Expert consensus statement on the use of fractional flow reserve, intravascular ultrasound, and optical coherence tomography
The rationale for use of intracoronary physiology assessment and imaging arises from the limitations of coronary angiography, the traditional method for determining the severity of coronary stenoses. The visual assessment of percent diameter reduction has significant interobserver variability 1-3, even among experienced angiographers 4. Computer-assisted quantitative coronary angiography only marginally improves diagnostic accuracy and its estimate of functional significance 5. Fractional flow reserve (FFR) is used to determine the functional significance of a coronary stenosis 6. Intravascular ultrasound (IVUS) offers excellent visualization of intraluminal and transmural coronary anatomy. Optical coherence tomography (OCT) further improves vascular visualization. There is now persuasive evidence regarding intracoronary diagnostic lesion assessments using physiology and anatomy. These adjunctive diagnostic procedures may influence the decision for coronary revascularization, guide the performance of percutaneous coronary interventions (PCI), and optimize procedural outcomes. There are substantial long-term outcome data showing benefit associated with FFR-guided decision-making. However, these techniques are underutilized in contemporary practice: the rates of use of IVUS and FFR during PCI for intermediate coronary stenoses (40–70% diameter stenosis) are 20.3% and 6.1% respectively 7. In 2011, the ACCF/AHA/SCAI PCI guidelines 8 assigned levels of evidence for the use of these modalities in various clinical situations (Table 1). The purpose of this consensus statement is to review recent studies, to develop a consensus of how these procedures are best utilized in practice, and to support their incorporation into guideline and appropriate use documents. A trans-lesional functional assessment is an important adjunct to coronary angiography for providing an objective evaluation of stenosis severity. FFR is the ratio of mean distal coronary pressure (Pd) to mean aortic pressure (Pa) during maximum hyperemia, usually induced by adenosine i.c. bolus or i.v. infusion, and represents the percentage of normal flow across a coronary stenosis. If the patient has active obstructive airways disease, i.c. adenosine can be used safely instead of IV adenosine. Alternative pharmacologic agents include nitroprusside, dobutamine, and regadenoson. Physiologic stenosis assessment by FFR is a lesion-specific index of epicardial conductance, which is independent of the microvasculature and hemodynamic changes induced by variations in heart rate, blood pressure or myocardial contractility. The FFR threshold for detecting ischemia has been corroborated by multiple tests for myocardial ischemia and reflects the functional significance (i.e. ischemic potential) of an epicardial stenosis. To establish an ischemic threshold, FFR was validated in patients with single vessel intermediate lesions and compared with the combination of three different noninvasive stress tests 9. FFR was first validated using a cutoff value of 0.75. With further experience with the technique, investigators appreciated that by extending the cutoff value to 0.80, the sensitivity of FFR could be improved without greatly compromising the specificity. For this reason, a cutoff value of ≤0.80 was used in FAME 1 and FAME 2 and shown to be clinically valid. This is now the recommended ischemic reference standard for the invasive assessment of myocardial ischemia 10 (Fig. 1). Three prospective randomized trials have demonstrated the clinical utility of FFR. To determine the safety of deferring PCI based on nonsignificant FFR, the Percutaneous Coronary Intervention of Functionally Non-significant Stenosis (DEFER) Study 12 randomized 181 patients with stable ischemic heart disease (SIHD) with FFR ≥0.75 across an intermediate stenosis to PCI or to deferral of PCI with medical treatment. At 5-year follow-up, the deferred group had a rate of death or myocardial infarction (MI) that was less than half the rate in the PCI group. To evaluate the utility of FFR for guiding the performance of PCI, the Fractional Flow Reserve versus Angiography for Multivessel Evaluation (FAME) trial 12 randomized 1005 patients with multivessel disease (including SIHD, unstable angina, and NSTEMI) to either FFR-guided PCI or to angiography-guided PCI. The primary outcome, the composite rate of death, MI, or repeat revascularization at 1 year, was significantly lower (13.2% vs. 18.3%, P = 0.02) in patients who received FFR-guided PCI (Fig. 2). This was due to non-significant reductions in each component of the primary endpoint and a significant reduction in the combined rate of death or MI (7.3% vs. 11.1%, P = 0.04) in the FFR-guided group. At 2-year follow-up, the combined rate of death and MI remained significantly lower. An economic evaluation verified that FFR-guided PCI is a cost-saving strategy 13, with significantly fewer stents deployed and significantly less contrast media used. Additionally, patients treated with the FFR-guided strategy had similar rates of freedom from angina compared with the angiography-guided strategy. To compare outcomes in ischemia-guided PCI with medical therapy, the Fractional Flow Reserve versus Angiography for Multivessel Evaluation 2 (FAME 2) trial 14 randomized 888 patients with single or multivessel SIHD to FFR-guided PCI with optimal medical therapy or optimal medical therapy alone. The key difference between FAME 2 and other studies evaluating PCI for SIHD, such as COURAGE 15, is that to be included in the randomized portion of FAME 2, patients had to have at least one lesion with FFR ≤0.80. Enrollment in FAME 2 was stopped early because there was a highly significant difference in the primary endpoint of death, MI and urgent revascularization favoring the FFR-guided PCI arm. This was due to a significantly greater rate of urgent revascularization in the medical therapy arm (11.1% vs. 1.6%, P < 0.001); there was no difference in death or MI. A landmark analysis suggested a higher rate of spontaneous MI in the medical therapy arm starting 1 week after randomization. Patients with angiographic disease that was not hemodynamically significant (based on FFR), and therefore not enrolled in the randomized trial but followed in a registry, had a very low event rate, which supports the role of medical therapy in this group. Thus, clinical outcome studies indicate that measuring FFR optimizes the benefit of PCI and distinguishes stenoses responsible for ischemia from functionally insignificant ones. FFR improves clinical outcomes and saves resources compared with angiography-guided PCI. For optimal outcomes, FFR should be employed when decisions regarding the need for PCI are ambiguous based on the coronary angiogram and available noninvasive data. FFR is especially useful when noninvasive testing is absent, equivocal, or does not provide objective evidence of ischemia in the myocardial segment subtended by the targeted lesion. FFR can justify a procedure if the clinical context suggests a potential benefit. Judgment should always be used when assessing the risk:benefit ratio in clinical decision-making, however, and thus an abnormal FFR does not compel a revascularization procedure. In the presence of intermediate stenoses, or when there is an apparent discordance between lesion severity, location of ischemia by noninvasive testing and clinical symptoms, FFR provides valuable data for clinical decision-making. The limitations of angiography in characterizing intermediate severity stenoses are another important lesson from FAME 12. In lesions with 50 to 70% diameter narrowing, only 35% were hemodynamically significant based on FFR. In lesions with 71 to 90% diameter stenosis, for which many operators would perform PCI, 20% were not hemodynamically significant based on FFR and did not require PCI. Therefore, FFR can be useful in guiding revascularization decisions even in more severe angiographic stenoses when noninvasive data is not available or discordant with coronary angiography. These findings have implications for determining the optimal treatment strategy in patients with multivessel CAD. By measuring FFR and discounting non-ischemic lesions, the Functional SYNTAX Score can be calculated and angiographic three-vessel CAD can be reclassified as one- or two-vessel CAD, which could benefit from PCI and not require CABG 16. Table 2 17-23 summarizes the studies that assess revascularization in left main coronary stenosis (LMCA) based on FFR. FFR <0.75 is a clinically effective tool when used as a cutoff for guiding revascularization decisions and is the preferred technique for evaluating intermediate LMCA lesions. FFR can be used to evaluate the significance of serial stenoses to guide the strategy for determining which lesion(s) should be revascularized and which should be managed medically 24. Observational studies have reported favorable outcomes using FFR in specific anatomic subsets, including diffuse disease 25, bifurcations 26, stent-jailed side branches 27, and nonculprit lesions in ST-segment elevation myocardial infarction 28, but none of these subsets have been the subject of randomized trials. FFR is valid in all nonculprit vessels in non-ST-elevation ACS and valid in most nonculprit vessels in STEMI, with the caveat that STEMI with markedly elevated LVEDP and impaired global microcirculatory function may result in transiently and falsely elevated FFR. The reason is that microvascular impairment reduces the flow across the stenosis, elevating FFR. Therefore, in this setting, a low FFR indicates hemodynamic significance of the non-culprit lesion but a normal FFR is not definitive. The application of FFR in these clinical situations remains to be fully elucidated. The potential role of "functional angioplasty" (i.e., performing PCI on lesions responsible for ischemia and treating medically those that are not), as opposed to complete anatomic revascularization (performing PCI on all lesions that appear angiographically significant) as tested in the FAME trials, would constitute a substantial change from traditional practice. It is quite likely that in the future, functional PCI using FFR or an equivalent physiologic guide will be the measure of PCI appropriateness 10. Instead of relying solely on angiographic criteria of severity when there is no stress test present, or the stress test/anatomy results are discordant, FFR would be the final arbiter, irrespective of lesion severity 29. IVUS is a catheter-based imaging modality that provides high-resolution cross-sectional images of the coronary artery, enabling measurements of luminal and vessel areas 30. Plaque morphology and constituents are identified by the amplitude and frequency of reflected ultrasound signals that correspond to normal tissue, fibrosis, calcium, and necrotic core. The axial resolution is 100 to 200 µm and lateral resolution is 250 µm with frequencies of 20 to 40 MHz 31. IVUS quantification of a stenosis has fewer anatomic limitations than angiography. IVUS algorithms accurately measure minimum luminal area (MLA), a more accurate dimension than luminal diameter. The main limitation is ensuring imaging in a coaxial position. IVUS is an excellent method for determining plaque volume; however, despite numerous investigative applications, its precise role in clinical decision-making has not been defined. IVUS can be especially useful in situations in which angiographic imaging is considered unreliable, such as the presence of ostial lesions or segments with multiple overlapping vessels. Bifurcation lesions are particularly difficult to assess by angiography because overlapping side branches often obscure the lesion. IVUS may provide an optimal assessment in these subsets. Perhaps the most important use of IVUS is as an adjunct before and during PCI. It is especially useful in planning PCI strategy in high-risk subsets, such as left main stenosis, calcified lesions, and bifurcations 32. It is also helpful during stent placement to assess stent sizing, expansion, and apposition. Although IVUS has not been definitively shown to impact procedural mortality or MI when routinely used during PCI, there are data suggesting that IVUS-guided stent placement reduces stent thrombosis, restenosis, and repeat revascularization 33. A meta-analysis by Zhang et al. 34 (Table 3) showed improved clinical outcomes using IVUS to direct PCI. While there was no significant reduction in MI rate, stent thrombosis, and mortality were significantly reduced in the IVUS guided group 32, 35-43. Recently, the ADAPT-DES study evaluated 1-year clinical outcomes in patients undergoing PCI with DES 44. IVUS-guided PCI changed the interventional strategy in 74% of cases; at 1 year, there was a significant reduction in definite/probable stent thrombosis (0.52% vs. 1.04%, P = 0.01) and MI (2.5% vs. 3.7%, P = 0.002) but no reduction in mortality (1.0% vs. 1.4%, P = 0.14). For LMCA stenosis, IVUS is an important procedural adjunct, both pre- and postintervention, if stenting is being considered 45. OCT uses the scattering and absorption of near-infrared light. The OCT light source operates on a wavelength range of 1,250 to 1,350 nm, providing tissue penetration of 1 to 3 mm, and a spatial resolution at the cellular level 46. It has very high axial and lateral resolution, which provides accurate characterization of plaque morphology and composition in real-time, including thin fibrous caps, lipid pools, and fibrocalcific plaques 47. OCT has better resolution than IVUS, but less power of penetration. OCT is better at looking at fine detail in the near field, around the lumen and stent edges, but is less valuable for imaging plaque size or determining tissue characteristics. The improved spatial resolution compared to IVUS has raised the possibility that OCT could ultimately replace IVUS. However, its limited depth of penetration mitigates its ability to visualize the external elastic lamina, especially in large or proximal vessels. There are no prospective randomized trials of OCT for guiding PCI, but there are several single center trials that suggest the potential for improving poststent clinical decision-making 48-51. OCT can demonstrate thrombus, unrecognized plaque rupture, stent underexpansion, significant edge dissections, and excessive plaque at the stent edges treatable with further stent expansion or the placement of additional stents. IVUS can also detect these conditions, but OCT provides improved resolution. In the multicenter CLI-OPCI trial 52, OCT identified adverse features requiring further intervention in 35% of cases. The OCT arm had a significantly lower risk of death and MI at 1 year. OCT has potential benefit as an adjunct means of evaluating the anatomy and composition of stenoses of uncertain severity or morphology, but this application needs further evaluation. Of particular interest is its capability to determine fibrous cap thickness, potentially identifying vulnerable plaque, and perhaps predict impending rupture 53-55. The identification of thrombus is more accurate with OCT than with IVUS. A recent study 52 identified patients with ST-elevation-MI who could be treated with thrombus aspiration alone based on an OCT finding of plaque erosion rather than fibrous cap rupture, suggesting unique medical and cost saving possibilities. To assess the correlation between OCT luminal dimensions and FFR, Shiono et al. 56 evaluated 62 intermediate coronary lesions in 59 patients. An OCT-derived MLA = 1.91 mm2 (sensitivity = 93.5%, specificity = 77.4%), MLD = 1.35 mm (sensitivity = 90.3%, specificity = 80.6%), and percent lumen area stenosis >70% (sensitivity = 96.8%, specificity = 83.9%) had the best cutoff values for a FFR <0.75. Gonzalo et al. 57 compared OCT and IVUS with FFR to determine the accuracy of OCT in identifying hemodynamically severe coronary stenoses in 61 stenoses studied in 56 patients. Although OCT and IVUS demonstrated a similar diagnostic accuracy in detecting lesions with FFR <0.80, OCT was superior in the subgroup of vessels <3 mm diameter. To date, no prospective randomized studies have been performed to demonstrate improvement in clinical outcomes from this technology. There is no proven impact on mortality, MI, stent thrombosis, or restenosis rates. There are also no prospective randomized trials assessing the role of OCT-guided PCI. Since OCT, like IVUS, evaluates anatomic dimensions rather than functional significance, its use in assessing lesion physiology will probably be limited. As a result, this imaging modality remains investigational in terms of improving clinical outcomes associated with the performance of PCI. Accurate quantification of the severity of LMCA stenoses can be a valuable adjunct to diagnosis when coronary angiography gives equivocal or ambiguous images. When using IVUS to determine LMCA severity, the most widely used parameter is MLA. IVUS assessment has a relatively strong correlation with FFR in evaluating intermediate LMCA stenoses. A normal FFR can be predicted reasonably well with IVUS dimensions 58, 59. Limited variability in LMCA length, diameter, and amount of supplied myocardium explains the better correlation in LMCA than non-LMCA stenoses. However, both techniques have theoretical and practical limitations. Proximal LAD and/or LCX disease can impact FFR of LMCA stenoses. With IVUS, distal LMCA lesions can be difficult to accurately image, and often requires pullback from both the LCX and LAD. Table 4 58, 60 summarizes the studies that correlated IVUS MLA in LMCA stenoses with FFR. Jasti et al. 58 showed good correlation between FFR and IVUS, with sensitivities and specificities >0.90. In a study of 55 intermediate LMCA lesions, an MLA <5.9 mm2 and an MLD <2.8 mm correlated well with FFR<0.75 61. In 354 intermediate left main stenoses, an MLA >6.0 mm2 identified patients at low risk for adverse events with deferred revascularization 32. A prospective application of these criteria was tested in the LITRO study 62. LMCA revascularization was performed in 90.5% (152 of 168) of patients with an MLA <6 mm2 and was deferred in 96% (179 of 186) of patients with an MLA >6 mm2. In a 2-year follow-up period, cardiac death-free survival was 97.7% in the deferred group versus 94.5% in the revascularized group (P = ns), and event-free survival was 87.3% versus 80.6%, respectively (P = ns). At 2-year follow-up, only eight (4.4%) patients in the deferred group required subsequent LMCA revascularization, none of who had an MI. Thus, it is safe to defer LMCA revascularization with MLA >6 mm2. Additionally, the data confirms that MLA<6.0 mm2 is clinically significant, correlates with FFR <0.75, and may warrant intervention to improve 1-year mortality 60. In Asian populations, with smaller normal coronary diameters, an MLA cutoff <4.8 mm2 correlates better with reduced FFR <0.8 and <4.1 mm2 with FFR <0.75 36, 60. The primary limitation of IVUS MLA in predicting hemodynamic significance in non-LMCA lesions is that the functional effects of a lesion are dependent on additional factors besides dimension. These include lesion location in the coronary tree, lesion length, eccentricity, entrance and exit angles, shear forces, reference vessel dimensions, and the amount of viable myocardium subtended by the lesion 25. Therefore, in non-LMCA lesions there is only moderate correlation between anatomic dimensions by IVUS and ischemia by physiological assessment. The IVUS and FFR correlation is best in demonstrating nonsignificant lesions 59; their correlation in demonstrating significant stenoses is weaker. Part of the reason for this deficiency is that attempting to determine a critical MLA without considering the reference vessel MLA leads to inaccuracy. An MLA = 3.0 mm2 in a proximal versus distal arterial segment has entirely different effects on flow and subsequent clinical implications. In non-LMCA stenosis, IVUS MLA <4.0 mm2 correlates with ischemia on single-photon emission computed tomography and also correlates moderately well with an FFR <0.75 (sensitivity and specificity 92% and 56%, respectively). Importantly, low event rates are observed in intermediate lesions when intervention is deferred with an IVUS MLA ≥4 mm2 63-65. In the largest study to date, IVUS was compared with FFR in 544 lesions 66. The optimal cut-off value for predicting an FFR ≤0.80 was an MLA = 2.9mm2 by IVUS, but the overall accuracy was only 66%. Moreover, of the 240 lesions that had an MLA <2.9 mm2, only 47% were hemodynamically significant by FFR. Similarly concerning, 19% of lesions with an MLA >2.9 mm2 had an FFR <0.80, limiting the utility of IVUS for lesion assessment. Kang et al. 60 evaluated 236 angiographically intermediate coronary lesions in which both IVUS and FFR measurements were performed. An IVUS MLA ≤2.4 mm2 had the maximum accuracy for predicting FFR <0.80. However, the overall diagnostic accuracy was 68% with a confidence interval ranging from 1.8 to 2.6 mm2. FIRST was a multicenter prospective registry of patients who underwent elective coronary angiography and had intermediate coronary stenoses (40–80%) 67. An IVUS-measured MLA <3.07 mm2 had the best sensitivity and specificity (64% and 64.9%, respectively) for correlating with FFR <0.80. Thus, FFR is better validated than IVUS as a physiologic assessment and should be considered the standard for assessing the hemodynamic significance of intermediate non-LMCA lesions. An MLA ≥4.0 mm2 has reasonable accuracy in identifying non-significant lesions for which PCI can be safely deferred. However, an MLA <4.0 mm2 does not accurately predict a hemodynamically significant lesion and should not be used in the absence of supporting functional data to recommend revascularization 25. An MLA <3.0 mm2 is most likely a significant stenosis, but due to its only modest sensitivity and specificity, physiologic testing is desirable before proceeding with revascularization. The writing group recommends that the following conclusions be adopted in clinical practice and in future guidelines and appropriateness documents. In SIHD, when noninvasive stress imaging is contraindicated, discordant, nondiagnostic, or unavailable, FFR should be used to assess the functional significance of intermediate coronary stenoses (50–70%) and more severe stenoses (<90%). In patients with multivessel coronary disease, PCI guided by FFR measurement improves outcomes and saves resources when compared to PCI guided by angiography alone. In patients with three-vessel coronary disease, measuring FFR could allow reclassification of number of vessels diseased and/or SYNTAX score, thereby guiding decisions regarding revascularization by CABG or PCI. In SIHD, PCI of lesions with FFR <0.80 improves symptom control and decreases the need for hospitalization requiring urgent revascularization when compared with medical therapy alone. In SIHD, medical therapy is indicated for an angiographically intermediate stenosis (LMCA or non-LMCA) of unclear clinical significance when FFR >0.80. FFR measurement of the culprit vessel in a patient with an acute ST segment elevation myocardial infarction or any unstable acute coronary syndrome presentation should not be performed. IVUS is an accurate method for determining optimal stent deployment (complete stent expansion and apposition and lack of edge dissection or other complications after implantation), and the size of the vessel undergoing stent implantation. IVUS can be used to appraise the significance of LMCA stenosis and, employing a cutoff MLA = 6 mm2, assess whether revascularization is warranted. IVUS can be useful for the assessment of plaque morphology. IVUS measurements for determination of non-LMCA lesion severity should not be relied upon, in the absence of additional functional evidence, for recommending revascularization. Determination of optimal stent deployment (sizing, apposition, and lack of edge dissection), with improved resolution compared with IVUS. OCT can be useful for the assessment of plaque morphology. OCT should not be performed to determine stenosis functional The writing group with guidelines that these modalities are not indicated when imaging and angiographic data are or when the result of the additional procedure will not the treatment strategy or of stent implantation.