online Appendix and Online Figure 1. TheiFR/FFR (Verrata and PrimeWire Prestige, Volcano,San Diego, California) was measured distal to eachtarget lesion, and the site of iFR/FFR assessmentwas documented with angiography.

In the SYNTAX II trial, coronary angiography wasperformed prior to PCI for anatomic SYNTAX scoringbut without specific acquisition guidelines for QFRanalysis. Angiography was preceded by an intra-coronary injection of isosorbide dinitrate ornitroglycerin.

In the trial, an independent clinical eventscommittee adjudicated adverse events. To investi-gate the clinical impact of fSS, major adverse car-diac events (patient-oriented composite endpoint

[POCE]) were defined as a composite of all-causedeath, any myocardial infarction, or any revascu-larization. The definition of myocardial infarctionused in the SYNTAX II trial has been describedpreviously (8).

QFR ANALYSIS AND ANGIOGRAPHIC PARAMETERS.

All lesions with wire-based physiological assessmentin the SYNTAX II trial were eligible for the presentanalysis. Lesions were excluded from the analysis ifthey: 1) were located <3 mm from the aorta; 2) had areference luminal diameter <2.0 mm by visualassessment; 3) presented slow coronary blood flow(TIMI [Thrombolysis In Myocardial Infarction] flowgrade 1 or 2); 4) were filmed with <2 projections withisocenter calibration information; 5) had severevessel overlap at the stenotic segments; or 6) hadpoor angiographic image quality precluding precisecontour delineation.

TABLE 1 Baseline Characteristics of the Study Patients

QFR Analyzable(n ¼ 386)

QFR Not Analyzable(n ¼ 68) p Value

fSS Analyzable(n ¼ 138)

fSS Not Analyzable(n ¼ 248) p Value

Age (yrs) 66.8 � 9.5 66.0 � 10.7 0.53 65.9 � 9.5 67.2 � 9.5 0.18

Male 93.3 (360/386) 92.6 (63/68) 0.80 94.2 (130/138) 92.8 (231/248) 0.59

Hypertension 76.6 (294/384) 79.4 (50/63) 0.62 75.4 (104/138) 76.9 (190/247) 0.73

Dyslipidemia 77.0 (292/379) 79.0 (49/62) 0.73 71.5 (98/137) 79.8 (194/243) 0.07

Diabetes mellitus type 1 or 2 31.9 (123/386) 19.0 (12/63) 0.04 34.8 (48/138) 30.1 (75/249) 0.35

Current smoker 13.5 (52/386) 19.0 (12/63) 0.24 13.0 (18/138) 13.7 (34/249) 0.87

Previous myocardial infarction 12.7 (49/385) 11.3 (7/62) 0.75 11.6 (16/138) 13.3 (33/248) 0.63

Ejection fraction (%) 58.0 � 8.2 58.4 � 7.6 0.51 58.4 � 8.3 57.6 � 8.2 0.36

Classic anatomic SYNTAX score 20.2 � 6.4 21.0 � 6.4 0.34 19.5 � 5.7 20.5 � 6.7 0.13

SYNTAX score II PCI 30.2 � 8.5 30.2 � 9.6 0.98 29.1 � 7.9 30.8 � 8.8 0.06

Predicted 4-yr mortality with PCI (%) 8.9 � 8.4 9.5 � 10.8 0.57 7.7 � 6.1 9.5 � 9.4 0.06

SYNTAX score II CABG 29.2 � 10.1 28.5 � 12.2 0.64 28.1 � 10.2 29.7 � 10.1 0.12

Predicted 4-year mortality with CABG (%) 8.8 � 8.8 9.8 � 12.1 0.45 8.2 � 8.3 9.2 � 9.0 0.27

Values are mean � SD or % (n/N).

CABG ¼ coronary artery bypass grafting; fSS ¼ functional SYNTAX (Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery) score;PCI ¼ percutaneous coronary intervention; QFR ¼ quantitative flow ratio.

TABLE 2 Baseline Ch

VesselRCALADLCXLeft main coronary a

Severe calcification

Severe tortuosity

Bifurcated lesionMain branch

Medina 0,1,0Medina 0,1,1Medina 1,0,0Medina 1,1,0Medina 1,1,1

Side branchMedina 0,0,1Medina 0,1,1Medina 1,0,1Medina 1,1,1

Trifurcated lesionMain branchSide branch

Lesion length (mm)*

Reference luminal diam

Area stenosis (%)*

Values are % (n/N) or mea

LAD ¼ left descendingQFR ¼ quantitative flow ra

Asano et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 3 , 2 0 1 9

The SYNTAX II QFR Substudy F E B R U A R Y 1 1 , 2 0 1 9 : 2 5 9 – 7 0

262

Off-line QFR analysis was performed by experi-enced analysts certified for use of the software withthe QAngio XA 3D version 1.1 software package (MedisMedical Imaging Systems, Leiden, the Netherlands).

aracteristics of the Study Lesions

QFR Analyzable(836 Lesions)

QFR Not Analyzable(341 Lesions) p Value

0.3927.5 (230/836) 26.7 (91/341)38.9 (325/836) 34.9 (119/341)33.5 (280/836) 38.1 (130/341)

rtery 0.1 (1/836) 0.3 (1/341)

8.6 (72/836) 8.0 (27/341) 0.80

2.3 (19/836) 2.8 (10/341) 0.72

12.0 (100/834) 11.1 (38/341) 0.780.89

17.6 (13/74) 15.3 (4/26)5.4 (4/74) 7.7 (2/26)9.5 (7/74) 11.5 (3/26)

32.4 (24/74) 23.1 (6/26)35.1 (26/74) 42.3 (11/26)

0.3542.3 (11/26) 33.3 (4/12)15.4 (4/26) 16.7 (2/12)11.5 (3/26) 8.3 (1/12)30.8 (8/26) 41.7 (5/12)

0.5 (4/834) 0.2 (1/341) 0.66100.0 (4/4) 0.0 (0/0)0.0 (0/4) 100.0 (1/1)

19.7 � 12.1 NA NA

eter (mm)* 2.48 � 0.52 NA NA

70.0 � 14.8 NA NA

n � SD. *Value derived from 3-dimensional angiography in QFR analysis.

coronary artery; LCX ¼ left circumflex coronary artery; NA ¼ not applicable;tio; RCA ¼ right coronary artery.

For computation of QFR, contrast QFR without hy-peremic setting along with the frame-countingmethod was applied. The analysts were blinded toiFR and FFR results. Details regarding the QFRcalculation have been reported elsewhere (10,17).Briefly, the QFR calculation is based on the 3-dimensional quantitative coronary angiogram recon-structed from 2 angiographic projections withangles $25� apart and volumetric flow rate calculatedby using contrast bolus frame count (10). The volu-metric flow from the proximal to the distal part of thequantified segment of the coronary artery wasassessed by the product of area and flow velocity onthe basis of frame count as assessed using the previ-ously reported method (18).

QFR was analyzed from the ostium of the mainvessels (left anterior descending, right, and leftcircumflex coronary arteries) to the anatomic sitewhere iFR and FFR were interrogated. Whenever thelocation of the iFR/FFR sensor was not identified, theendpoint of the analysis segmentwas set at a landmark(e.g., side branch) located distal to the lesion. Theautomatic reference interpolation function was usedto establish the reference for the calculation. When-ever a proper reference-interpolated line could not beestablished, it was adjusted by using nondiseasedproximal or distal segments. The cutoff value of QFRfor physiological significance was defined as 0.80 (10).

Reference vessel diameter (RVD), lesion length,and percentage area stenosis were derived from the3-dimensional quantitative coronary angiogramsimultaneously analyzed with QFR in QAngio XA3D (7).

TABLE 3 Results of Physiological Indexes

Physiological Index

Pressure derived

iFR (817 lesions) 0.84 (0.69–0.93)

FFR (243 lesions) 0.78 (0.73–0.84)

FFR with iFR between 0.86 and0.93 (181 lesions)

0.78 (0.73–0.83)

Angiography derived

QFR (836 lesions) 0.78 (0.66–0.88)

Values are median (interquartile range).

FFR ¼ fractional flow reserve; iFR ¼ instantaneous wave-free ratio;QFR ¼ quantitative flow ratio.

FIGURE 2 Receiver-Operating Characteristic Curves for Quantitative Flow Ratio and

Percentage Area Stenosis

AS ¼ area stenosis; AUC ¼ area under the curve; QFR ¼ quantitative flow ratio; ROC ¼receiver-operating characteristic.

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 3 , 2 0 1 9 Asano et al.F E B R U A R Y 1 1 , 2 0 1 9 : 2 5 9 – 7 0 The SYNTAX II QFR Substudy

263

fSSQFR AND HYBRID iFR AND FFR ASSESSMENT.

The classic anatomic SYNTAX score (cSS) was calcu-lated on site on the basis of a visual evaluation ofsignificant lesions (diameter stenosis >50% invessels $1.5 mm) and reported using an electroniccase report form. An online calculator, Syntax Scoreversion 2.28 (Syntax Score Working Group), was usedto derive total and per lesion cSS.

In the present study, 2 fSS were calculated on thebasis of both hybrid iFR/FFR and QFR assessment(14). The fSS derived from hybrid iFR and FFRassessment (fSSiFR/FFR) was calculated by summingthe individual points of physiologically significantlesions and excluding physiologically nonsignificantlesions on the basis of the hybrid iFR and FFRassessment. In the case of sequential lesions, the iFRand FFR measured in the most distal segment of thevessel were used to assess the physiological signifi-cance of the sequential lesions (19). The scores from atotal occlusion were included in the fSS calculation.The fSSQFR was calculated in a similar manner usingQFR values. The scores from lesions in small vessels(<2.0 mm) were included in the fSS calculation,despite the unavailability of QFR measurement.

STATISTICAL METHODS. Data are expressed as mean� SD or median (interquartile range). Categorical var-iables were compared using the Pearson chi-squaretest or Fisher exact test, as appropriate. Continuousvariables were compared using the Student’s t-test.Unless otherwise specified, a 2-sided p value <0.05was considered to indicate statistical significance.

The levels of correlation and agreement betweenQFR and iFR were determined using the Pearsoncorrelation coefficient, Passing-Bablok regressionanalysis, and the Bland-Altman method (20,21). Thediscrimination ability of QFR was quantified usingthe receiver-operating characteristic (ROC) curveand the area under the curve (AUC) was comparedwith percentage area stenosis by using the DeLongmethod (22). For a reference, a positive testoutcome of wire-based physiological assessmentwas defined as iFR <0.86 or FFR #0.8, and anegative test outcome was defined as iFR >0.93 orFFR >0.8 (Online Figure 1).

As an ancillary analyses, to assess the predictorsof diagnostic discordance between QFR and hybridiFR and FFR assessment (QFR false positive or QFRfalse negative), multivariate logistic regressionanalysis was conducted. The detailed methodologyis described in the Online Appendix.

The prediction capability of cSS and fSS for the2-year POCE was assessed using ROC curve analysiswith AUC. The risk reclassification of the SYNTAX

score terciles from the classic anatomic model to thefunctional model was assessed using the net reclas-sification index comparing the predictive accuracy ofthe cSS and fSS for the POCE during 2-year follow-up(23). Agreement of the classification according to thetertiles between fSSQFR and fSSiFR/FFR was assessedusing Cohen’s kappa.

All statistical analyses were performed using Rversion 3.4 (R Foundation for Statistical Computing,Vienna, Austria) and SPSS version 24.0 (IBM, Armonk,New York).

FIGURE 3 2-by-2 Table for Diagnostic Performance of Quantitative Flow Ratio Against Instantaneous Wave-Free Ratio/Fractional Flow

Reserve Hybrid Assessment

In the hybrid physiological assessment, when instantaneous wave-free ratio (iFR) was between 0.86 and 0.93 (iFR gray zone), fractional flow

reserve (FFR) was measured to confirm functional significance. A positive test outcome was therefore defined as iFR <0.86 or FFR #0.80,

and a negative test outcome was defined as iFR >0.93 or FFR >0.80. L ¼ number of lesions; NPV ¼ negative predictive value; PPV ¼ positive

predictive value.

Asano et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 3 , 2 0 1 9

The SYNTAX II QFR Substudy F E B R U A R Y 1 1 , 2 0 1 9 : 2 5 9 – 7 0

264

RESULTS

In the SYNTAX II trial, 447 patients with 1,177 lesionswere interrogated with iFR or FFR (only iFR per-formed, 840 lesions; iFR and FFR performed, 310 le-sions; only FFR performed, 27 lesions). Among thoselesions, QFR was analyzable in 836 lesions (analyz-ability 71.0%). In 386 patients (86.3%), QFR wasanalyzable in at least 1 lesion, whereas in 109 patients(28.2%), QFR was analyzable in angiographic anatomicstenoses in 3 vessels. QFR was not analyzable in 341lesions, mainly because of the absence of 2 appropriateprojections (severe vessel overlaps or tortuosity atlesion, and so on) (Figure 1). The baseline characteris-tics of the study population are shown in Tables 1 and 2.

Median QFR was 0.78 (IQR: 0.66 to 0.88; n ¼ 836),and median iFR and FFR were 0.84 (IQR: 0.69 to 0.93;

n ¼ 817) and 0.78 (IQR: 0.73 to 0.84; n ¼ 181),respectively (Table 3). The distribution of QFR andiFR is shown in Online Figure 2.

CORRELATION BETWEEN iFR AND QFR. The corre-lation between iFR and QFR is shown in OnlineFigure 3. In the Passing-Bablok linear regressionanalysis, there were no systematic or proportionaldifferences between QFR and iFR (slope 1.00 [95% CI:0.89 to 1.14], intercept 0.05 [95% CI: �0.07 to 0.14]),respectively. The details of the results are shown anddiscussed in the Online Appendix, Online Figure 4.

DIAGNOSTIC PERFORMANCE OF QFR AGAINST

BINARY PHYSIOLOGICAL ASSESSMENT WITH

HYBRID iFR/FFR APPROACH. In the ROC curveanalysis, the AUC for QFR predicting significantischemia using the hybrid wire-based approach was

FIGURE 4 Representative Case of Functional SYNTAX Score Calculation Derived From Quantitative Flow Ratio

A patient with 3-vessel disease yielded a classic anatomical SYNTAX (Synergy Between Percutaneous Coronary Intervention With Taxus and

Cardiac Surgery) score of 27, which was classified in the high-risk group. In the patient, the functional SYNAX score derived from quantitative

flow ratio (QFR) was as low as 9 (low-risk group) because the functional assessment using QFR revealed that the lesions in the right coronary

artery (RCA) and left anterior descending coronary artery (LAD) were not significant.

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 3 , 2 0 1 9 Asano et al.F E B R U A R Y 1 1 , 2 0 1 9 : 2 5 9 – 7 0 The SYNTAX II QFR Substudy

265

0.81 (95% CI: 0.78 to 0.84), and the AUC ofpercentage area stenosis was 0.73 (95% CI: 0.69to 0.76) (difference 0.08; 95% CI: 0.05 to 0.11;p < 0.001) (Figure 2).

The diagnostic accuracy of QFR against hybridwire-based physiologic assessmentwas 73.8% (95% CI:70.6% to 76.8%), and the sensitivity, specificity,positive predictive value (PPV), and negative pre-dictive value (NPV) of QFR were 73.7% (95% CI:69.8% to 77.4%), 73.9% (95% CI: 68.1% to 79.2%),85.9% (95% CI: 82.4% to 88.9%), and 56.7% (95% CI:51.2% to 62.1%), respectively (Figure 3). There were212 lesions (26.2%) with discordant diagnosis be-tween QFR and hybrid iFR and FFR assessment (falsepositive 8.3% [67 of 809], false negative 17.9% [145of 809]).

Sensitivity analysis was performed using iFR with acutoff value of 0.89 as a reference (n ¼ 817). Thisanalysis showed a similar AUC of 0.80 (95% CI: 0.77 to0.83), with accuracy of 72.9% (95% CI: 69.8% to76.0%), sensitivity of 74.6% (95% CI: 70.6% to 78.2%),

specificity of 70.1% (95% CI: 64.5% to 75.2%), PPV of81.6% (95% CI: 77.8% to 85.0%), and NPV of 60.8%(95% CI: 55.3% to 66.0%).

In logistic regression analyses, lesion location inside branches (segments 4, 16, 9, 10, 12, and 14) wasan independent predictor if false-positive QFR, withan odds ratio of 2.07 (95% CI: 1.14 to 3.76), and smallvessel (RVD #2.25 mm) and bifurcation or trifurcationwere independent predictors of increased incidenceof false-negative QFR, with odds ratios of 1.67 (95%CI: 1.14 to 2.44) and 1.81 (95% CI: 1.10 to 2.98) (OnlineTable 1). Stratified analyses according to lesion char-acteristics assessing the diagnostic performance ofQFR are shown and discussed in the OnlineAppendix, Online Figures 5 and 6.FUNCTIONAL SYNTAX SCORE DERIVED FROM QFR

AND HYBRID iFR/FFR ASSESSMENT. In 138 patients(35.8%), fSS was analyzable with both methodologies(fSSiFR/FFR and fSSQFR). The patient flow forfSS calculation is presented in Online Figure 7. Inthose patients, cSS was 19.5 � 5.6, while fSSQFR and

FIGURE 5 Reclassification by Functional SYNTAX Score Derived From Quantitative Flow Ratio and Hybrid Instantaneous Wave-Free

Ratio/Fractional Flow Reserve Assessment

cSS ¼ classic anatomic SYNTAX (Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery) score; fSSiFR/FFR ¼functional SYNTAX score derived from hybrid instantaneous wave-free ratio/fractional flow reserve assessment; fSSQFR ¼ functional SYNTAX

score derived from quantitative flow ratio; NRI ¼ net reclassification index.

Asano et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 3 , 2 0 1 9

The SYNTAX II QFR Substudy F E B R U A R Y 1 1 , 2 0 1 9 : 2 5 9 – 7 0

266

fSSiFR/FFR were reduced to 14.8 � 7.0 and 16.0 � 6.5,respectively. A representative case of fSS calculationon the basis of QFR is shown in Figure 4.

Risk classification of patients was performed ac-cording to tertiles of cSS (<17, low-risk group; 17 to 22,intermediate-risk group; >22, high-risk group). Afterthe calculation of fSS (fSSQFR and fSSiFR/FFR), 26.1% ofpatients were reclassified from the high- orintermediate-risk group into the low-risk group byQFR, and 20.3% of patients were reclassified into thelow-risk group by iFR/FFR (Figure 5). Substantialagreement of the risk classification on the basis offSSQFR with that based on fSSiFR/FFR was observed(kappa ¼ 0.740).

IMPACT OF fSSQFR AND fSSiFR/FFR ON 2-YEAR

CLINICAL ENDPOINTS. The median follow-up dura-tion was 788 days (IQR: 738 to 1,098 days). Theincidence of the 2-year POCE stratified according toSYNTAX score tercile (cSS, fSSQFR, and fSSiFR/FFR) ispresented in Figure 6. In the classification according

to cSS, the POCE occurred 4.4%, 9.8%, and 9.5% ofpatients in the low-, intermediate-, and high-riskgroups, respectively, at 2 years (p ¼ 0.57, chi-squaretest). Regarding fSS, POCE rates were 3.7%, 11.0%,and 19.0% of patients in the low-, intermediate-, andhigh-risk groups with fSSQFR, respectively (p ¼ 0.05,chi-square test) whereas POCE rates were 4.1%, 9.5%,and 17.0% of patients in the low-, intermediate-, andhigh-risk groups stratified by fSSiFR/FFR (p ¼ 0.11, chi-square test).

The reclassification table with the incidence of the2-year POCE is presented in Table 4. Both fSSQFR andfSSiFR/FFR yielded significantly improved risk classi-fication compared with cSS (net reclassification indexfor fSSQFR 0.32 [95% CI: 0.24 to 0.40; p < 0.001], netreclassification index for fSSiFR/FFR 0.19 [95% CI: 0.01to 0.34; p ¼ 0.004).

The ROC curves for cSS, fSSQFR, and fSSiFR/FFR areshown with AUCs predicting the 2-year POCE inFigure 7. The AUC of fSSQFR was significantly greaterthan that of cSS (0.68 [95% CI: 0.50 to 0.87] for

FIGURE 6 Incidence of Patient-Oriented Composite Endpoint Stratified According to

Classic Anatomic SYNTAX Score, Functional SYNTAX Score Derived From

Quantitative Flow Ratio, and Functional SYNTAX Score Derived From

Instantaneous Wave-Free Ratio/Fractional Flow Reserve

FFR ¼ fractional flow reserve; iFR ¼ instantaneous wave-free ratio; QFR ¼ quantitative

flow ratio, SS ¼ SYNTAX (Synergy Between Percutaneous Coronary Intervention With

Taxus and Cardiac Surgery) score.

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 3 , 2 0 1 9 Asano et al.F E B R U A R Y 1 1 , 2 0 1 9 : 2 5 9 – 7 0 The SYNTAX II QFR Substudy

267

fSSQFR, 0.56 [95% CI: 0.37 to 0.75] for cSS; p ¼ 0.002).A similar trend was observed in the comparison ofAUCs between fSSiFR/FFR and cSS, although the dif-ference was not statistically significant (0.62 [95% CI:0.42 to 0.82] for fSSiFR/FFR, 0.56 [95% CI: 0.37 to 0.75]for cSS; p ¼ 0.159).

DISCUSSION

In this post hoc substudy of the SYNTAX II trialinvestigating the feasibility and diagnostic perfor-mance of QFR in patients with 3VD and the impact offSSQFR, the main findings are summarized as follows:1) QFR was analyzable in 71.0% of lesions, whileper-patient-level QFR analysis of all angiographicstenoses in the 3 vessels was feasible in 28.2% ofpatients; 2) the diagnostic performance of QFR topredict binary wire-based ischemia was substantial(AUC 0.81), with a PPV of 85.9%; 3) independentpredictors of diagnostic discordance were lesions inside branches, involvement of bifurcation or trifur-cation, and small vessel (RVD #2.25 mm); and 4)according to the 2-year POCE, fSSQFR appropriatelyreclassified high-risk patients to lower categories.

FEASIBILITY OF QFR IN PATIENTS WITH 3VD. In thepresent study, QFR demonstrated acceptable lesionanalyzability. However, without specific acquisitionguidelines, analyzability of the entire coronary treeby QFR was as low as 28.1% of patients. The majorreasons for nonanalyzable QFRs were the lack of 2appropriate projections and the analysis of smallvessels with RVDs <2 mm. How analyzability will beimproved by the implementation of specific guide-lines for angiographic acquisition should beinvestigated.

DIAGNOSTIC PERFORMANCE OF QFR IN PATIENTS

WITH 3VD AND FACTORS AFFECTING DIAGNOSTIC

DISCORDANCE. In the present study in patients with3VD, substantial diagnostic performance of QFR wasobserved, with a high PPV (AUC 0.81, PPV 85.9%). Theprognostic value of QFR in patients with 3VDmay havea considerable impact on the treatment decision.

However, in the present study, the overall NPV waslow (56.7%). In the logistic regression analysis, inde-pendent factors for false negatives were small vessels(#2.25 mm) and involvement of bifurcation or trifur-cation (odds ratios: 1.67 [95% CI: 1.14 to 2.44] and 1.81[95% CI: 1.10 to 2.98], respectively). In those lesionswith small vessels or bifurcation or trifurcation, theNPV of QFR was 42.1% (95% CI: 33.9% to 50.8%). Thelimited diagnostic performance of QFR in thoselesions should be acknowledged.

RISK RECLASSIFICATION BY fSSQFR. In the presentstudy, fSSQFR yielded a significant improvementin risk classification (0.32; 95% CI: 0.24 to 0.40;p < 0.001) as well as fSSiFR/FFR (0.19; 95% CI: 0.01 to0.34; p ¼ 0.004). The AUC of fSSQFR predicting the2-year POCE was greater than that of cSS (0.68 forfSSQFR and 0.56 for cSS; p ¼ 0.002). These resultssuggest that risk classification based on cSS inpatients with 3VD can be properly reclassified byusing QFR.

Risk assessment of patients with 3VD is of para-mount important because the assessment influencesheart team decision making, such as whether toperform PCI or surgical treatment (16,24). Physiolog-ical assessment before revascularization can alter notonly treatment selection between PCI or surgery butalso the number of treated lesions in patients withmultivessel disease. In the total population of theSYNTAX II trial (n ¼ 454), pressure wire–based phys-iological interrogation of the 3 coronary vessels,performed in 82.8% of the patients, reduced thenumber of 3-vessel interventions to 37.2% (8).

However, physiological assessment with a pressurewire has several limitations, such as cost, time, andrelated complications. Angiography-derived FFRpotentially reduces these limitations. In the FAVOREurope-Japan trial, prospectively investigating thefeasibility and diagnostic performance of online QFR,

TABLE 4 Reclassification Table of Functional SYNTAX Scores for the 2-Year Patient-Oriented Composite Outcome

Anatomic SYNTAX Score Event

fSSQFR fSSiFR/FFR

<17 17–22 >22 % Reclassified <17 17–22 >22 % Reclassified

<17 Yes 3 0 0 0 3 0 0 0No 52 0 0 0 52 0 0 0

17–22 Yes 0 4 0 0 1 3 0 25No 20 17 0 54 17 20 0 46

>22 Yes 0 0 4 0 0 0 4 0No 12 9 17 55 8 11 19 50

fSSiFR/FFR ¼ functional SYNTAX (Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery) score derived from hybrid instantaneous wave-free ratioand fractional flow reserve assessment; fSSQFR ¼ functional SYNTAX score derived from quantitative flow ratio.

Asano et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 3 , 2 0 1 9

The SYNTAX II QFR Substudy F E B R U A R Y 1 1 , 2 0 1 9 : 2 5 9 – 7 0

268

the procedure time for QFR assessment was shorterthan that for FFR with intravenous or intracoronaryadenosine (5.0 min [IQR: 3.5 to 6.1 min] for QFR vs.7.0 min [IQR: 5.0 to 10.0 min]; p < 0.001) (25). WhenQFR is applied to functional assessment of multiplevessels for risk assessment of a patient, the advantageof angiography-derived functional assessment maybe enhanced with fewer complications, lower cost,and short procedure time.

Furthermore, the residual SYNTAX score after PCIwas reported to have a substantial impact on long-term clinical outcome (26). The improved discrimi-nant capability of the residual “functional” SYNTAX

FIGURE 7 Receiver-Operating Characteristic Curves of Classic Anato

Quantitative Flow Ratio, and Functional SYNTAX Score Derived From

Predicting 2-Year Patient-Oriented Composite Endpoint

AUC ¼ area under the curve; cSS ¼ classic anatomic SYNTAX (Synergy B

Surgery) score; FFR ¼ fractional flow reserve; fSSiFR/FFR ¼ functional SY

fractional flow reserve assessment; fSSQFR ¼ functional SYNTAX score d

score on the basis of FFR for clinical outcome wasreported in comparison with anatomic or physiolog-ical assessment alone (27). The prognostic value ofthe residual SYNTAX score combined with functionalassessment on the basis of angiography-derivedphysiological index will be investigated in futurestudies.

STUDY LIMITATIONS. The present study was aretrospective and non-pre-specified analysis. Theangiography was not prospectively acquired ac-cording to specific acquisition protocol to fulfillthe technical requirement of QFR analysis. The

mic SYNTAX Score, Functional SYNTAX Score Derived From

Instantaneous Wave-Free Ratio/Fractional Flow Reserve

etween Percutaneous Coronary Intervention With Taxus and Cardiac

NTAX score derived from hybrid instantaneous wave-free ratio/

erived from QFR.

PERSPECTIVES

WHAT IS KNOWN? Functional assessment refines the prog-

nostic value of the SYNTAX score only on the basis of anatomic

J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 3 , 2 0 1 9 Asano et al.F E B R U A R Y 1 1 , 2 0 1 9 : 2 5 9 – 7 0 The SYNTAX II QFR Substudy

269

iFR/FFR measurement site, which was used forthe endpoint of QFR analysis, was filmed only in59.0%, which could be cause of the discrepancybetween QFR and iFR/FR because of anatomicmismatch.

information in patients with 3VD. QFR is a novel angiography-

derived and therefore less invasive functional index that has

demonstrated substantial diagnostic accuracy with FFR as a

reference in patients with relatively simple lesions.

WHAT IS NEW? This is the first study investigating the

diagnostic performance of QFR compared with iFR in patients

with 3VD and the impact of the fSSQFR on clinical events.

According to the results of the present study, QFR yielded

substantial diagnostic accuracy in those patients, although

diagnostic performance was impaired in bifurcated and trifur-

cated lesions or lesions in small vessels. The fSSQFR improved risk

classification for the 2-year POCE compared with the cSS.

WHAT IS NEXT? Risk assessment using QFR may influence the

treatment decision for patients with 3VD. It is warranted to

investigate the impact of the fSSQFR on the treatment decision of

a heart team for patients with 3VD using angiographic acquisition

guidelines for QFR.

CONCLUSIONS

In the present analysis, investigating patientswith 3VD without acquisition guideline, the feasi-bility of QFR was achieved in 71% of lesions and in28.2% of these patients (feasible in the entirecoronary tree). QFR demonstrated substantialdiagnostic performance with high PPV, but lowNPV was observed, especially for lesions in smallvessels (RVD #2.25 mm) and bifurcation or trifur-cation. The fSSQFR has the potential to furtherrefine prognostic risk estimation in a less invasivefashion.

ADDRESS FOR CORRESPONDENCE: Prof. Patrick W.Serruys, ThoraxCenter, Erasmus Medical Center,Westblaak 98, 3012 KM Rotterdam, the Netherlands.E-mail: patrick.w.j.c.serruys@gmail.com.

RE F E RENCE S

1. Bech GJ, De Bruyne B, Pijls NH, et al. Frac-tional flow reserve to determine the appropri-ateness of angioplasty in moderate coronarystenosis: a randomized trial. Circulation 2001;103:2928–34.

2. Tonino PA, De Bruyne B, Pijls NH, et al. Frac-tional flow reserve versus angiography for guidingpercutaneous coronary intervention. N Engl J Med2009;360:213–24.

3. Fearon WF, Nishi T, De Bruyne B, et al. Clinicaloutcomes and cost-effectiveness of fractionalflow reserve-guided percutaneous coronaryintervention in patients with stable coronary ar-tery disease: three-year follow-up of the FAME 2trial (Fractional Flow Reserve Versus Angiographyfor Multivessel Evaluation). Circulation 2018;137:480–7.

4. Davies JE, Sen S, Dehbi HM, et al. Use of theinstantaneous wave-free ratio or fractional flowreserve in PCI. N Engl J Med 2017;376:1824–34.

5. Gotberg M, Christiansen EH, Gudmundsdottir IJ,et al. Instantaneous wave-free ratio versus frac-tional flow reserve to guide PCI. N Engl J Med2017;376:1813–23.

6. Authors/Task Force Members, Windecker S,Kolh P, et al. 2014 ESC/EACTS guidelines onmyocardial revascularization: the Task Force onMyocardial Revascularization of the EuropeanSociety of Cardiology (ESC) and the EuropeanAssociation for Cardio-Thoracic Surgery(EACTS) developed with the special contribu-tion of the European Association of

Percutaneous Cardiovascular Interventions(EAPCI). Eur Heart J 2014;35:2541–619.

7. Fihn SD, Gardin JM, Abrams J, et al. 2012 ACCF/AHA/ACP/AATS/PCNA/SCAI/STS guideline for thediagnosis and management of patients with stableischemic heart disease: a report of the AmericanCollege of Cardiology Foundation/American HeartAssociation Task Force on Practice Guidelines, andthe American College of Physicians, American As-sociation for Thoracic Surgery, Preventive Cardio-vascular Nurses Association, Society forCardiovascular Angiography and Interventions,and Society of Thoracic Surgeons. J Am Coll Car-diol 2012;60:e44–164.

8. Escaned J, Collet C, Ryan N, et al. Clinical out-comes of state-of-the-art percutaneous coronaryrevascularization in patients with de novo threevessel disease: 1-year results of the SYNTAX IIstudy. Eur Heart J 2017;38:3124–34.

9. Serruys PW, Morice MC, Kappetein AP, et al.Percutaneous coronary intervention versuscoronary-artery bypass grafting for severe coro-nary artery disease. N Engl J Med 2009;360:961–72.

10. Tu S, Westra J, Yang J, et al. Diagnostic ac-curacy of fast computational approaches to derivefractional flow reserve from diagnostic coronaryangiography: the international multicenter FAVORpilot study. J Am Coll Cardiol Intv 2016;9:2024–35.

11. Papafaklis MI, Muramatsu T, Ishibashi Y, et al.Fast virtual functional assessment of intermediate

coronary lesions using routine angiographic dataand blood flow simulation in humans: comparisonwith pressure wire – fractional flow reserve.EuroIntervention 2014;10:574–83.

12. Xu B, Tu S, Qiao S, et al. Diagnostic accuracy ofangiography-based quantitative flow ratio mea-surements for online assessment of coronary ste-nosis. J Am Coll Cardiol 2017;70:3077–87.

13. Westra J, Tu S, Winther S, et al. Evaluation ofcoronary artery stenosis by quantitative flow ratioduring invasive coronary angiography: the WIFI IIstudy (Wire-Free Functional Imaging II). Circ Car-diovasc Imaging 2018;11:e007107.

14. Nam CW, Mangiacapra F, Entjes R, et al.Functional SYNTAX score for risk assessment inmultivessel coronary artery disease. J Am CollCardiol 2011;58:1211–8.

15. Escaned J, Banning A, Farooq V, et al. Ratio-nale and design of the SYNTAX II trial evaluatingthe short to long-term outcomes of state-of-the-art percutaneous coronary revascularisation inpatients with de novo three-vessel disease. Euro-Intervention 2016;12:e224–34.

16. Farooq V, van Klaveren D, Steyerberg EW,et al. Anatomical and clinical characteristics toguide decision making between coronary arterybypass surgery and percutaneous coronary inter-vention for individual patients: development andvalidation of SYNTAX score II. Lancet 2013;381:639–50.

17. Tu S, Barbato E, Koszegi Z, et al. Fractionalflow reserve calculation from 3-dimensional

Asano et al. J A C C : C A R D I O V A S C U L A R I N T E R V E N T I O N S V O L . 1 2 , N O . 3 , 2 0 1 9

The SYNTAX II QFR Substudy F E B R U A R Y 1 1 , 2 0 1 9 : 2 5 9 – 7 0

270

quantitative coronary angiography and TIMI framecount: a fast computer model to quantify thefunctional significance of moderately obstructedcoronary arteries. J Am Coll Cardiol Intv 2014;7:768–77.

18. Gibson CM, Cannon CP, Daley WL, et al. TIMIframe count: a quantitative method of assessingcoronary artery flow. Circulation 1996;93:879–88.

19. Kikuta Y, Cook CM, Sharp ASP, et al. Pre-angioplasty instantaneous wave-free ratiopullback predicts hemodynamic outcome inhumans with coronary artery disease: primaryresults of the international multicenter iFRGRADIENT registry. J Am Coll Cardiol Intv2018;11:757–67.

20. Passing H, Bablok. A new biometrical pro-cedure for testing the equality of measurementsfrom two different analytical methods. Applicationof linear regression procedures for method com-parison studies in clinical chemistry, part I. J ClinChem Clin Biochem 1983;21:709–20.

21. Bland JM, Altman DG. Statistical methodsfor assessing agreement between two methodsof clinical measurement. Lancet 1986;1:307–10.

22. Min JK, Leipsic J, Pencina MJ, et al. Diag-nostic accuracy of fractional flow reserve fromanatomic CT angiography. JAMA 2012;308:1237–45.

23. Grunkemeier GL, Jin R. Net reclassification in-dex: measuring the incremental value of adding anew risk factor to an existing risk model. AnnThorac Surg 2015;99:388–92.

24. Mohr FW, Morice MC, Kappetein AP, et al.Coronary artery bypass graft surgery versuspercutaneous coronary intervention in patientswith three-vessel disease and left main coro-nary disease: 5-year follow-up of the rando-mised, clinical SYNTAX trial. Lancet 2013;381:629–38.

25. Westra J, Andersen BK, Campo G, et al. Diag-nostic performance of in-procedure angiography-derived quantitative flow reserve compared to

pressure-derived fractional flow reserve: theFAVOR II Europe-Japan study. J Am Heart Assoc2018;7.

26. Genereux P, Palmerini T, Caixeta A, et al.Quantification and impact of untreated coronaryartery disease after percutaneous coronary inter-vention: the residual SYNTAX (Synergy BetweenPCI with Taxus and Cardiac Surgery) score. J AmColl Cardiol 2012;59:2165–74.

27. Choi KH, Lee JM, Koo BK, et al. Prognosticimplication of functional incomplete revasculari-zation and residual functional SYNTAX score inpatients with coronary artery disease. J Am CollCardiol Intv 2018;11:237–45.

KEY WORDS 3-vessel disease, functionalSYNTAX score, quantitative flow ratio

APPENDIX For supplemental methods,figures, and tables, please see the onlineversion of this paper.