ORIGINAL RESEARCH

published: 12 October 2021

doi: 10.3389/fcvm.2021.714077

Diagnostic Performance of
Angiography-Based Fractional Flow
Reserve for Functional Evaluation of
Coronary Artery Stenosis
Changling Li 1 , Xiaochang Leng 2 , Jingsong He 2 , Yongqing Xia 2 , Wenbing Jiang 3 ,
Yibin Pan 4 , Liang Dong 1 , Yong Sun 1 , Xinyang Hu 1 , Jian’an Wang 1 , Jianping Xiang 2* and
Jun Jiang 1*
1
Department of Cardiology, The Second Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China,
2
ArteryFlow Technology Co., Ltd., Hangzhou, China, 3 Department of Cardiology, The Third Clinical Institute Affiliated to
Wenzhou Medical University, Wenzhou, China, 4 Department of Cardiology, Affiliated Jinhua Hospital, Zhejiang University
School of Medicine, Jinhua, China

Edited by: Background: A new method for calculating fraction flow reserve (FFR) without pressure-
Sebastian Kelle,
wire (angiography-derived FFR) based on invasive coronary angiography (ICA) images
Deutsches Herzzentrum
Berlin, Germany can be used to evaluate the functional problems of coronary stenosis.
Reviewed by: Objective: The aim of this study was to assess the diagnostic performance of a
Jan Brüning,
Charité—Universitätsmedizin
novel method of calculating the FFR compared to wire-based FFR using retrospectively
Berlin, Germany collected data from patients with stable angina.
Haikun Qi,
ShanghaiTech University, China Methods: Three hundred patients with stable angina pectoris who underwent ICA
Masafumi Ono, and FFR measurement were included in this study. Two ICA images with projections
Cardialysis, Netherlands
>25◦ apart at the end-diastolic frame were selected for 3D reconstruction. Then, the
*Correspondence:
Jun Jiang
contrast frame count was performed in an angiographic run to calculate the flow velocity.
[email protected] Based on the segmented vessel, calculated velocity, and aortic pressure, AccuFFRangio
Jianping Xiang
distribution was calculated through the pressure drop equation.
[email protected]
Results: Using FFR ≤ 0.8 as a reference, we evaluated AccuFFRangio performance for
Specialty section: 300 patients with its accuracy, sensitivity, specificity, positive predictive value (PPV), and
This article was submitted to
Cardiovascular Imaging,
negative predictive value (NPV). Comparison of AccuFFRangio with wire-measured FFR
a section of the journal resulted in an area under the curve (AUC) of 0.954 (per-vessel, p < 0.0001). Accuracy for
Frontiers in Cardiovascular Medicine
AccuFFRangio was 93.7% for Pa set from measurement and 87% for Pa = 100 mmHg
Received: 24 May 2021
in this clinical study. Overall sensitivity, specificity, PPV, and NPV for per-vessel were 90,
Accepted: 15 September 2021
Published: 12 October 2021 95, 86.7, 96.3, and 57.5, 97.7, 90.2, 86.3%, respectively. Overall accuracy, sensitivity,
Citation: specificity, PPV, and NPV for 2-dimensional (2D) quantitative coronary angiography (QCA)
Li C, Leng X, He J, Xia Y, Jiang W, were 63.3, 42.5, 70.9, 34.7, and 77.2%, respectively. The average processing time of
Pan Y, Dong L, Sun Y, Hu X, Wang J,
Xiang J and Jiang J (2021) Diagnostic
AccuFFRangio was 4.30 ± 1.87 min.
Performance of Angiography-Based Conclusions: AccuFFRangio computed from coronary ICA images can be an accurate
Fractional Flow Reserve for Functional
Evaluation of Coronary Artery and time-efficient computational tool for detecting lesion-specific ischemia of coronary
Stenosis. artery stenosis.
Front. Cardiovasc. Med. 8:714077.
doi: 10.3389/fcvm.2021.714077 Keywords: fractional flow reserve, invasive coronary angiography, coronary artery, ischemia, stenosis

Frontiers in Cardiovascular Medicine | www.frontiersin.org 1 October 2021 | Volume 8 | Article 714077

Li et al. Diagnostic Performance of Angio-Based FFR

INTRODUCTION numerical calculation of pressure drop, angio-based FFR enables

interventional cardiologists and researchers to obtain accurate
Compared with the anatomical stenosis of the coronary anatomical quantifications of one or more lesions in the analyzed
artery, functional assessment can more accurately evaluate and coronary segment, to determine the functional significance of
predict the progression of coronary heart disease (1). In the the individual and consecutive multiple lesions. These methods
catheterization laboratory, invasive coronary angiography (ICA) can be helpful for optimal percutaneous coronary intervention
images can only qualitatively assess the degree of stenosis but (PCI) treatment of the lesion of coronary disease. Several studies
cannot evaluate the physiological function of coronary arteries. have shown that angio-based FFR is highly correlated with
Therefore, it may overestimate or underestimate the severity of invasive FFR compared to coronary computed tomography
the disease, leading to the untreated or over-treatment of lesions angiography (CTA) and ICA assessment (4, 8–10). It is also more
(2). Fractional flow reserve (FFR) has become a recognized advantageous in formulating treatment strategies for coronary
index for the functional evaluation of coronary stenosis, which artery disease under circumstances that screening people with
is defined as a ratio of the pressure of the distal end of the suspected chest pain for the presence of myocardial ischemia.
stenosis and the cardiac aorta at hyperemia (1). The current In this study, coronary angiography was used to calculate the
method of measuring FFR requires a pressure wire inserted average volume flow using TIMI (thrombolysis in myocardial
into the distal end of the stenosis, which will bring additional infarction) frame count combined with three-dimensional
procedure-related risks causing adverse effects to the blood quantitative coronary angiography (QCA). Subsequently,
vessel and increase the treatment time and cost (3, 4). A new applying computational fluid dynamics theory, a new
method of non-pressure wire FFR (angio-based FFR) calculation angiography-based FFR calculation method AccuFFRangio
method based on ICA images can reflect functional problems was proposed. The FFR measured by the pressure-wire was used
of coronary stenosis (5–7). By using two angiograms greater as a reference standard to evaluate the diagnostic performance
than 25◦ through independent 3D vessel reconstruction and of AccuFFRangio.

FIGURE 1 | AccuFFRangio analysis of intermediate stenosis of a left anterior descending artery (LAD). (a,b) Coronary angiography images from two different angles of
view. (c) The above chart shows the change of lumen diameter (red) along LAD with the computed reference diameter (green); the chart below shows the diameter
stenosis (red) and AccuFFRangio pullback (green). (d) Computed AccuFFRangio distribution; the AccuFFRangio value was 0.86. (e) The FFR measured by pressure
wire was 0.85.

Frontiers in Cardiovascular Medicine | www.frontiersin.org 2 October 2021 | Volume 8 | Article 714077

Li et al. Diagnostic Performance of Angio-Based FFR

MATERIALS AND METHODS Invasive Coronary Angiography and

Study Design 2D-QCA Analysis
The present study is a retrospective, single-center, observational ICA was performed using the X-ray system (Allura Xper
study performed at The Second Affiliated Hospital, Zhejiang FD20/10; PHILIPS Medical Systems, the Netherlands). These
University School of Medicine. This study aims to evaluate the angiographic images were recorded at 15 frames/s. The contrast
diagnostic accuracy, sensitivity, and specificity of AccuFFRangio medium was injected manually with a forceful and stable
in identifying functionally significant stenosis by using pressure injection or by the pump at a rate of ∼4 ml/s. 2D-QCA
wire-based FFR as the reference. AccuFFRangio and 2D-QCA was conducted by using angiogram vendor-integrated QCA
were analyzed and compared in the core laboratory of the software (Allura Xper FD20/10; PHILIPS Medical Systems,
Department of Cardiology at The Second Affiliated Hospital, the Netherlands).
Zhejiang University School of Medicine. After receiving ethics
approval from the institutional review board, this study was Wire-Based FFR Measurement
conducted with a written informed consent form waived. FFR was measured in all patients using coronary pressure
wire (St. Jude Medical, St. Paul, MN, USA). After calibration
Patient Population and equalization, the pressure wire was advanced distally
Since this was a retrospective study, consecutive patients to the stenosis. Maximum hyperemia was induced with i.v.
with stable angina pectoris who underwent ICA and FFR adenosine triphosphate at a concentration of 180 µg/kg/min.
measurement were eligible for enrollment. Angiographic Both the distal coronary pressure at the pressure sensor and
inclusion criteria were (1) percentage diameter stenosis of the the proximal pressure at the coronary artery ostium were
coronary artery between 30% and 90% in a vessel ≥2 mm by recorded simultaneously. The FFR measurement was performed
visual estimation; (2) angiographic projections ≥25◦ apart were by physicians in The Second Affiliated Hospital, Zhejiang
recorded. Exclusion criteria include (1) overlapping interrogated University School of Medicine (Y.P., L.D., W.J., Y.S.). Pressure
vessels with too much shortening without preferred references sensor was pulled back to the catheter tip to check or correct the
in proximal or distal vessels; (2) insufficient injected contrast for pressure drift (Figure 1e).
QCA analysis; (3) location of the target lesion at the ostium of the
left or right coronary artery; (4) wire-position not documented. AccuFFRangio Computation
Exclusion criteria on patient level contain (1) acute myocardial AccuFFRagnio was computed with the AccuFFRangio V1.0
infarction within 72 h; (2) severe asthma or severe chronic software (ArteryFlow Technology, Hangzhou, China) by
obstructive pulmonary disease; (3) allergy to contrast media or participating physicians and technicians (F.M., Y.Z., M.H.)
adenosine; or (4) atrial fibrillation. blinded to FFR. Two angiographic images with projections

FIGURE 2 | Study enrollment flow chart.

Frontiers in Cardiovascular Medicine | www.frontiersin.org 3 October 2021 | Volume 8 | Article 714077

Li et al. Diagnostic Performance of Angio-Based FFR

>25◦ apart at the end-diastolic frame were selected for Contrast flow rate velocity for FFR computation was derived
three-dimensional (3D) reconstruction (6, 7, 11). To simplify from the TIMI frame counting method for the segmented
the geometry calibration procedure and achieve a reliable vessel. With the calculated velocity and input aortic pressure
correspondence in centerline points for 3D reconstruction, we from the measurement of the pressure at the coronary
have implemented three pairs of reference points to eliminate ostium, AccuFFRangio distribution can be calculated (Figure 1).
the isocenter offset and rotational angle parameter errors. Since AccuFFRangio value was taken at the same position of wire-
the pressure drop has a positive relationship with the coronary based FFR using angiography images as a reference. To
vessel flow rate, the frame count method is a relatively feasible compare the diagnostic accuracy of AccuFFRangio by using
solution (12). This method hypothesizes that the blood flow a fixed value of aortic pressure and to study the influence
velocity is proportional to the vessel cross-section diameter of fixed value on the performance of our approach in case
dimension. Typically, the pressure drop from proximal to distal some patient-specific pressures cannot be obtained, Pa =
stenosis is caused by two factors. The first is the viscous pressure 100 mmHg was set for each calculation of angio-based FFR
drop associated with friction. The second is the expansion (13, 14).
pressure drop due to the rapid change in radius, which is
usually characterized by narrowing. Pressure drop PR is related Statistical Analysis
to viscosity loss coefficient CVis , expansion loss coefficient Continuous variables with normally distributed were expressed
CExpan , and flow rate Q: PR = (CVis + CExpan • Q) • Q. More as mean ± standard deviation (SD) and non-normal distributed
detail of the derived equations can be seen in our previous variables as the median. Categorical variables were expressed as
study (11). percentages and data were analyzed on a per vessel basis. Pearson

FIGURE 3 | Percent distribution of invasive FFR.

Frontiers in Cardiovascular Medicine | www.frontiersin.org 4 October 2021 | Volume 8 | Article 714077

Li et al. Diagnostic Performance of Angio-Based FFR

correlation was used to quantify the correlation between FFR TABLE 1 | Baseline patient characteristics (n = 300).
and AccuFFRangio. Agreement between FFR and AccuFFRangio
Age, y 64.1 ± 9.6
was assessed on the Bland-Altman plot. Using FFR ≤ 0.8 as
Male 67% (201)
the reference standard, the performance of AccuFFRangio for
Weight (kg) 68.5 ± 34.5
predicting functionally significant stenosis was evaluated by
Height (cm) 165 ± 7.3
diagnostic accuracy, sensitivity, specificity, positive predictive
BMI, kg/m2 25.2 ± 13.9
value (PPV), and negative predictive value (NPV). The area
Systolic blood pressure (mm Hg) 130 ± 20
under the curve (AUC) of receiver operating characteristic
Diastolic blood pressure (mm Hg) 78 ± 15
(ROC) analysis was used to assess the diagnostic accuracy of
AccuFFRangio. All the statistical analyses were performed by Diabetes 21% (64)

using MedCalc (MedCalc Software Inc., Belgium). Hypertension 45% (135)

Hyperlipidemia 13% (40)

RESULTS
Patient Characteristics TABLE 2 | Vessel characteristics (n = 300).
Figure 2 presents the study enrollment flow chart. Three
hundred eighteen patients with 318 vessels were included in Vessels

this clinical study from May 2016 to July 2019. Due to the LAD 61.7% (185)
incomplete data from six patients, 312 patients underwent ICA LCX 7.3% (22)
procedure and wire-FFR waveform analysis. Among them, 12 RCA 29.7% (89)
patients were excluded due to predefined exclusion criteria, Anatomy
including undocumented wire-position, poor image quality, Diameter stenosis, % 44 ± 12%
excessive vessel overlap, insufficient contrast, projections <25 <50% 67.3% (202)
degrees, excessive pressure wire drift, and left main coronary ≥50% 32.7% (98)
artery lesions. In the end, 300 patients with 300 vessels were Physiology
included in the final analysis. FFR (per vessel) 0.84 ± 0.10
Mean FFR was 0.84 ± 0.10, as shown in Figure 3, and mean Vessels with FFR ≤ 0.8 26.7% (80)
percentage diameter stenosis (%DS) form 2D-QCA was 44 ± Vessels with FFR > 0.8 73.3% (220)
12%. FFR ≤ 0.80 was found in 80 (26.7%) vessels and the mean Diffuse or serial lesions 32.3% (97)
contrast flow rate velocity was 0.17 ± 0.05 m/s. Baseline patient Bifurcation lesions 2.7% (8)
and procedural characteristics are listed in Tables 1, 2. Calcified lesions 2% (6)
Myocardial bridge 5.7% (17)
Correlation and Agreement Between
AccuFFRangio and FFR
Good correlations were observed in Figure 4 with a correlation
coefficient of r = 0.83 (p < 0.001). There were good agreements
In addition, the mean processing time of AccuFFRangio
between AccuFFRangio and FFR in the Bland-Altman plot with
assessment was 4.30 ± 1.87 min including the 3D anatomic
a mean difference value of −0.001 (limits of agreement: −0.124
model reconstruction and AccuFFRangio calculation for
to 0.122) when Pa measured at the coronary ostium and −0.030
each patient.
(limits of agreement: −0.155 to 0.095) when Pa = 100 mmHg
was used, as shown in Figure 5. The number of patients with the
absolute difference between AccuFFRangio and FFR falling out of
the 95% CI was 9 (3%) when Pa was set according to the patients DISCUSSION
and 18 (6%) when Pa was set equal to a fixed value.
Wire-based FFR has potential risks in measurement procedures
Diagnostic Performance of AccuFFRangio and vasodilator complications, and the complexity of the
and 2D-QCA operation is also a challenge. In this situation, this study had
Accuracy for AccuFFRangio was 93.7% in this clinical study. demonstrated a reliable and efficient computational method
Overall sensitivity, specificity, PPV, and NPV were 90, 95, 86.7, AccuFFRangio for the functional evaluation of lesion-specific
and 96.3%, respectively (Table 3). Meanwhile, these values for ischemia of coronary artery stenosis based on ICA images
AccuFFRangio, when Pa = 100 mmHg was implemented in the without injecting vasodilators. Thus, instead of using invasive
calculation, were 87, 57.5, 97.7, 90.2, and 86.3%. Comparison of wire-based FFR for evaluating the severity of suspected coronary
AccuFFRangio and 2D-QCA with pressure wire measured FFR as heart disease, AccuFFRangio uses a combination of the 3D
reference resulted in an AUC for AccuFFRangio of 0.954 (95%CI: structure of the coronary vessel and computational fluid
0.924–0.975) and 0.934 (95%CI: 0.900–0.960, when Pa = 100 dynamics (CFD)-based equations on account of TIMI frame
mmHg) and 2D-QCA of 0.567 (95% CI: 0.509–0.624), as shown count to analyze the functional performance in a short time
in Figure 6. of 5 min. The diagnostic accuracy of AccuFFRangio was 93.7%

Frontiers in Cardiovascular Medicine | www.frontiersin.org 5 October 2021 | Volume 8 | Article 714077

Li et al. Diagnostic Performance of Angio-Based FFR

FIGURE 4 | Correlation between AccuFFRangio computation and conventional pressure wired measured FFR. (A) Pa from the measurement of the pressure at the
coronary ostium; (B) Pa set as equal to 100 mmHg.

FIGURE 5 | Agreement between AccuFFRangio computation and conventional pressure wired measured FFR. (A) Pa from the measurement of the pressure at the
coronary ostium; (B) Pa set as equal to 100 mmHg.

compared to pressure wire-derived FFR, which shows a higher coronary tree construction based on the geometry of two
accuracy compared to 2D-QCA with an accuracy of 63.3%. or more projections with a minimum separation of 30◦
For assessment of FFR without pressure-wire, many and application of an automatic resistance-based lumped
research groups have made significant efforts and developed model of the entire coronary tree, FFRangio (17–19) showed a
different angiography-based FFR methods. Morris et al. high concordance between pressure-wire measured FFR. By
(15) described that the construction of arteries was based reconstructing a 3D QCA model of the target vessel using two
on two projections from similar phases of the cardiac cycle angiographic projections recorded at least 25◦ intervals, the
with good vessel opacification and contrast. Meanwhile, QFR was computed through mathematic equations incorporated
the virtual FFR was calculated from CFD simulations with with contrast flow velocity determined using frame count
generic downstream boundary conditions applied to the analysis (6, 20). It represented a high diagnostic accuracy
arterial outlet with a Windkessel model (16). With a 3D with FFR, reducing the number of patients for pressure-wire

Frontiers in Cardiovascular Medicine | www.frontiersin.org 6 October 2021 | Volume 8 | Article 714077

Li et al. Diagnostic Performance of Angio-Based FFR

TABLE 3 | Diagnostic performance of AccuFFRangio for per-vessel.

AccuFFRangio ≤ 0.8 AccuFFRangio ≤ 0.8 (Pa = 100 mmHg) Diameter stenosis by QCA ≥ 50%

Accuracy 93.7% (89.9–95.9%) 87% (81.9–90.0%) 63.3% (57.94–69.1%)

Sensitivity 90.0% (84.6–97.2%) 57.5% (45.3–67.8%) 42.5% (33.8–56.5%)
Specificity 95.0% (89.5–96.5%) 97.7% (94.1–99.0%) 70.9% (63.9–76.4%)
PPV 86.7% (76.3–89.9%) 90.2% (77.3–94.5%) 34.7% (28.8–43.2%)
NPV 96.3% (94.1–98.7%) 86.3% (82.6–88.7%) 77.2% (73.9–81.4%)

Data are shown in percentage with 95% confidence interval in brackets.

FIGURE 6 | ROC Curve between AccuFFRangio and QCA.

measurements. In the current study of AccuFFRangio, the The accuracy of AccuFFRangio in the present study was 93.7%
3D geometry model construction and calculation of FFR for per-vessel bias, which is comparable to the clinical trials
were different from the methods described above. We used with patients over 200, such as 83% for WIFI II Study (21),
three physiological points to do the vessel calibration to 86.8% for FAVOR II Europe-Japan Study (7), 92.7% for FAVOR
eliminate the error during the reprojection procedure, as the II China Study (4), and 93.5% for FAST-FFR Study (18). The
same processes in our previous studies (11). Moreover, the sensitivity, specificity, and AUC for the four clinical trials were
velocity of the inlet of the blood vessel was set according to 86, 77%, and 0.86 for WIFI II Study (21), 86.5, 86.9%, and 0.92
the TIMI frame count. The blood pressure at the aorta was for FAVOR II Europe-Japan Study (7), 94.6, 91.7%, and 0.96
chosen equal to the value from measuring the pressure at the for FAVOR II China Study (4), and 91.2, 92.2%, and 0.944 for
coronary ostium. FAST-FFR Study (18). Those for AccuFFRangio were 90, 95%,

Frontiers in Cardiovascular Medicine | www.frontiersin.org 7 October 2021 | Volume 8 | Article 714077

Li et al. Diagnostic Performance of Angio-Based FFR

and 0.954. Compared to Pa taken from the measurement at the 93.7, 90, and 95%, respectively. Those were better than the
coronary ostium, as Pa set equal to a fixed value of 100 mmHg, diagnostic performance of AccuFFRangio calculated based on Pa
the diagnostic performance decreased to 87% for accuracy setting equal to 100 mmHg. AccuFFRangio bears the potential
and 57.5% for sensitivity, respectively. Angiography-based FFR of improving angiography-based identification of functionally
can improve the low sensitivity of 2D-QCA in evaluating significant stenosis during coronary angiography procedure.
hemodynamically significant of coronary stenosis, from about
42.5–62.5–86.5–94.6% (4, 7, 21–23). Similarly, this method will Impact on Daily Practice
also increase the specificity from the original 58.1–76.5–86.9– AccuFFRangio can quickly and accurately calculate FFR values
92.2% (4, 7, 21–23). Thus, the implementation of angiography- based on coronary angiography images and can be used for
based FFR can avoid unnecessary revascularization of many functional assessment of patients with coronary heart disease,
interrogated vessels when performed coronary angiography. It is avoiding unnecessary PCI treatment.
also useful to optimize the strategies of percutaneous coronary
intervention (PCI) to reduce the number of the implanted stents
and improve the clinical outcome for patients who plan to DATA AVAILABILITY STATEMENT
perform PCI.
The time for calculating angiography-based FFR is also The original contributions presented in the study are included
essential for evaluating superiority when there is only limited in the article/supplementary material, further inquiries can be
time during the PCI operation. For vFFR (15), it took ∼ directed to the corresponding authors.
24 h for the CFD simulation of one case, which cannot be
implemented in the condition during PCI performance. Another ETHICS STATEMENT
method FFRangio took nearly 10 min for the whole procedure,
including reconstruction of the 3D geometry model of the entire The studies involving human participants were reviewed and
coronary tree and calculation of the FFR values based on lumped approved by Department of Cardiology, Second Affiliated
model (17, 18, 24, 25). But, it only took 4.3 ± 3.4 min to Hospital of Zhejiang University School of Medicine. Written
perform an analysis for one lesion (24). By constructing the informed consent for participation was not required for this
3D geometry model for only the target vessel and CFD-based study in accordance with the national legislation and the
equations on account of TIMI frame count to calculate the institutional requirements.
FFR values, the entire procedure was completed in a short time
of 5 min 59 s on average (6), which can be used during the
PCI operation. In addition, the entire process of calculating AUTHOR CONTRIBUTIONS
FFR with a similar algorithm used in this paper took about
CL and XL: concept and design of the study, analysis of the
5 min, which met the requirement of clinical application in
data, and drafting of the manuscript. JH: case calculation and
PCI surgery.
drafting of the manuscript. CL, LD, and XH: image annotation
It is worth noting the limitations of this clinical study.
of ICA and critical review of the manuscript. YX and JH:
Firstly, this was a retrospective study at one single center.
implementation of the algorithm and drafting of the manuscript.
Secondly, the study may have selection bias because of
JW and JX: conception and design of the study, analysis of
the relatively small number of positive cases (80 vessels,
the data, drafting of the manuscript, and final approval of the
26.7%) compared with the negative ones (220 vessels, 73.3%).
manuscript submitted. WJ, YP, YS, and JJ: have contributed to
Third, this study was an observational study. In the future,
the submitted work. All authors contributed to the article and
prospective, multi-center, and follow-up studies will be
approved the submitted version.
performed in the post-market clinical researches. Fourth,
abnormal pressure curves such as wave form distortion or
ventricularization were not found in this study due to the FUNDING
nature of our study population, while this could influence
the measurement of FFR; thus, further assessment should This work was supported by the National Natural Science
be considered. Foundation of China (Nos. 82170332, 81320108003, 31371498,
81100141, and 81570322), Zhejiang Provincial Public
CONCLUSIONS Welfare Technology Research Project (No. LGF20H020012),
Zhejiang Provincial key research and development plan
This clinical study demonstrates that AccuFFRangio is clinically (No. 2020C03016), the Major projects in Wenzhou of China
feasible. The performance is superior to angiographic assessment (No. 2019ZG0107), the Major projects in Jinhua of China
by 2D-QCA for evaluating coronary artery stenosis when (No. 2020A31003), Scientific research project of Zhejiang
using FFR as a reference. The accuracy, sensitivity, and Education Department (No. Y201330290), and Major medical
specificity of AccuFFRangio in identifying hemodynamically and health science and technology plan of Zhejiang Province
significant of coronary stenosis using 300 patient data were (No. WKJ-ZJ-1913).

Frontiers in Cardiovascular Medicine | www.frontiersin.org 8 October 2021 | Volume 8 | Article 714077

Li et al. Diagnostic Performance of Angio-Based FFR

REFERENCES Flow Reserve From Coronary Angiography) study. JACC Cardiovascular

Interventions. (2013) 6:149–57. doi: 10.1016/j.jcin.2012.08.024
1. Kern MJ, Samady H. Current concepts of integrated coronary physiology 16. Westerhof N, Lankhaar J-W, Westerhof BE. The arterial Windkessel. Med Biol
in the catheterization laboratory. J Am Coll Cardiol. (2010) 55:173– Eng Comput. (2009) 47:131–41. doi: 10.1007/s11517-008-0359-2
85. doi: 10.1016/j.jacc.2009.06.062 17. Pellicano M, Lavi I, De Bruyne B, Vaknin-Assa H, Assali A,
2. Lindstaedt M, Spiecker M, Perings C, Lawo T, Yazar A, Holland-Letz T, et Valtzer O, et al. Validation study of image-based fractional flow
al. How good are experienced interventional cardiologists at predicting the reserve during coronary angiography. Circ Cardiovasc Interv. (2017)
functional significance of intermediate or equivocal left main coronary artery 10:e005259. doi: 10.1161/CIRCINTERVENTIONS.116.005259
stenoses? Int J Cardiol. (2007) 120:254–61. doi: 10.1016/j.ijcard.2006.11.220 18. Fearon WF, Achenbach S, Engstrom T, Assali A, Shlofmitz R,
3. Douglas PS, Pontone G, Hlatky MA, Patel MR, Norgaard BL, Byrne RA, et Jeremias A, et al. Accuracy of fractional flow reserve derived
al. Clinical outcomes of fractional flow reserve by computed tomographic from coronary angiography. Circulation. (2019) 139:477–
angiography-guided diagnostic strategies vs. usual care in patients with 84. doi: 10.1161/CIRCULATIONAHA.118.037350
suspected coronary artery disease: the prospective longitudinal trial of 19. Kornowski R, Lavi I, Pellicano M, Xaplanteris P, Vaknin-Assa H, Assali A, et
FFRCT: outcome and resource impacts study. Eur Heart J. (2015) 36:3359– al. Fractional Flow Reserve Derived From Routine Coronary Angiograms. J
67. doi: 10.1093/eurheartj/ehv444 Am Coll Cardiol. (2016) 68:2235–7. doi: 10.1016/j.jacc.2016.08.051
4. Xu B, Tu S, Qiao S, Qu X, Chen Y, Yang J, et al. Diagnostic accuracy 20. Smit JM, Koning G, van Rosendael AR, El Mahdiui M, Mertens
of angiography-based quantitative flow ratio measurements for online BJ, Schalij MJ, et al. Referral of patients for fractional flow reserve
assessment of coronary stenosis. J Am Coll Cardiol. (2017) 70:3077– using quantitative flow ratio. Eur Heart J Cardiovasc Imaging.
87. doi: 10.1016/j.jacc.2017.10.035 (2018). doi: 10.1093/eurheartj/ehy563.P4635
5. Zhou J, Yang J-J, Yang X, Chen Z-Y, He B, Du L-S, et al. Impact of 21. Westra J, Tu S, Winther S, Nissen L, Vestergaard MB, Andersen
clinical guideline recommendations on the application of coronary computed BK, et al. Evaluation of coronary artery stenosis by quantitative flow
tomographic angiography in patients with suspected stable coronary artery ratio during invasive coronary angiography: The WIFI II study (wire-
disease. Chin Med J (Engl). (2016) 129:135–41. doi: 10.4103/0366-6999.173434 free functional imaging II). Circulation Cardiovascular Imaging. (2018)
6. Koltowski L, Zaleska M, Maksym J, Tomaniak M, Solinski M, Puchta D, 11:e007107. doi: 10.1161/CIRCIMAGING.117.007107
et al. Quantitative flow ratio derived from diagnostic coronary angiography 22. Tu S, Westra J, Yang J, von Birgelen C, Ferrara A, Pellicano M, et al.
in assessment of patients with intermediate coronary stenosis: a wire- Diagnostic accuracy of fast computational approaches to derive fractional flow
free fractional flow reserve study. Clin Res Cardiol. (2018) 107:858– reserve from diagnostic coronary angiography: the international multicenter
67. doi: 10.1007/s00392-018-1258-7 FAVOR pilot study. JACC Cardiovascular Interventions. (2016) 9:2024–
7. Westra J, Andersen BK, Campo G, Matsuo H, Koltowski L, Eftekhari 35. doi: 10.1016/j.jcin.2016.07.013
A, et al. Diagnostic performance of in-procedure angiography-derived 23. Asano T, Katagiri Y, Collet C, Tenekecioglu E, Miyazaki Y, Sotomi Y,
quantitative flow reserve compared to pressure-derived fractional flow et al. Functional comparison between the BuMA Supreme biodegradable
reserve: The FAVOR II Europe-Japan study. J Am Heart Assoc. (2018) polymer sirolimus-eluting stent and a durable polymer zotarolimus-eluting
7:e009603. doi: 10.1161/JAHA.118.009603 coronary stent using quantitative flow ratio: PIONEER QFR substudy.
8. Min JK, Koo BK, Erglis A, Doh JH, Daniels DV, Jegere S, et al. EuroIntervention. (2018) 14:e570–e9. doi: 10.4244/EIJ-D-17-00461
Usefulness of noninvasive fractional flow reserve computed from coronary 24. Omori H, Witberg G, Kawase Y, Tanigaki T, Okamoto S, Hirata
computed tomographic angiograms for intermediate stenoses confirmed T, et al. Angiogram based fractional flow reserve in patients with
by quantitative coronary angiography. Am J Cardiol. (2012) 110:971– dual/triple vessel coronary artery disease. Int J Cardiol. (2019) 283:17–
6. doi: 10.1016/j.amjcard.2012.05.033 22. doi: 10.1016/j.ijcard.2019.01.072
9. Min JK, Koo B-K, Erglis A, Doh J-H, Daniels DV, Jegere S, et al. Effect of image 25. Kornowski R, Vaknin-Assa H, Assali A, Greenberg G, Valtzer O, Lavi I. Online
quality on diagnostic accuracy of noninvasive fractional flow reserve: Results angiography image-based FFR assessment during coronary catheterization: a
from the prospective multicenter international DISCOVER-FLOW study. J single-center study. J Invasive Cardiol. (2018) 30:224–9.
Cardiovasc Comput Tomogr. (2012) 6:191–9. doi: 10.1016/j.jcct.2012.04.010
10. Nørgaard BL, Leipsic J, Gaur S, Seneviratne S, Ko BS, Ito H, et al. Diagnostic Conflict of Interest: XL is employed by ArteryFlow Technology and receives
performance of noninvasive fractional flow reserve derived from coronary grants from National Natural Science Foundation of China (No. 11802113). JH
computed tomography angiography in suspected coronary artery disease: and YX was employed by ArteryFlow Technology. JX is the CEO of ArteryFlow
the NXT trial (Analysis of Coronary Blood Flow Using CT Angiography: Technology and receives grants from National Natural Science Foundation of
Next Steps). J Am Coll Cardiol. (2014) 63:1145–55. doi: 10.1016/j.jacc.2013. China (No. 81771242).
11.043
11. Jiang J, Feng L, Li C, Xia Y, He J, Leng X, et al. Fractional flow reserve for The remaining authors declare that the research was conducted in the absence of
coronary stenosis assessment derived from fusion of intravascular ultrasound any commercial or financial relationships that could be construed as a potential
and X-ray angiography. Quant Imaging Med Surg. (2021) 11:4543–55. conflict of interest.
doi: 10.21037/qims-20-1324
12. Gibson CM, Cannon CP, Daley WL, Dodge JT. Jr, Alexander B, Jr, Marble SJ, Publisher’s Note: All claims expressed in this article are solely those of the authors
et al. TIMI frame count: a quantitative method of assessing coronary artery and do not necessarily represent those of their affiliated organizations, or those of
flow. Circulation. (1996) 93:879–88. doi: 10.1161/01.CIR.93.5.879
the publisher, the editors and the reviewers. Any product that may be evaluated in
13. Papafaklis MI, Muramatsu T, Ishibashi Y, Lakkas LS, Nakatani S, Bourantas
this article, or claim that may be made by its manufacturer, is not guaranteed or
CV, et al. Fast virtual functional assessment of intermediate coronary lesions
using routine angiographic data and blood flow simulation in humans: endorsed by the publisher.
comparison with pressure wire - fractional flow reserve. EuroIntervention.
(2014) 10:574–83. doi: 10.4244/EIJY14M07_01 Copyright © 2021 Li, Leng, He, Xia, Jiang, Pan, Dong, Sun, Hu, Wang, Xiang and
14. Tu S, Bourantas CV, Norgaard BL, Kassab GS, Koo BK, Reiber JH. Image- Jiang. This is an open-access article distributed under the terms of the Creative
based assessment of fractional flow reserve. EuroIntervention. (2015) 11:V50– Commons Attribution License (CC BY). The use, distribution or reproduction in
4. doi: 10.4244/EIJV11SVA11 other forums is permitted, provided the original author(s) and the copyright owner(s)
15. Morris PD, Ryan D, Morton AC, Lycett R, Lawford PV, Hose DR, et al. Virtual are credited and that the original publication in this journal is cited, in accordance
fractional flow reserve from coronary angiography: modeling the significance with accepted academic practice. No use, distribution or reproduction is permitted
of coronary lesions: results from the VIRTU-1 (VIRTUal Fractional which does not comply with these terms.

Frontiers in Cardiovascular Medicine | www.frontiersin.org 9 October 2021 | Volume 8 | Article 714077