Detecting lesion-specific ischemia in patients with coronary artery disease with computed tomography fractional flow reserve measured at different sites

Coronary computed tomography angiography (CCTA) has been widely accepted as a reliable noninvasive assessment modality for excluding the presence of coronary artery significant obstructive disease (≥ 50% luminal narrowing) in low-to-intermediate-risk populations with a high negative predictive value [1‐3]. However, its diagnostic specificity for assessing obstructive coronary artery disease (CAD) is still suboptimal. For example, in severely calcified plaques, luminal stenosis is often overestimated owing to calcium blooming. More importantly, CCTA cannot provide hemodynamic information to determine whether a particular stenotic lesion is associated with hemodynamically significant ischemia i.e., lesion-specific ischemia [4]. While exercise treadmill testing and stress echocardiography can assess overall ischemic burden, they are limited in locating lesion-specific ischemia on a per-vessel basis [5]. Therefore, accurate and prompt assessment of lesion-specific ischemia is critical in the management of stable CAD to improve its clinical outcomes and benefits as myocardial blood flow can be improved by medical therapy or revascularization procedures such as percutaneous coronary intervention (PCI) or coronary artery bypass graft surgery (CABG) [6].

Fractional flow reserve (FFR) is the gold standard for the assessment of lesion-specific ischemia to guide revascularization in stable CAD patients [7]. Clinically, FFR is measured along with invasive coronary angiography (ICA) by placing a pressure guidewire beyond a stenotic lesion and measuring the ratio of mean distal coronary pressure to mean aortic pressure ratio under conditions of adenosine infusion to the maximum hyperemia. At present, invasive FFR has become a cornerstone in determining lesion-specific ischemia and appropriate decision-making [8]. Whereas, invasive FFR requires additional expensive instruments, and clinical application of invasive FFR-informed treatment decision-making is relatively limited. Only 10% to 20% of revascularization procedures have incorporated invasive FFR results into the treatment decisions [9].

Recently, FFR can be noninvasively calculated from anatomical CCTA data based on computational fluid dynamics (CFD) [10]. This CT-based FFR (FFR_CT) does not need additional imaging and vasodilator administration. FFR_CT, the ratio of the maximum coronary blood flow through a stenotic artery to the blood flow assumed to be free of stenosis in that artery, has been validated to be useful in diagnosing and excluding lesion-specific ischemia [11‐13]. Compared with invasive FFR, FFR_CT can provide FFR information almost at any site along the entire epicardial coronary artery tree [14]. Whereas the diagnostic accuracy of FFR_CT was previously determined by comparing a single measurement site corresponding to invasive FFR at a specific location within the coronary artery [11‐13]. Inconsistent measurement sites between FFR_CT and invasive FFR can lead to inconclusive diagnostic results and confuse decision-making [15, 16]. Therefore, how to select an appropriate site along the coronary artery tree to measure FFR_CT is clinically relevant for the management of CAD. However, the optimal site to measure FFR_CT for a target lesion remains to be adequately determined.

In this single-center retrospective cohort study, CAD patients who had undergone both CCTA and invasive FFR were included. The diagnostic performances of FFR_CT measured at different sites at 1 cm, 2 cm, and 3 cm distal to the lower border of the target lesion of the artery (FFR_CT-1 cm, FFR_CT-2 cm, FFR_CT-3 cm), and the lowest FFR_CT value at the distal vessel tip in which diameter larger than 1.5 mm (FFR_CT-lowest) were assessed and compared with invasive FFR as the reference standard. The purpose of this study was to determine the optimal site to measure FFR_CT for a target lesion in detecting lesion-specific ischemia in CAD patients.

In this study, FFR_CT measured at four different sites along the coronary arteries was used to identify lesion-specific ischemia using invasive FFR as the reference standard. Our results showed that FFR_CT-2 cm had the highest accuracy (87.5%) and AUC (0.857) in identifying lesion-specific ischemia and FFR_CT-lowest has the highest sensitivity and NPV than FFR_CT-1 cm, FFR_CT-2 cm, FFR_CT-3 cm. CCTA (DS ≥ 50%) combined with either FFR_CT-1 cm, FFR_CT-2 cm, FFR_CT-3 cm, or FFR_CT-lowest showed higher AUC than CCTA alone in detecting lesion-specific ischemia.

Commonly, invasive FFR measurement is performed at one position selected 2-to-3 cm distal to the target lesion along coronary artery tree during ICA [19]. Unlike invasive FFR, FFR_CT values can be obtained along the entire coronary arterial tree. Inappropriate FFR_CT measurement can mislead the clinical decision. However, there is no consensus about the best location to measure FFR_CT in clinical practice. Previously, Cami et al. [16] used a CFD-based algorithm to measure FFR_CT value 10.5 mm (IQR 7.3–14.8 mm) distal to the stenosis in 26 patients with proximal LAD stenosis and found that it was a reliable location for measuring FFR_CT using invasive FFR measured 2-3 cm distal to the stenosis as reference standard. An expert panel [23] based on the finding of Cami et al. [16] advised to use the minimum FFR_CT values measured 1 to 2 cm distal to the stenosis as the result to judge the coronary lesion-specific ischemia. However, the measurement site proposed by Cami et al. [16] was defined as the distance from the distal end of the target stenosis to the place where FFR_CT declined to a lower plateau. In a similar study by Omori et al. [15], the diagnostic performance of FFR_CT measured at 1 to 2 cm distal to the stenosis was also found to be higher than that of FFR_CT-lowest (0.86 vs. 0.80, p = 0.002) in identifying lesion-specific ischemia using invasive FFR as reference standard. In addition, Nozaki et al. [18] used a CFD-based algorithm and found that the AUC of FFR_CT-2 cm was higher than that of FFR_CT-lowest (0.80 vs. 0.68, p = 0.002) in identifying lesion-specific ischemia and was comparable with that of FFR_CT-1 cm (0.80 vs. 0.79, p = 0.73). In a recent study by Chen et al. [24] where a machine learning-based algorithm was applied, the AUC of FFR_CT-2 cm was found to be comparable with that of FFR_CT-1 cm (0.91 vs. 0.91, p = 0.663) and was higher than that of FFR_CT-3 cm (0.91 vs. 0.88, p = 0.002) and FFR_CT-4 cm (0.91 vs. 0.88, p = 0.008) in identifying lesion-specific ischemia using invasive FFR as reference standard. Based on these results, FFR_CT measured at 1-to-2 cm distal to the stenosis is better than FFR_CT-lowest in identifying lesion-specific ischemia in patients with CAD.

In our study, invasive FFR was also used as the reference standard. Our results showed that there was a good correlation (r = 0.80, p < 0.001) and a very mild difference (mean difference 0.0001, 95% limits of agreement: -0.1222 to 0.1220) between invasive FFR and FFR_CT-2 cm. Further ROC analysis showed that the diagnostic performance of FFR_CT measured at 2 cm distal to the target lesion (FFR_CT-2 cm) was higher than that measured at 1 cm distal to the target lesion (FFR_CT-1 cm) and that measured at the vessel terminus (FFR_CT-lowest) (AUC: 0.857, 0.768, 0.770, respectively), and was comparable with FFR_CT-3 cm (AUC: 0.856). These findings suggest that FFR_CT measured at 2 cm distal to the target lesion had the highest performance in identifying lesion-specific ischemia. Overall, our results are consistent with the findings of previous study [15, 16, 18, 24]. In our study, FFR_CT was measured at four different sites along the same artery and our results showed that FFR_CT measured at 2 cm distal to the target lesion are the optimal site for FFR_CT measurement, which is in line with previous studies [15, 16, 18, 24]. It should be noted that the measurement site of the invasive FFR and the definition of the target lesion were not totally consistent between our study and previous studies [15, 16, 18, 24]. Indeed, the invasive FFR was measured 2 to 3 cm distal to the target lesion in our study which was based on the recommendation as previously described [19],which is as same as the study by Cami et al. [16]. Nonetheless, Omori et al. [15] used invasive FFR measured 2–4 cm distal to the target lesion as the reference standard, Nozaki et al. [18] used invasive FFR measured distal to the stenosis as far as possible as the reference standard. Chen et al. [24] used invasive FFR which was measured at a minimum of 2 cm distal to the stenosis in vessel segments ≥ 2 mm as the reference standard. As regards the target lesion, the definition of the target lesion of a serial lesion in the study by Nozaki et al. [18] was similar to our study, i.e., the most distal lesion in the vessel with 30%-90% diameter stenosis selected as the target lesion. However, Omori et al. [15] and Chen et al. [24] selected the most severe stenosis in a serial lesion as the target lesion. These differences likely result in the mild discrepancy in the diagnostic performances of FFR_CT measured at 1 to 2 cm distal to the target lesion. It is known that there is a gradual decrease in FFR_CT from proximal to distal along the vessel even without focal stenosis [16]. The lowest value of FFR_CT measured at 1 to 2 cm distal to the target lesion probably is not significantly different from FFR_CT measured 2 cm distal to the lower border of the target lesion. Thus, it is reasonable that 2 cm distal to the target lesion could be used as the exact measurement site for FFR_CT. Our results indicate that FFR_CT-2 cm is the optimal for identifying lesion-specific ischemia. This finding might impact the clinical decision-making and patient outcomes. For example, if a patient has a lesion with 30–90% vessel diameter stenosis while no lesion-specific ischemia as determined by FFR_CT-2 cm, this patient could avoid invasive FFR and unnecessary interventional revascularization.

Due to the presence of a gradual decrease in FFR_CT from proximal to distal along the vessel even without focal stenosis [16], FFR_CT measures at different sites along the same coronary arterial might have different clinical indication. Our study showed that FFR_CT-1 cm has the highest specificity (93.6%) in diagnosing lesion-specific ischemia compared with FFR_CT-2 cm, FFR_CT-3 cm and FFR_CT-lowest, but had the lowest sensitivity (60%); FFR_CT-lowest has the highest sensitivity (88%) and NPV (91.2%) in identifying lesion-specific ischemia but had the lowest specificity (66%) and PPV (58%). 18% vessels positive for FFR_CT-lowest were reclassified as negative when determined by FFR_CT-2 cm. These results indicate that FFR_CT-1 cm could underestimate the severity of the lesion and the FFR_CT-lowest could overestimate the severity of the lesion. Similarly, Kueh et al. [25] also found that FFR_CT-lowest overestimated the severity of the lesion when compared to FFR_CT measured within 20 mm of the stenotic lesion in identifying lesion-specific ischemia and false positive results of FFR_CT-lowest could be effectively reclassified by FFR_CT measured within 20 mm of the stenotic lesion. This might be associated with the gradual decrease in FFR_CT from proximal to distal along the vessel even without focal stenosis, which is more significant with FFR_CT than with invasive FFR due to pressure loss by frictional losses according to Poiseulle’s equation [16]. Taken together, FFR_CT-1 cm and FFR_CT-lowest both are not optimal site for FFR_CT measurement and cannot be used as FFR_CT result in clinical decision-making.

CCTA can overestimate the severity of stenosis in CAD. It has been reported that less than a half of severe coronary artery disease diagnosed by CCTA can really result in lesion-specific ischemia [26]. This raised the concern that the widespread use of CCTA may encourage unnecessary ICA [27]. In our study, the AUC of CCTA (DS ≥ 50%) alone had only a moderate diagnostic performance (AUC = 0.576) for identifying lesion-specific ischemia, which was lower than that of FFR_CT-1 cm, FFR_CT-2 cm, FFR_CT-3 cm, and FFR_CT-lowest. When CCTA was combined with either FFR_CT-1 cm, FFR_CT-2 cm, FFR_CT-3 cm, or FFR_CT-lowest, its AUC was improved in identifying lesion-specific ischemia. These results suggest that the addition of FFR_CT can improve the diagnostic performance of CCTA in identifying lesion-specific ischemia and may reduce unnecessary ICA, thereby enhance its role as a gatekeeper for ICA.

Our study has some limitations. First, it is a retrospective study from a single center and the sample size was not large. It has selection bias inherent in a retrospective study. Second, patients with previous revascularization were excluded from the study. Thus, the validity of FFR_CT parameters in these patients needs further investigation. Third, a per-vessel analysis was performed in our study. In some patients, more than one vessel was included for analysis. Fourth, FFR_CT can be calculated using a machine learning-based algorithm or a CFD-based algorithm. In our study, only CFD-based algorithm was investigated. Fifth, the long-term effect of FFR_CT on the adverse cardiac events was not investigated in this study. Further prospective clinical studies are warranted to validate the impact of FFR_CT on the clinical outcome in patients with CAD.

Our study demonstrates that FFR_CT measured at 2 cm distal to the lower border of the target lesion is the optimal measurement site in identifying lesion-specific ischemia in patients with CAD. The addition of FFR_CT to CCTA can improve the diagnostic performance in in identifying lesion-specific ischemia. FFR_CT-2 cm could be used as an alternative imaging biomarker in identifying lesion-specific ischemia. The use of FFR_CT-2 cm can avoid unnecessary invasive FFR in a patent who has a 30%-90% diameter stenosis but negative FFR_CT-2 cm. This may aid in the decision-making in patients with CAD. In our study, the invasive FFR was used as a reference standard to assess the diagnostic performances of FFR_CT measured at different sites. Whereas the sample size was relatively small. Future multicenter prospective studies are needed to validate the clinical role of FFR_CT-2 cm.