Today, I review, link to and excerpt from:
CT Fractional Flow Reserve: A Practical Guide to Application, Interpretation, and Problem Solving [PubMed Abstract] [Full-text HTML] [Full-Text PDF] [Supplemental Material]. Radiographics. 2022 Mar-Apr;42(2):340-358. doi: 10.1148/rg.210097. Epub 2022 Feb 4.
All that follows is from the above resource.
Who Needs FFRCT?
FFRCT can be performed at all coronary CTA examinations but determined on the basis of clinical and coronary CTA findings. Coronary stenosis at CTA can be classified as low-, intermediate-, or high-risk anatomy. Low-risk anatomy is defined as when coronary CTA is either normal or the stenosis is less than 30%. These patients generally do not require FFRCT, and the stenosis can be managed with optimal medical therapy (OMT). High-risk anatomy is defined as significant left main (≥50%) stenosis, high-grade (≥70%) left anterior descending artery (LAD) stenosis, or three-vessel stenosis. These patients also do not generally require FFRCT and can be referred for ICA. However, FFRCT may be useful in this group in some circumstances, specifically for revascularization planning, especially in a multivessel setting as discussed in the following section. Occlusive lesions at CTA do not require FFRCT as the diagnostic performance of FFRCT in this setting is not well known. In patients with occlusive lesions who are not fit for surgery and are being evaluated for PCI, FFRCT may be still performed to provide physiologic information on other vessels. Some institutions perform FFRCT even in single-vessel disease greater than 70% and safely defer ICA if the FFRCT is greater than 0.8 with a low event rate.
Intermediate-risk anatomy describes the presence of one or two intermediate stenotic lesions; that is, 30%–69% luminal stenosis reduction or greater than or equal to 70% stenosis vessels other than the left main or LAD arteries.
FFRCT is most useful in the intermediate-risk category for determining subsequent management (OMT vs ICA and revascularization). If the FFRCT is normal (>0.8) in this category, the patient can undergo management with OMT
(Fig 4) (33). ICA and revascularization can be safely deferred in this group, with lower rates of myocardial infarction, major adverse cardiovascular events (MACE), cardiac death, and revascularization after 90 days (34). However, if the FFRCT is abnormal (≤0.75) in the intermediate-risk anatomy, the patient may require ICA and revascularization (Figs 5, E2). Abnormal FFRCT in small vessels, distal vessels, or side branches can be managed medically since they are not suitable targets for revascularization (35). In moderately stenotic lesions, FFRCT improves the net reclassification of CAD, that is, it converts false-positives to true negatives in up to 68% of patients who do not have lesion-specific ischemia (36). The prevalence of nonobstructive disease at ICA is only 12% with the use of FFRCT compared with 73% in patients who undergo ICA without noninvasive testing (37). Reserving ICA only for patients with FFRCT of 0.8 or less would decrease nonobstructive disease at ICA by 44% (38), making CT a more effective gatekeeper for ICA without increasing MACE (37,38).
Borderline FFRCT Values
FFRCT values between 0.76 and 0.80 are considered borderline or in the gray zone. Ischemia is present in 55% of these cases (Fig 6) (33). The type of management needed for borderline FFRCT values is debatable, with options including OMT, an additional functional test, ICA, or revascularization. We favor a nuanced personalized management plan in this scenario, with consideration of other factors to risk-stratify the patient. The vessel territory (LAD vs others), location (proximal vs distal, main vessel vs side branch), number (multiple vs single), ∆FFRCT, plaque burden, and high-risk plaque features (low attenuation, napkin ring, spotty calcification, positive remodeling) should all be considered (35). ICA may be recommended only for patients with additional high-risk features, whereas OMT would be suitable for the other risk categories (35). Some groups suggest OMT for 3 months for all patients with borderline values in a stenotic lesion, as data show that there is no improved outcome following revascularization in these patients (39). If symptoms persist at 3 months, the patient can be referred for ICA, whereas if the symptoms resolve at 3 months, OMT can be continued (Fig 7) (33). One study in a small number of patients recommended using a lower cutoff of 0.71 to call a lesion borderline (40). Using a cutoff of 0.70 or less for abnormal FFRCT increases the specificity and positive predictive value of FFRCT.
Figure 7. Flowchart shows management of atypical chest pain and the integration of FFRCT. A patient with low-risk anatomy at coronary CTA is referred for OMT, whereas a patient with high-risk anatomy at CTA is directly referred for ICA. Patients with intermediate-risk anatomy or indeterminate stenosis grade are referred for FFRCT. If the FFRCT is abnormal, the patient undergoes ICA; if the FFRCT is normal, the patient receives OMT. If the FFRCT is borderline, some groups advocate OMT with a 3-month follow-up (3-mo FU). If symptoms persist at 3 months, the patient is then sent for ICA, but if symptoms resolve in 3 months, OMT is continued. Another option for borderline FFRCT values is to evaluate for additional risk factors such as vessel territory, location, number of lesions, ∆FFRCT, plaque burden, and high-risk plaque features. ICA is recommended only for patients with additional high-risk features, whereas OMT is suitable for the other risk categories.
FFRCT in Multivessel Disease and Revascularization Planning
Multivessel disease is associated with incomplete revascularization in 25% of patients and a higher MACE rate (41). Performing an accurate noninvasive test is beneficial in minimizing the hazard of incomplete revascularization. Because of its low specificity, coronary CTA alone is usually inadequate to provide this information. CTA and FFRCT provide valuable anatomic and physiologic information of multiple vessels, identifying patients who may benefit from ICA and revascularization, and decrease inappropriate revascularizations at an overall lower cost than ICA (Fig 8) (42,43). Even in a high-risk population, ICA was cancelled more frequently, and there was a higher PCI/ICA ratio on the basis of FFRCT (75%) versus CTA alone (45%) (44). Data from the Assessing Diagnostic Value of Noninvasive FFRCT in Coronary Care, or ADVANCE, registry show that FFRCT modified the treatment strategy in 67% of patients with fewer revascularizations at 1 year in the FFRCT group greater than 0.8 than in the FFRCT group less than 0.8 (38.4%) (45).
Discordant Coronary CTA and FFRCT Values
Discordant values between coronary CTA and FFRCT can be seen (ie, abnormal FFRCT in a mild stenotic lesion and normal FFRCT in a severe stenotic lesion) (53). This is not surprising since anatomy and physiology do not always have a linear relationship. In addition, there are multiple factors that determine a low FFRCT, including minimal luminal diameter; length, shape and location of the lesion; reference vessel size proximal and distal branches; serial lesions; diffuse plaque; and the myocardial mass (53,54). From ICA results, we know that a severely narrowed coronary artery does not always result in bad consequences, as there may be good collateralization or the vessel distal to the obstruction may be supplying only a small amount of the myocardium.
On the other hand, lesions that do not appear severe may cause more ischemia if they are serial or involve a long segment of the vessel. Smaller luminal narrowing in larger proximal vessels causes greater physiologic derangement than does larger narrowing in smaller vessels. At CTA, 21% of patients with mild stenosis in the 30%–49% range have positive FFRCT values, more common in the LAD (Fig 10). On the other end of the spectrum, 28% of patients with severe stenosis in the 71%–90% range can have negative FFRCT values (Fig 11) (53). A good match between CTA severity and FFRCT is seen only for lesions with less than 30% or greater than 90% stenosis (53,54). Hence, interpretation of the FFRCT is always made in the context of clinical history, symptoms, location of abnormal FFRCT, and revascularization target.
Patients with severe stenosis but normal FFRCT are managed medically since there is no ischemia to treat. There is no clear consensus on management of patients with abnormal FFRCT without significant stenosis, but they require further correlation with clinical features, anatomic features, and suitability for revascularization. Proximal and significantly symptomatic patients may undergo ICA for further management, whereas distal lesions without a suitable revascularization target can be managed by medical therapy and followed up. FFRCT is occasionally abnormal without an obvious CTA abnormality. This is usually encountered in a vessel that was not the original reason for sending the study for FFRCT. It is possible that a subtle lesion may have been missed in the original CTA interpretation, including occlusion of small branches. The significance of this is unknown and may have an impact on proximal FFRCT values. Modeling errors can also cause this, as discussed in a subsequent section of this article. The management of this situation should be based on the principles described previously, that is, to correlate management with anatomic and clinical features.
Gradual Decrease of FFRCT and Low Distal FFRCT
Gradual decrease of FFR from the proximal to distal course of a normal coronary tree is demonstrated both at coronary CTA and ICA. In normal and less than 25% stenosis, there is approximately a 0.08–0.13 difference in the FFRCT between the ostium and distal-most vessel. The FFR and pressure decrease are dependent not only on diameter reduction but also on length of the lesion. A gradual pressure drop may also be seen in diffuse atherosclerosis without a focal obstructive stenosis (Fig 12).
Serial stenosis and myocardial hypertrophy are other causes of gradual pressure drop. These entities, particularly atherosclerosis, can also cause an abnormal value in the distal-most (nadir) vessel of less than 0.8 (Fig 13). Sometimes diffuse atherosclerosis without focal stenosis and gradual pressure drop can cause ischemia (55). The management of such cases is complex and challenging, but there is no suitable target for revascularization. One study showed the safety of OMT in the intermediate follow-up of patients with low nadir FFRCT (56). However, note that abnormal distal FFRCT may be significant when there is a proximal lesion without a significant drop immediately distal to the lesion, provided that this lesion is supplying a large myocardial territory (35).
Serial Stenosis
Assessing FFRCT is challenging in serial stenosis or diffuse disease, which can be seen in 25%–40% of patients, especially those with diabetes. Assessment of the individual contribution or hemodynamic significance of each lesion is challenging owing to the complex physiologic interplay and difficulty in quantifying the contribution of cumulative pressure loss (Fig 14). The most proximal lesion results in a pressure drop distal to it owing to increased resistance, which is magnified in hyperemic conditions. In invasive FFR with another downstream stenosis or diffuse disease, the additional resistance makes it challenging to achieve maximal hyperemia, thus underestimating the true individual stenosis contribution. ∆FFR should theoretically be able to estimate the impact of an individual lesion. However, a recent study showed that ∆FFRCT and invasive ∆FFR also underestimate the contribution of individual stenosis in serial lesions (57).
Unlike ICA, CT provides geometric information on the serial lesions, which can be used to model the complex hemodynamic interplay between each serial stenosis (57). To evaluate the significance of each serial stenosis, an FFRCT-based revascularization planner tool can be used, which is an idealized model that simulates postrevascularization pressure loss. The effect of revascularization on a specific lesion can be evaluated and the FFRCT values reanalyzed. This PCI planner tool accurately predicts true FFR contribution of each stenosis (57), correlating well with invasive FFR obtained after PCI (58).
Pitfall Modeling
FFRCT relies on accurate anatomic and physiologic modeling. Any error in modeling will result in downstream errors in FFRCT values and interpretation. Therefore, it is good practice to correlate FFRCT values with the original coronary CTA images as well as the anatomic model of FFRCT, particularly with discrepant values. For example, a focal motion artifact may affect modeling and cause artifactual decrease in FFRCT (Fig 16). If the FFRCT anatomic modeling is tighter than the extent of stenosis seen at coronary CTA, the FFRCT values may be artifactually abnormal (Fig 17). If such discrepancy is noted, the modeling should be redone and FFRCT reevaluated. On the flip side, one should always check if an abnormal stenotic area at CTA has been modeled appropriately in FFRCT; otherwise, artifactual normal FFRCT values may be returned (Fig 18). Sometimes the anatomic area may have been left out of modeling, especially distal vessels or small branch vessels (<2 mm) either owing to their small size or because the anatomic area is clipped out of modeling (Fig 19). Therefore, FFRCT may not identify the stenosis and/or occlusion in small branch vessels. Modeling abnormalities without an anatomic target for revascularization can be managed with OMT. Tables 1 and 2 summarize the key features and pitfalls of FFRCT.
Challenges of FFRCT
FFRCT does not require any additional radiation dose or change in CT protocol. However, it requires good-quality contrast-enhanced CT images without artifacts or noise and nitroglycerine administration. Approximately 2.9%–33% of CT examinations are rejected for FFRCT analysis because of suboptimal quality (74). Motion artifact accounts for most rejections (78%). Other image quality issues include misalignment, metallic artifact, blooming, noise, and suboptimal contrast (Figs E16–E20). Technical reasons for rejections include data characteristics such as (a) not including the whole heart in the field of view, (b) section thickness 1 mm or greater, (c) section spacing 1 mm or greater, (d) pixel spacing 0.5 mm or greater, (e) missing sections, (f) series thickness less than 5 cm, and (g) duplicate data. High cost is also an issue, with HeartFlow costing approximately $1100 U.S. dollars per patient. Currently, this is reimbursed by Medicare and several insurance companies. Processing times for HeartFlow are long, ranging from 1 to 5 hours, which limits its use in the acute setting. On-site solutions may be faster, but they are not yet validated or commercially available. On-site solutions will also require radiologist involvement and time.
Conclusion
FFRCT improves the specificity of coronary CTA in the evaluation of CAD by providing the hemodynamic significance of a stenotic lesion. This can be used to effectively triage patients for ICA, decrease nonobstructive disease in ICA, and guide revascularization. For lesion-specific ischemia, FFRCT is measured 2 cm distal to the stenosis. FFRCT is always interpreted in correlation with clinical and anatomic features. Knowledge of pitfalls and limitations is essential to avoid misinterpretation.
