Coronary CT Angiography-derived Fractional Flow Reserve Testing in Patients with Stable Coronary Artery Disease: Recommendations on Interpretation and Reporting

Noninvasive fractional flow reserve derived from coronary CT angiography (FFRCT) is increasingly used in patients with coronary artery disease as a gatekeeper to the catheterization laboratory. While there is emerging evidence of the clinical benefit of FFRCT in patients with moderate coronary disease as determined with coronary CT angiography, there has been less focus on interpretation, reporting, and integration of FFRCT results into routine clinical practice. Because FFRCT analysis provides a plethora of information regarding pressure and flow across the entire coronary tree, standardized criteria on interpretation and reporting of the FFRCT analysis result are of crucial importance both in context of the clinical adoption and in future research. This report represents expert opinion and recommendation on a standardized FFRCT interpretation and reporting approach. Published under a CC BY 4.0 license.

S ince the first study on coronary CT angiographyderived fractional flow reserve (CT FFR) diagnostic performance by Koo and colleagues in 2011 (1), an abundance of data pertaining to this modality has been published. Several tools have been introduced for the calculation of CT FFR (1-3); however, the majority of existing evidence and clinical experience is based on the HeartFlow FFR CT method (HeartFlow, Redwood City, Calif ), which is the only CT FFR cleared by the United States Food and Drug Administration (4) and endorsed by the National Institute for Health and Care Excellence in the United Kingdom (5). Comprehensive reviews of the principle of FFR CT have been described previously (6)(7)(8). FFR CT assessment is increasingly used in mainstream clinical practice (9)(10)(11)(12)(13)(14) and is likely to further expand with the increased utilization of coronary CT angiography as a first-line test in patients suspected of having coronary artery disease (CAD). While there has been much focus on the diagnostic performance and potential clinical utility of FFR CT in patients with moderate CAD (9)(10)(11)(12)(13)(14)(15)(16)(17)(18), there has been less focus on interpretation, reporting, and integration of FFR CT results into routine clinical practice (19). A broadly adopted standardized FFR CT interpretation and reporting approach providing rich and consistent information may facilitate more appropriate clinical implementation and stimulate further high-quality research. Thus, this report, which was written by an independent group of physicians with years of clinical experience with FFR CT , proposes standardized criteria for FFR CT interpretation and reporting for application in clinical practice and for clinical research.

FFR versus FFR CT
FFR CT provides simultaneous calculation of pressure and flow across the entire coronary tree (Fig 1). In contrast, information pertaining to invasively measured FFR is only available in vessels that have been interrogated with the pressure wire, which is typically decided during invasive coronary angiography at the discretion of the interventionists (20). While anatomic percentage of stenosis is evaluated at the location of the lesion, invasive FFR is typically measured by positioning the pressure sensor in the distal part of the vessel and then manually pulling the pressure sensor back to the ostium to assess the distribution of abnormal epicardial resistance along the course of the vessel (20). In both invasive FFR and FFR CT , the distal values in any given vessel reflect the cumulative pressure loss and impact of all disease proximal to the measurement location. Values obtained by both techniques may vary depending on the measurement location within a vessel. Accordingly, in vessels that have been assessed using both techniques, if the measurement locations of invasive FFR and FFR CT are not matched, their values can be different and may not closely correlate. significant and that the patient can be safely treated with optimal medical treatment without further downstream testing (1,(12)(13)(14)(15)(16)(17)(18)22,23). A poststenotic FFR CT value less than or equal to 0.80 indicates the possibility of hemodynamic significance (1,(15)(16)(17)(18). The use of this dichotomous FFR CT threshold to guide treatment decisions, namely to avoid further downstream testing or consider invasive angiography and revascularization, remains controversial, as it is well known from the invasive literature that the greatest benefit of revascularization is obtained in patients with the most severe pressure loss (24,25). We recommend a dichotomous interpretation strategy to be considered in lesions with FFR CT greater than 0.80 or less than or equal to 0.75 (ie, values >0.80 are "normal" and values ≤0.75 are associated with high likelihood of hemodynamic significance) (Table, Figure 2). Several factors support this strategy. First, FFR CT values are lower than measured FFR (with a bias ranging between 0.03 and 0.05) (16,18). Second, among patients with FFR CT values less than or equal to 0.80, there is a graded correlation between FFR CT and invasive FFR, with the highest FFR CT uncertainty in the range between 0.76 and 0.80 and the highest agreement when FFR CT is less than or equal to 0.75 (12,18). Third, FFR CT , similar to FFR, exhibits a continuous relationship between its numerical value and clinical outcomes, with the worst outcome at lower FFR CT values (14,22,23). Finally, symptomatic patients with moderate CAD determined at CT angiography and FFR CT values greater than 0.80 and in whom invasive angiography is deferred have a favorable prognosis (12,14,22,23).
Clinical decision making in patients with FFR CT ranging between 0.76 and 0.80 is nuanced and may benefit from consideration of additional risk stratification information (Fig 2). Identifying patients at incrementally higher cardiovascular risk, who may benefit from an early coronary angiography approach, can be done by assessing several factors: high-risk plaque features (low attenuation, positive remodeling, napkin-ring sign) (26)(27)(28), plaque burden (27,28), stenosis location (proximal vs distal; main vessel vs side branch) (25,29,30), vessel territory (left anterior descending artery [LAD] vs non-LAD) (29), ratio of coronary vessel volume to myocardial mass (31), and/ or the translesional FFR CT gradient (FFR CT ) (32). It is the opinion of the present author group that in certain instances with FFR CT values less than or equal to 0.75 (eg, small vessels, distal lesions, side branches), patients may be treated with optimal medical therapy without referral to invasive angiography as a first-line strategy (14,33).
In a recent retrospective study, a large pressure drop (FFR CT  0.06) was a stronger predictor of culprit lesions for future acute coronary syndromes than FFR CT measured distal to the lesion alone (32). Ongoing studies are assessing the potential diagnostic value of FFR CT in clinical practice. Overall, the results of FFR CT , as for invasive FFR, must always be evaluated in their clinical context, taking into account patient symptoms and comorbid conditions, which inform the goals of coronary intervention, in combination with the coronary anatomy and suitability of revascularization.

FFR CT Interpretation
As for CT angiography, FFR CT interpretation should be performed by the local imaging experts determined by level of clinical knowledge and practical experience with the technique. This may include cardiologists and/or radiologists. It is recommended that downstream management decision making beyond FFR CT takes into account both the clinical scenario (symptoms, risk profile, and/or comorbid conditions) and the coronary anatomy.

Evaluation of CT Angiography and Lesion Location
The first step in the interpretation of FFR CT is to re-examine the original coronary CT angiography study with particular focus on the location and severity of detailed anatomic lesions (Table). Because FFR CT declines along the length of the vessel with serial focal lesions or areas of diffuse disease, it is important to correlate the pressure loss to specific lesions, which can only be established by direct comparison between the CT angiography lesion location and the FFR CT three-dimensional coronary tree model in relation to identifiable vessel landmarks, such as origin, branches, and segments. It is recommended that this first step be performed by using the Society of Cardiovascular Computed Tomography (SCCT) coronary segmentation model (21).

FFR CT Threshold
There is high per-patient and per-vessel agreement between FFR CT and invasive FFR using the threshold of 0.80 for both techniques (1,(15)(16)(17)(18). An FFR CT value greater than 0.80 indicates that the lesion is unlikely to be hemodynamically Abbreviations CAD = coronary artery disease, FFR = fractional flow reserve, FFR CT = CT angiography-derived FFR, LAD = left anterior descending artery, SCCT = Society of Cardiovascular Computed Tomography Summary Expert opinion and recommendation was given by an independent group of physicians on a standardized interpretation and reporting approach for CT-derived fractional flow reserve testing supported by years of clinical experience.

Key Points
n Standardized criteria on interpretation and reporting of CTderived fractional flow reserve (FFR CT ) analysis results are of importance both in context of their clinical adoption and in future research.
n Use of the FFR CT value 10-20 mm distal to the lower border of the stenosis for decision making is recommended.
n We recommend for clinical decision making a dichotomous interpretation strategy to be considered only in lesions with FFR CT greater than 0.80 or lower than or equal to 0.75, whereas, in patients with FFR CT ranging between 0.76 and 0.80, additional risk stratification information is needed.
n The results of FFR CT must be evaluated in their clinical context, taking into account patient symptoms, the coronary anatomy, and suitability of revascularization. dilatation, resulting in reduced flow velocity and pressure recovery (Fig 3). In a recent study, it was suggested that a reliable location at which to assess FFR CT was 1 cm distal to the end of a stenosis (36). For clinical decision making, we recommend using the FFR CT value 1-2 cm distal to the lower border of the stenosis, avoiding the pressure recovery phenomenon.

Nørgaard et al
Distal Vessel FFR CT Values FFR CT provides simultaneous computation of pressure and flow in the entire coronary tree, thus exposing both lesion-specific pressure as well as nadir FFR CT values across the coronary system, which in various settings may drop less than or equal to 0.80 (14,19,(34)(35)(36) (Fig 1). Low terminal vessel FFR CT values (rather than a value distal to stenosis) may include effects unrelated to the stenosis (19,(35)(36)(37). These low values remote from a focal lesion may be due to diffuse CAD or reflect the sum of serial flow-limiting lesions (35)(36)(37). In recent studies, 35%-44% of patients with stable CAD and terminal vessel FFR CT values less than or equal to 0.80 were reclassified as negative when the FFR CT point of reading was 1-2 cm distal to stenosis (14,35). In one observational single-center study, the intermediate follow-up clinical outcome was favorable in patients with terminal FFR CT values less than or equal to 0.80 who were treated with optimal medical treatment (14). In vessels without a significant pressure loss within 2 cm distal to the lesion of interest, but with FFR CT values less than or equal to 0.80 in nearby (eg, mid coronary) segments, we recommend assessment for extent of upstream disease including both CT angiography and FFR CT . FFR CT values less than or equal to 0.80 in such circumstances may be clinically relevant (especially when present distal to a lesion in a proximal segment supplying a large myocardial territory). The group recognizes that more research is needed, particularly in large vessels that have discordance between lesion-specific FFR CT and values taken 2 cm beyond an upstream lesion.

Serial Lesions
The individual contribution of a given lesion in the event of serial stenosis cannot be assessed with FFR CT , similar to measured FFR, in any straightforward way because of the complex physiologic interplay between stenoses (Fig 3). At present, there is no accepted way to identify the lesion that contributes most to this cumulative pressure loss. Intuitively, the intrinsic impact of a given lesion should relate to FFR CT of that individual lesion, and previous data have in fact demonstrated excellent correlation between FFR CT and invasive FFR (38). However, in a recent study, it was demonstrated that FFR CT (as well as FFR) may underestimate the physiologic contribution of stenosis in vessels with serial lesions (39). An interactive revascularization FFR CT -based planner tool (HeartFlow) may more accurately predict the invasive FFR contribution of each stenosis in serial CAD (39). The

Standardized Interpretation of Hemodynamically Significant Lesions
In patients with CAD, as for measured FFR, FFR CT values decline from the ostium to the distal vessel irrespective of the vessel territory, stenosis severity, and location (14,19,(34)(35)(36). In FFR practice, it is advised that the FFR value within the throat of the lesion (which may correspond to the minimum FFR CT value) is not used clinically and that the pressure is assessed at least 2-3 cm distal to the stenosis of interest (20). Likewise, for clinical decision making, we recommend using the FFR CT value distal to the lesion. With the interactive three-dimensional coronary model tool (HeartFlow), it is possible to obtain multiple values across the vessel. Hence, after localizing the stenosis, the vessel should be serially interrogated downstream from the lesion. Notably, the FFR CT value may transiently rise immediately after the stenosis because of poststenotic vessel

Results
We recommend FFR CT values to be reported for each major coronary branch by specific coronary segments (diameter greater than 1.8 mm) using the standardized SCCT guidelines for coronary segmentation classification (21), and that the values be related to specific lesions within a given segment. Any lesion identified in the original coronary CT angiography report as a potential source of pressure loss should be specifically reported in the FFR CT report and its standard SCCT coronary segment identified. A given FFR CT value may have different therapeutic implications if located in a proximal segment as opposed to either a distal location or within a minor side branch (25,29,30). If no FFR CT value ongoing Precise Percutaneous Coronary Intervention plan (P 3 ) study (ClinicalTrials.gov: NCT03782688) investigates the diagnostic value of the FFR CT revascularization planner tool.

FFR CT Reporting
Coronary CT angiography and FFR CT uniquely provide simultaneous anatomic and functional information in a noninvasive fashion. To provide useful, actionable guidance for medical or invasive management, the FFR CT report must relate the observed anatomic coronary CT angiography findings with lesionspecific FFR CT values. The principal purpose of the report is to communicate these findings and their clinical implications (Fig 4).

Indications
The indications for the FFR CT analysis should include clinical information from the original coronary CT angiography report, as well as specific anatomic details from the impression of the report that motivated the performance of FFR CT analysis. Mention should be made of factors pertinent to the FFR CT indication and suitability for analysis, such as angiographic degree of stenosis, extent of calcifications, and overall image quality (signal-to-noise ratio, motion artifacts, luminal contrast opacification). The indications should specify the anatomic lesions from the original coronary CT angiography report that were of particular concern in ordering the FFR CT analysis. The present author group finds FFR CT testing appropriate in patients with intermediate anatomic stenosis. FFR CT values may be less than or equal to 0.80 in lesions of less than 50% diameter stenosis. Physiologic characterization with FFR CT may be relevant in a small proportion of such lesions when located in proximal coronary segments supplying a large myocardial territory because they may have prognostic implications (40). On the other hand, even high-grade anatomic lesions with stenosis severity greater than 70% or even greater than 90%, which are generally considered flow limiting, may overestimate the physiologic significance (41,42). Therefore, we commonly use FFR CT testing in the setting of more severe anatomic disease and multivessel disease to help guide decision making on downstream catheterization and potential revascularization planning (Fig 2). As with any test, the appropriateness is often determined on a case-by-case basis and commonly related to many factors beyond stenosis severity (Fig 2). Finally, because the impact of coronary occlusion on the diagnostic performance of FFR CT is unknown, we do not recommend FFR CT analysis to be prescribed in such circumstances.  the parent vessel. The impact will depend on the size of the branch relative to other vessels. Occluded segments should be identified and referenced. In recognition of the fact that FFR CT is a mathematically derived analysis rather than an actual measurement of flow and pressure, it is recommended that results be described as demonstrating low, borderline, or high likelihood of hemodynamic significance rather than ischemia (Table).

Impression
The report summary should focus on the presence of a low, borderline, or high likelihood of hemodynamic significance of the lesions identified in the impression section of the original coronary CT angiography report. In addition, any other lesion that has a borderline or high likelihood of hemodynamic significance should be reported even if it was not identified in the original coronary CT angiography interpretation. In particular, areas of diffuse coronary disease that produce low FFR CT values distal to the affected segments should be described.

Format
It is recognized that institutional requirements may dictate the specific reporting format required. Ideally, the coronary CT angiography and FFR CT reports can be combined into a single uniform report that will most clearly relate anatomic and functional information. However, it is important to interrogate the anatomy to assess the extent and severity of CAD to determine the need for FFR CT analysis. Given the time gap between the coronary CT angiography and FFR CT results, either a preliminary CT angiography report may be finalized after the FFR CT results are available or an FFR CT report may subsequently be added to the original coronary CT angiography report. Both of these formats closely incorporate the most detailed description of anatomy and functional significance with minimal repetition. If institutions require a separate freestanding report, additional details in the indications should be provided to emphasize the severity, morphology, and location of lesions suspected of causing flow limitation.

FFR CT Images
It is recommended that relevant images from the FFR CT report should be included if technically possible to more accurately convey the location of FFR CT values at a specific anatomic location. This will help other physicians understand the location and extent of the pressure loss and the location for potential confirmatory invasive FFR measurement and will facilitate medical or invasive treatment planning. Providing images combining FFR CT values and their specific location can rapidly and succinctly convey the extent of pressure loss and facilitate therapeutic decision making more easily than if textual description was offered alone.

Management Recommendations
The decision of whether FFR CT interpretation reports should contain management recommendations (ie, consideration for of 0.80 or less was reported in a given artery territory, we recommend the lowest value for that territory be reported. It is not necessary to provide FFR CT values greater than 0.80 for minimal (1%-24% stenosis) or mild (25%-49%) lesions unless located in the left main or proximal LAD or when containing high-risk plaque features, in which case FFR CT values should be provided. Any lesion with an abnormal FFR CT value should be reported even if not considered as a likely source of significant pressure loss in the original coronary CT angiography report. We recommend that an FFR CT value be provided for all moderate (50%-69%) and all severe (>70% to 99%) stenoses.
FFR CT values 0.80 or lower that are measured more than 2 cm beyond a lesion not causing a significant focal pressure loss (FFR CT > 0.80) should be reported when present in large vessels. The clinical significance of FFR CT values 0.80 or lower in the distal coronary tree remote from any focal lesion is unknown. These may be reported; however, it should be stated that the values are remote from angiographic stenosis and are of uncertain clinical significance.
In the event of serial lesions, we recommend that the value of FFR CT 10-20 mm distal to each lesion should be reported. If this is not possible, FFR CT values between lesions should be reported, including information on the distance between stenosis and the FFR CT value.
Occlusion of small vessels that were overlooked in the primary CT angiography assessment (typically involving distal segments or small side branches) may be revealed by the FFR CT analysis process. While this may or may not be clinically relevant, an occluded branch may have some slight impact on FFR CT in  Fig 1b). CAD = coronary artery disease, CTA = coronary CT angiography.
invasive coronary angiography or optimal medical therapy alone) will be determined by local institutional practices. If management recommendations are typically included in reports, note should be made that FFR CT values should not be considered in isolation but are integrated with clinical and other imaging factors such as symptoms, plaque morphology, and lesion location. This is particularly important in cases of borderline FFR CT values between 0.76 and 0.80 (Fig 2).

Limitations
The diagnostic performance and utility of FFR CT has been studied only in patients suspected of having stable CAD. At present, the use of FFR CT in patients with stents or bypass grafts, microvascular dysfunction, prior myocardial infarction, or suspected or known acute coronary syndromes cannot be recommended. FFR CT analysis cannot be performed in all patients. Coronary CT angiography-related artifacts, such as motion, misalignment, low contrast, or blooming from coronary calcification, may impair the diagnostic reliability of CT angiography and FFR CT (43)(44)(45). It is our experience that FFR CT has high diagnostic performance in patients with coronary calcification. However, our experience with FFR CT testing in patients with severe calcification (Agatston score > 1000) is limited, and in two previous studies demonstrating high diagnostic performance of FFR CT in vessels and patients with high calcium scores, the number of such patients were low (44,45). In previous multicenter studies of FFR CT diagnostic performance, CT angiographic images were not of sufficient quality for FFR CT analysis in 11%-13% of patients (15,16), whereas in more recent single-center studies that assessed the clinical utility of FFR CT , less than 4% of the patients did not meet the image quality requirements (10)(11)(12)14).

Conclusion
By virtue of the complexity of the FFR CT analysis providing information on pressure and flow across the entire coronary tree, standardized criteria on interpretation and reporting of the FFR CT analysis results are of crucial importance both in context of clinical adoption of the test and in future research. For assessment of the hemodynamic significance of lesions, we recommend using the FFR CT value 10-20 mm distal to the lower border of the stenosis. For clinical decision making, we recommend a dichotomous interpretation strategy be considered only in lesions with FFR CT greater than 0.80 or less than or equal to 0.75, whereas in patients with FFR CT ranging between 0.76 and 0.80, additional risk stratification information is needed. The results of FFR CT must always be evaluated in their clinical context, taking into account patient symptoms, the coronary anatomy, and suitability of revascularization. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: has received unrestricted institutional research grants from Siemens and HeartFlow; travel expenses covered at TCT 2017 by HeartFlow (no personal payment). Other relationships: disclosed no relevant relationships. T.A.F. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: has been an invited speaker sponsored by HeartFlow. Other relationships: disclosed no relevant relationships. R.D.S. disclosed no relevant relationships. M.G.R. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: author is consultant for HeartFlow. Other relationships: disclosed no relevant relationships. B.K. Activities related to the present article: has been an invited speaker at symposiums sponsored by Canon Medical, Medtronic, and St Jude, and has received research funding from Canon Medical. Activities not related to the present article: disclosed no relevant relationships. Other relationships: disclosed no relevant relationships. J.M.J. disclosed no relevant relationships. K.N. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: unrestricted institutional research support from HeartFlow, Siemens, GE, and Bayer; coverage of travel expenses to the present study results at the TCT in 2018 by HeartFlow; Steering committee ADVANCE registry for HeartFlow (unpaid). Other relationships: disclosed no relevant relationships. K.M.C. Activities related to the present article: institution receives grant from HeartFlow. Activities not related to the present article: Medical Advisory Board for HeartFlow (no personal compensation); institution receives grants from HeartFlow. Other relationships: disclosed no relevant relationships. N.P.S. disclosed no relevant relationships. H.M. disclosed no relevant relationships. J.L. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: has received speaker honoraria from GE Healthcare, served as a consultant for Edwards Lifesciences, and served as a consultant and has stock options in Circle CVI and HeartFlow; has core laboratory contracts with Edwards Lifesciences, Medtronic, Abbott, Noevasc; research support from Edwards. Other relationships: disclosed no relevant relationships. G.R. disclosed no relevant relationships.