Evidence robustly demonstrates that ischemia, rather than anatomy, is the optimal
target for coronary revascularization. In the cardiac catheter laboratory, fractional
flow reserve (FFR) and corresponding diastolic indices are regarded as the gold standard
for physiological lesion assessment and ischemia detection (Table 1). Yet, despite
a wealth of supporting data and indications in international guidelines, the use of
FFR remains surprisingly low in the diagnostic assessment of coronary artery disease
across the world.1, 2 To address this, multiple groups have developed methods for
computing FFR from invasive angiography, without the need for passing a pressure wire
or inducing hyperemia, thus removing the main barriers to uptake. Angiography‐derived
FFR therefore has the potential to extend the benefits of physiological coronary lesion
assessment to considerably more patients. Given the size of the interventional cardiology
market, clinical and commercial motivation to deliver these tools as quickly as possible
could hardly be greater. Several models are now approved as medical devices. Imminently,
physicians and healthcare providers will have to decide whether to use these tools.
But do they truly deliver physiology, and are they accurate enough? There are 3 particular
areas of that deserve close scrutiny.
Table 1
Angiography‐Based Coronary Physiological Assessment Techniques
Index
Abbreviation
Calculated
Equipment
Potential Benefits
Pitfalls/Limitations
Fractional flow reserve
FFR
Whole cardiac cycle Pd/Pa at hyperemia
Pressure wire
Predicts percentage improvement in flow with PCI. Good clinical outcomes data
Does not measure absolute flow and microvascular resistance
Instantaneous wave‐free ratio/resting full‐cycle ratio
iFR/RFR
Pd/Pa during diastolic phase
Pressure wire
Good clinical outcome data, relative to FFR
Does not measure absolute flow and microvascular resistance
Index of myocardial resistance
IMR
(Pd) · (thermodilution derived mean transit time)
Thermo‐ and pressure‐sensitive wire
Microvascular resistance becoming of increasing interest (eg, PCI nonresponders, ANOCA,
AMI, HFpEF)
Thermodilution not widely used
Hyperemic microvascular resistance
HMR
Pd/Doppler flow velocity
Doppler and pressure wire
Microvascular resistance becoming of increasing interest (eg, PCI nonresponders, ANOCA,
AMI, HFpEF)
Doppler flow velocity challenging to measure. Doppler wires not widely used
Hyperemic stenosis resistance
HSR
(Pa‐Pd)/Doppler flow velocity
Doppler and pressure wire
Objective, direct measure of the resistance of proximal disease
Doppler flow velocity challenging to measure. Doppler wires not widely used. Surrogate
index
Angiography‐derived FFR
vFFR/FFRangio/QFR
Fluid dynamics equations informed by anatomy
Computational fluid dynamics software
Delivering clinical benefits of FFR without factors that limit the invasive technique
Relatively wide Bland–Altman limits of agreement compared with FFR. Requires excellent
angiography. Less accurate in those with nonaverage microvascular resistance
CT‐derived FFR
CTFFR
Fluid dynamics equations informed by anatomy
Computational fluid dynamics software (offline)
Reduce the number of unnecessary invasive catheterizations
Relatively wide Bland–Altman limits of agreement compared with FFR
Coronary flow reserve
CFR
(Hyperemic flow surrogate)/(baseline flow surrogate)
Flow derived from Doppler velocity or thermodilution mean transit time
Doppler or thermosensitive wire
A surrogate for flow and vasodilatory reserve. Flow more important than pressure,
but hard to measure
Prone to same limitations as those for Doppler wire or thermodilution. Variability
in baseline measurement can impair interpretation
Absolute coronary flow
Qb
Infusion flow · (infusion temp/sensor temp) · 1.08
During continuous saline infusion
Thermosensitive wire, pressure wire, monorail infusion catheter
Predicts absolute (not percentage) coronary flow changes and microvascular resistance
Additional time, expertise, and hardware
All physiological indices are surrogate markers of physiology derived from other measures.
AMI indicates acute myocardial infarction; ANOCA, angina and no obstructive coronary
artery disease; FFR, fractional flow reserve; HFpEF, heart failure with preserved
ejection fraction; MVR, microvascular resistance; Pa, proximal pressure; PCI, percutaneous
coronary intervention; Pd, distal coronary pressure; and QFR, quantitative flow ratio.
John Wiley & Sons, Ltd
Simplification
Methods for computing angiography‐derived FFR are software based. Three‐dimensional
arterial anatomy is reconstructed from paired 2‐dimensional angiogram images. Mathematical
equations that define hemodynamic laws are then applied to the reconstructed artery
to predict the pressure dynamics along the artery, which are displayed as a color‐mapped
3‐dimensional artery. In an effort to rationalize these models to make them practical
and expedient for clinical use, many groups have abandoned complex, numerical, computational
fluid dynamics simulation in favor of analytical solutions based broadly upon the
laws of Bernoulli and/or Poiseuille. These simpler physical laws characterize pressure
losses attributable to convective acceleration and viscous friction, respectively.
They are quick and simple to execute and perform well under steady (nonpulsatile),
laminar flow conditions, in straight conduits. Coronary arteries, however, are not
straight, and flow is pulsatile. Furthermore, these laws are unable to accurately
characterize complex translesional pressure dynamics, particularly poststenosis pressure
recovery, which is the basis of FFR. Some stenosis models make empiric assumptions
or corrections for pressure loss and recovery. On average, these may perform adequately,
but cannot represent the potentially complex flow patterns in a specific case. Moreover,
they may be particularly vulnerable to inaccuracy in the context of serial lesions
and diffuse disease in which 3‐dimensional computational fluid dynamics computations
more reliably characterize interstenosis hemodynamic interaction. The impact this
has on accuracy, in all disease patterns, is yet to be fully determined.
Assumptions
The discordance between angiographic severity and physiological (FFR) significance
is well described and affects ≥30% of lesions. Discrepancies occur because, unlike
angiography, FFR elegantly and automatically incorporates the combined and inter‐related
effects of coronary flow and microvascular resistance. It is therefore imperative
that computational models of angiography‐derived FFR include adequate physiological
inputs or “tuning” to represent the maximum blood flow or minimum microvascular resistance;
the latter dictates the former, which, in turn, dictates the pressure gradient and
FFR. Hemodynamic equations are capable of accurately deriving a variety of physiological
parameters, but only if other appropriate physiological inputs, such as flow or microvascular
resistance, are included. A sensitivity analysis demonstrated that microvascular resistance
was the dominant influence on angiography‐derived FFR, above and beyond the severity
or anatomy of epicardial disease.3 Hyperemic flow and minimal microvascular resistance
are variable in health and disease and are hard to measure, even with invasive instrumentation.
Noninvasive models of angiography‐derived FFR therefore rely upon assumptions about
these parameters, or predict them from surrogate markers such as arterial diameter.
Again, empiric assumptions may be sufficient overall, for many cases, but will be
inaccurate in nonaverage cases with discordant anatomy and physiology, that is, the
very cases where FFR is superior to angiography. Therefore, unless models have an
accurate method for achieving this, on a patient‐specific basis, the “physiological”
prediction becomes simply a function of stenosis geometry and they cannot be a genuine
model of FFR at all (Figure). As an example, 1 study of angiographically derived FFR
observed a significant reduction in diagnostic accuracy in patients with elevated
microvascular resistance.4 Paradoxically, physiologically weak models will appear
more feasible relative to angiographic appearance, and a potential danger is that
user confidence may therefore be increased with poorer methods. FFR has enabled a
great stride forward in terms of physiologically guided revascularization. It would
be unfortunate if, in an attempt to increase physiological assessment, we were to
take half a step back toward assessment based on epicardial arterial anatomy. Table 2
summarizes major trials of angiography‐derived FFR.4, 5, 6, 7, 8, 9, 10, 11, 12, 13,
14, 15, 16, 17, 18
Figure 1
Error in angiography‐derived FFR.
(A) An anatomically severe circumflex case. In this case, the method applied an assumed
value for microvascular resistance based on a population average, which resulted in
considerable disagreement between angiography‐derived and invasive FFR (0.55 vs 0.82).
(B) Bland–Altman plot from a meta‐analysis of 13 studies (1842 vessels). There is
minimal bias (gray line), but the ±95% limits of agreement were FFR ±0.14. FFR indicates
fractional flow reserve. Reprinted from Collet et al20 with permission. Copyright
©2018, Oxford University Press.
Table 2
Major Trials/Studies of Angiographically Derived FFR
Author
Study
Year
N=Arteries
Surrogate/Software/Company
Mathematical Solution
Diagnostic Accuracy
95% Limits of Agreement
Morris et al5
VIRTU‐1
2013
35
vFFR from VIRTUheart (University of Sheffield)
Transient 3D CFD
97%
FFR ±0.16
Tu et al6
FAVOUR Pilot
2016
84
QFR from QAngio XA (Medis Medical Imaging Systems, NL)
Empiric flow velocity (fQFR), TIMI frame counting‐derived contrast velocity at baseline
(cQFR) and under hyperemia (aQFR). Analytical equations based on laws of Bernoulli
and Poiseuille
fQFR 80%
cQFR 86%
aQFR 87%
FFR ±0.14
FFR ±0.12
FFR ±0.13
Kornowski et al7
FFRangio FIM
2016
101
FFRangio (CathWorks, Israel)
Simple analytical equation, based on law of Poiseuille
94%
FFR ±0.10
Trobs et al8
FFRangio
2016
100
FFRangio from Syngo IZ3D and prototype software (Siemens Healthcare GmbH, Germany)
CFD based on BP, anatomy, and literature estimates of microvascular resistance
90%
FFR ±0.13
Pellicano et al9
FFRangio validation
2017
203
FFRangio (CathWorks, Israel)
Simple analytical equation, based on law of Poiseuille
93%
FFR ±0.10
Xu et al10
FAVOUR II China
2017
328
QFR from QAngio XA (Medis Medical Imaging Systems, NL)
TIMI frame counting‐derived contrast velocity at baseline (cQFR). Analytical equations
based on laws of Bernoulli and Poiseuille
93%
FFR ±0.13
Yazaki et al11
QFR in intermediate lesions
2017
151
QFR from QAngio XA (Medis Medical Imaging Systems, NL)
TIMI frame counting‐derived contrast velocity at baseline (cQFR). Analytical equations
based on laws of Bernoulli and Poiseuille
88%
FFR ±0.10
Westra et al12
WIFI II
2018
240
QFR from QAngio XA (Medis Medical Imaging Systems, NL)
TIMI frame counting‐derived contrast velocity at baseline (cQFR). Analytical equations
based on laws of Bernoulli and Poiseuille
83%
FFR ±0.16
Mejía‐Rentería et al4
QFR IMR study
2018
300
QFR from QAngio XA (Medis Medical Imaging Systems, NL)
TIMI frame counting‐derived contrast velocity at baseline (cQFR). Analytical equations
based on laws of Bernoulli and Poiseuille
IMR <23 =88%
IMR ≥23 =76%
FFR ±0.12
FFR ±0.15
Westra et al13
FAVOUR II EJ
2018
317
QFR from QAngio XA (Medis Medical Imaging Systems, NL)
TIMI frame counting‐derived contrast velocity at baseline (cQFR). Analytical equations
based on laws of Bernoulli and Poiseuille
87%
FFR ±0.12
Fearon et al14
FAST‐FFR
2019
319
FFRangio (CathWorks, Israel)
Simple analytical equation, based on law of Poiseuille
92%
FFR ±0.13
Omori et al15
FFRangio in multivessel disease
2019
118
FFRangio (CathWorks, Israel)
Simple analytical equation, based on law of Poiseuille
92%
FFR ±0.14
Stahli et al16
All comer QFR
2019
516
QFR from QAngio XA (Medis Medical Imaging Systems, NL)
TIMI frame counting‐derived contrast velocity at baseline (cQFR). Analytical equations
based on laws of Bernoulli and Poiseuille
93%
FFR ±0.07
Masdjedi et al17
FAST‐study
2019
100
vFFR from 3D QCA software, CAAS workstation (PIE Medical Imaging, NL)
Simple analytical equation, based on laws of Bernoulli and Poiseuille
AUC=0.93
FFR ±0.07
Li et al18
FLASH‐FFR
2019
328
caFFR from FlashAngio (Rainmed Ltd, China)
CFD based on postangiography TIMI frame counting of flow velocity
96%
FFR ±0.10
Listed in chronological order. Invasive FFR (threshold ≤0.80) was comparator in each
study. 3D indicates 3‐dimensional; aQFR, adenosine QFR; AUC, area under the curve;
BP, blood pressure; caFFR, coronary angiography–derived fractional flow reserve; CFD,
computational fluid dynamics; cQFR, contrast QFR; EJ, Europe and Japan; FFR, fractional
flow reserve; FFRangio, FFR derived from coronary angiography; FIM, first in man;
fQFR, fixed QFR; IMR, index of microcirculatory resistance; QFR, quantitative flow
ratio; TIMI, thrombolysis in myocardial infarction; and vFFR, virtual fractional flow
reserve.
John Wiley & Sons, Ltd
Accuracy and Error Range
Headline validation results report “diagnostic” accuracy. This quantifies how well
a method predicts physiological significance or nonsignificance (FFR ≤0.80), relative
to invasive FFR, expressed as sensitivity, specificity, positive, and negative predictive
values, area under a receiver operating curve, and overall diagnostic accuracy. Diagnostic
accuracy is a function of (1) the method's accuracy and (2) the cases included in
a particular study. The fewer cases close to the 0.80 threshold, the better the diagnostic
accuracy will appear and vice versa. This is nicely illustrated in a study of FFR
computed from computed tomography coronary angiography in which the diagnostic accuracy
was 82% overall, but only 46% in cases in FFR were 0.70 to 0.80, which is precisely
the range where most accuracy is required.19
The best test of how accurately angiography‐derived FFR agrees with invasive FFR is
to plot the differences between predicted and observed FFR values against the mean
(ie, a Bland–Altman plot). From this, the mean difference (delta), which quantifies
any bias in the angiography‐derived method, and the 95% limits of agreement, are calculated.
The limits of agreement (±1.96 SDs) comprise 95% of observed differences and are akin
to the 95% CI of a computed, angiography‐derived FFR result or an error range (Figure).
The wider the limits of agreement, the larger the method's error and vice versa. Unlike
diagnostic accuracy, the limits of agreement are only a function of how accurate a
method is. A recent meta‐analysis of 13 studies of angiography‐derived FFR demonstrated
impressive diagnostic accuracy (sensitivity, 89%; specificity, 90%), but more‐sobering
agreement, with limits of agreement of FFR ±0.14.20 This is remarkably similar to
FFR computed from computed tomography in the NXT trial (limits of agreement FFR ±0.15).21
FFR computed from computed tomography, however, is a noninvasive screening tool, best
used to reduce unnecessary invasive catheterization. Arguably, the accuracy “bar”
should be set far higher for a test in the catheter laboratory, where results directly
influence decisions regarding proceeding to percutaneous or surgical intervention.
Is FFR ±0.14 accurate enough for interventional decision making? It is likely that
noninferiority trials will be used to assess these methods. These should avoid the
usual pitfalls and be appropriate in terms of power, significance, analysis protocol,
sample size, patient population, and prespecified noninferiority margins. Moreover,
it remains to be seen how accurate and reproducible these methods are, beyond academic
core laboratories, in the hands of those who will be expected to use these tools (ie,
the interventional cardiologist operating in the catheter laboratory).
Conclusions
Angiography‐derived FFR has the potential to change clinical practice for the considerable
benefit of patients by providing routine physiological data, together with coronary
anatomy, to provide personalized management and improved clinical outcomes. However,
deriving physiology from anatomy is challenging and requires assumptions. Model simplification
and physiological assumptions, based on extrapolated or averaged data, are likely
to work in the majority of patients. However, much of FFR's success lies in its ability
to identify those cases where nonstandard microvascular resistance and/or flow result
in discordant physiology and anatomy. It is therefore important that models of angiography‐derived
FFR retain the same patient‐specific physiology that separates traditional FFR from
angiography, or at least that they highlight which cases require more‐reliable assessment.
Operators must understand how accuracy and error are defined in all patient groups.
Stringent validation is required to prove that models are accurate and physiologically
sound, in the hands of those who will be using them. If this can be achieved, clinicians
have the potential to achieve what could be a new level of patient‐specific medicine.
Sources of Funding
Dr Morris is funded by a Wellcome Trust Clinical Research Career Development Fellowship
(214567/Z/18/Z).
Disclosures
Dr Morris has previously received honoraria (speaking fees) from Abbott UK. Professor
Curzen has received unrestricted research grants from HeartFlow and Boston Scientific
for the FORECAST and RIPCORD2 trials, respectively. The remaining authors have no
disclosures to report.