Coronary computed tomography angiography (CCTA) has evolved as one of the most frequently utilized modalities for the evaluation of chest pain and the diagnosis of obstructive coronary artery disease in an outpatient (stable chest pain) or emergency/inpatient setting (unstable chest pain). However, severe coronary artery calcification (CAC) is known to be one of the most significant limitations of the technique, leading to false positive findings, mostly secondary to blooming artifact. Fractional flow reserve derived from coronary CT angiography (FFR-CT) has demonstrated acceptable correlation and agreement with invasive FFR, and due to its higher specificity and negative predictive value, enhances CCTA’s diagnostic performance. There are published reports showing that the diagnostic performance of FFR-CT is affected less by severe CAC compared with CCTA alone. In this review, we present the current evidence of how FFR-CT can increase the accuracy of CCTA for the evaluation of calcified coronary arteries.
Central Illustration Abbreviations: CACS = coronary artery calcium score; CCTA = coronary computed tomographic angiography; FFR = invasive fractional flow reserve; FFR-CT = fractional flow reserve derived from CCTA; ICA = invasive coronary angiography.
Coronary computed tomography angiography (CCTA) is an important diagnostic tool for the evaluation of obstructive coronary artery disease (CAD). It is particularly effective in assessing stable patients with low-to-intermediate pretest probability for CAD. The 2021 American College of Cardiology/American Heart Association Chest Pain Guidelines recommend CCTA as the initial test for this patient group to exclude significant CAD. Thanks to a high negative predictive value (NPV), CCTA reduces the need for invasive coronary angiography (ICA) while ensuring at least comparable safety outcomes with other noninvasive coronary evaluation modalities. ,
Coronary artery calcification (CAC) is a significant marker of atherosclerosis of coronary arteries. CAC is quantified using the Agatston score (AS). An AS score greater than 400 is indicative of severe calcification. Subclinical atherosclerosis from calcified or noncalcified plaque is common among middle-aged individuals, with studies reporting that over 40% of asymptomatic patients who undergo scanning have some degree of CCTA-proven atherosclerosis commonly accompanied by AS greater than 0. Among asymptomatic young adults aged 30 to 49 years undergoing CAC assessment for risk stratification, 21% to 34% are found to have a CAC score greater than zero. , Traditional cardiovascular risk factors such as age, hypertension, diabetes, and smoking are strongly associated with the presence and progression of CAC. Elevated CAC score is a well-established predictor for mortality and MACE in individuals with symptomatic and those with asymptomatic CAD. ,, Thus, CAC scoring (CACS) can risk-stratify patients across the cardiovascular risk spectrum and guide appropriate pharmacologic interventions.
However, CCTA diagnostic accuracy is often undermined in patients with calcified plaques and high CAC score. Calcium blooming artifacts usually lead to overestimation of stenosis significance, increasing the false positive results and lowering significantly the specificity and positive predictive value (PPV) of CCTA. Fractional flow reserve derived from coronary CT angiography (FFR-CT) is an emerging tool for noninvasive measurement of FFR and provides functional evaluation of CAD. FFR-CT possesses superior diagnostic performance compared with CCTA alone because it can predict lesion-specific ischemia. This is particularly important in patients with intermediate stenosis significance, where assessing the hemodynamic significance of stenosis can act as the decisive factor and gatekeeper for referral or not for ICA.
Given that FFR-CT is a product of the original CCTA image data, calcification in the coronaries could in theory affect the FFR-CT interpretation as well. This narrative review aims to elucidate the importance of FFR-CT in the evaluation of calcified coronary arteries and how FFR-CT can further optimize the gatekeeping ability of CCTA for referral for coronary angiogram among patients with severe CAC.
FFR-CT Diagnostic Performance/Clinical Utility
FFR-CT allows an estimation for lesion-specific ischemia and is supported by guidelines to evaluate patients with intermediate stenosis (coronary artery stenosis of 40% to 90% in a proximal or middle segment on CCTA,) with stable or acute chest pain. Among patients with intermediate lesions on CCTA which are inconclusive for the presence of obstructive CAD, FFR-CT > 0.8 justifies guideline-directed medical therapy over further evaluation with ICA.
Numerous studies have evaluated the diagnostic performance and additive value of FFR-CT. ,, In the PRECISE trial, 2,100 patients with stable chest pain were randomized to either CCTA with selective FFR-CT (for intermediate- and high-risk patients) and deferred testing (for low-risk patients) vs usual care. , The group undergoing FFR-CT experienced a reduction in “unnecessary” catheterizations (catheterizations where no significant obstruction was found) and an increase in revascularizations. Similar outcomes were seen in the TARGET and the FORECAST trials, where implementation of FFR-CT was shown to reduce the need for invasive coronary angiography as well as the proportion of “unnecessary” catheterizations. , The ADVANCE registry included clinically stable patients with documented stenosis of at least 30% on CCTA. FFR-CT resulted in revision of the clinical management in 66.9% of patients when compared with the initial CCTA-based treatment plan. Moreover, rates of nonobstructive coronary disease (stenosis <50% in ICA) were significantly lower in patients with FFR-CT ≤ 0.80 (14.4%) compared with patients with FFR-CT > 0.80 (43.8%), (p <0.001). Interestingly, no death, myocardial infarction (MI) or major adverse cardiovascular events (MACE) occurred within 90 days in any patient whose FFR-CT was > 0.80, while 14 (0.3%) death/MI and 19 (0.6%) MACE occurred in patients with an FFR-CT ≤ 0.80. Another study evaluated 224 patients with CCTA-only approach and 136 patients with FFR-CT performed on top the CCTA and assessed the difference in clinical decision-making and patient prognosis. Adding FFR-CT to CCTA led to reduced invasive diagnostic tests and fewer revascularizations, with similar MACE rates between the 2 groups in 1 year. Lastly, a meta-analysis of 23 studies revealed that FFR-CT ≤ 0.8 had per-patient sensitivity of 88% (95% Confidence Intervals (CI): 85% to 90%) and specificity of 79% (95% CI: 71% to 85%), for ischemia diagnosis with invasive FFR used as the reference standard, and per-vessel sensitivity of 85% (95% CI: 82% to 87%) and specificity of 81% (95% CI: 76% to 85%). Finally, in the prospective multicenter PLATFORM trial, CCTA and selective FFR-CT in symptomatic patients with suspected CAD and planned ICA were associated with equivalent clinical outcomes and quality of life, and lower costs, compared with usual care over 1-year follow-up.
CCTA and Calcification of the Coronary Vessels
CCTA diagnostic performance is affected by the presence of calcified coronary arteries, since they can obscure the visualization of the coronary lumen. Blooming artifacts caused by calcification lead to overestimation of the lesion and high false positive rates. The diagnostic accuracy of CCTA significantly decreases as the calcification burden increases. ,,,
In a study of 360 symptomatic patients with acute and stable anginal syndromes– patients with impaired image quality due to calcification were not excluded from the analysis– CCTA (with invasive FFR used as reference) specificity was shown to decrease and sensitivity to increase in the presence of severe calcifications. The Prospective Multicenter ACCURACY trial assessed the diagnostic performance of CCTA after stratification of patients by CACS, with a threshold of 400 Agatston units. While sensitivity remained high in both subgroups, specificity was reduced in patients with CACS ≥ 400 (specificity = 52.6% in CACS ≥ 400 vs 86.3% in CACS < 400; p = 0.0003). Similarly, another study reported a statistically significant reduction in specificity of CCTA in patients with CACS > 400. In patients with CACS < 10, the specificity was 89% (95% CI:76% to 96%), whereas in those with a CS > 400, the specificity dropped to 20% (95% CI: 5% to 49%). In a prospective, multicenter study investigating the characteristics of CCTA that lead to inaccuracy in diagnosing CAD, Yan et al. found that the presence of coronary calcification was the only patient-level factor independently associated with false positive diagnosis (OR = 5.22; 95% CI: = 1.11 to 24.61; referent: zero Agatston score). Finally, a meta-analysis of 7 studies aimed to determine the diagnostic accuracy of 64-slice computed tomography coronary angiography including 1190 patients revealed statistically significant drop in specificity (from 88.5% to 42%) and NPV of CCTA in patients with CACS < 400 versus CACS ≥ 400, while the sensitivity and accuracy of the test remained unchanged. The abovementioned data question the ability of CCTA to act as a gatekeeper for ICA in patients with severe CAC.
FFR-CT and Calcification of Coronary Arteries
Diagnostic performance
While CCTA tends to lose its discriminative ability and specificity as the calcification score increases, previous data suggest that FFR-CT results are affected to a lesser extent. , Relevant studies are presented in Table 1 . One study reported that the presence of coronary calcified plaque (vs not) was a factor significantly associated with false positivity of FFR-CT results compared with invasive FFR. However, there was no difference between different calcification degrees in terms of false positivity rates.
Table 1
Studies presenting impact of calcification on FFR-CT diagnostic performance
| Study | Year | Population | Agatston scores (AS) | N Patients/Vessels | Objective | Main outcomes regarding FFR-CT and calcification |
|---|---|---|---|---|---|---|
| Ding et al. | 2023 | Patients with suspected CAD | NR | 303/324 | Diagnostic performance of a novel computational fluid dynamics (CFD)-based algorithm for in situ FFR-CT |
|
| Tao et al. | 2022 | Patients with suspected or known CAD who underwent CCTA | per- patient: 212.0 (IQR: 59 to 626) | 128/NR | Effect of coronary artery calcification on diagnostic performance of FFR-CT |
|
| Zhao et al. | 2022 | Patients with clinically suspected CAD and scheduled invasive coronary angiography (ICA) | per-patient: 87.0 (range: 0 to 2895.0) per-vessel: 41.0 (range: 0 to 2012.6) | 305/348 | Effect of coronary artery calcification on diagnostic performance of FFR-CT |
|
| Dai et al. | 2022 | Patients with chest pain who were referred for dynamic CT-MPI and CCTA | NR | 180/229 | Effect of calcium burden and morphology, and stenotic extent on diagnostic performance of FFR-CT |
|
| Koo et al. | 2021 | Patients with suspected or known CAD | 311.4 ± 527.4 | 471/557 | Effect of coronary artery calcification on diagnostic performance of FFR-CT |
|
| Tesche et al. | 2020 | Patients suspected of CAD who underwent CCTA | per-patient: 492 ± 644 (range 0 to 3920) per-vessel: 187 ± 280 (range 0 to 2,066) | 314/482 | Effect of coronary artery calcification on diagnostic performance of FFR-CT |
|
| Di Jiang et al. | 2020 | Patients with suspected CAD | per-patient: 288 ± 386 (range 0 to 1800) per-vessel: 138 ± 201 (range 0 to 1000) | 442/544 | Effect of coronary artery calcification on diagnostic performance of FFR-CT |
|
| Tang et al. | 2019 | Patients with suspected CAD | NR | 338/422 | Validate novel CFD-FFR-CT method |
|
| Xu et al. | 2019 | Patients with suspected or known CAD | per-vessel: 36.8 (range 0 to 191.6) | 437/570 | Effect of CT quality on diagnostic performance of FFR-CT |
|
| van Hamersvelt et al. | 2019 | Patients with suspected or known CAD | per-patient: 363.5 | 57/77 | Performance of a prototype on-site FFR-CT algorithm, based on patient-specific lumped parameter models |
|
| Kawaguchi et al. | 2019 | Patients with 30–90% stenosis of at least 1 major epicardial vessel of 2-mm or larger | per-patient: = 424.7 ± 554.8 | 66/81 | Predictive factors of false-positive FFR-CT findings |
|
| Nørgaard et al. | 2015 | Patients suspected of CAD who underwent CCTA | per-patient: 302 ± 468 (range 0 to 3599) per-vessel: 95 ± 172 (range 0 to 1703) | 214/333 | Effect of coronary artery calcification on diagnostic performance of FFR-CT |
|
| Min et al. | 2012 | Patients with suspected or known CAD who underwent CCTA | NR | 42/66 | Effect of CT quality on diagnostic performance of FFR-CT |
|
AS = Agatston Score; AUC = area under the curve; CAD = coronary artery disease; CCTA = coronary computed tomography angiography; CFD = computational fluid dynamics; CT-MPI = computed tomography myocardial perfusion imaging; FFR = fractional flow reserve; FFR-CT = fractional flow reserve computed tomography; IQR = interquartile range; NPV = negative predictive value.
AS scores are presented as means ± standard deviations or medians.
In the substudy of the NXT Trial (Analysis of Coronary Blood Flow Using CT Angiography: Next Steps), 214 patients were divided based on their Agatston score in 4 quartiles and the diagnostic performance and ability to diagnosed ischemia of FFR-CT compared with CCTA was investigated. First, there was no statistical difference in diagnostic accuracy, sensitivity, or specificity of FFR-CT across Agatston Score (AS) groups. Additionally, discrimination of ischemia by FFR-CT in the per-patient analysis was excellent, regardless of the presence of severe or not CAC (cutoff of 416 AS) as proven by the absence of significant difference in area under the receiver-operating characteristic curve (AUC) (0.86 vs 0.92, p = 0.45).
Similarly, in the MACHINE (MACHINE Learning based CT Angiography derived FFR: a multicenter Registry) registry, no significantly different sensitivity, specificity and accuracy of FFR-CT to detect lesion-specific ischemia was found across AS categories in the per-vessel and the per-patient analysis. However, and contrary to NXT findings, discriminatory power of FFR-CT was found to be significantly higher in vessels not in the severe calcium score group (CACS > 0 to 400) compared to vessels with severe CAC (CACS ≥ 400) with AUC: 0.85 [95% CI: 0.82 to 0.89] vs 0.71 [95% CI: 0.57 to 0.85)], (p = 0.04), respectively. The authors note that, compared with the NXT trial, the burden of vessel calcification was higher in the MACHINE registry. Specifically, the “high” AS group in the per-vessel analysis of NXT trial had a median AS of 248, whereas the corresponding highest AS group in the MACHINE registry had a median AS of 661. Moreover, possible selection bias was introduced in the NXT trial, where 12% of patients were ineligible for FFR-CT calculation, nearly half of them due to artifacts related to calcium blooming—an issue not observed in the MACHINE study. Of note, the 2 studies used different approaches for FFR-CT measurement, with computational fluid dynamics modeling in the NXT trial and machine-learning–based FFR-CT in the MACHINE. Previous data show equal performance of these 2 technical approaches. ,
Another study assessing diagnostic performance of FFR-CT in calcified coronary arteries was performed by Tao et al. This Chinese multicenter study analyzed data from 128 patients with suspected CAD. At the per-patient level, there was no significant difference in the sensitivity, specificity, accuracy, PPV, and NPV of FFR-CT for detecting lesion-specific ischemia between the low-to-intermediate CACS group and the high CACS group. Likewise, Di Jiang et al. reported no significantly different accuracy, sensitivity and specificity across AS categories in per-vessel and per-patient analyses . Finally, in a subset of 42 patients from the Diagnosis of ISChemia-Causing Stenoses Obtained Via NoninvasivE FRactional FLOW Reserve (DISCOVER-FLOW) study, FFR-CT diagnostic accuracy did not decline with increasing Agatston score, and overall performance was not found to be influenced by calcium artifacts. Of note, in patients with severe coronary calcification (Agatston ≥400), a pilot study using a subtraction CCTA technique showed that FFR-CT computed on the subtracted dataset significantly improved diagnostic performance versus standard FFR-CT when invasive FFR was the reference (AUC 0.84 vs 0.70; specificity 75% vs 59%) while maintaining high sensitivity and NPV. Despite the discrepancy between different studies regarding the actual CACS cutoff that may affect FFR-CT results, the morphology of calcification appeared to significantly influence the diagnostic performance of FFR-CT. Decreased accuracy has been reported in lesions with calcium ring >180° compared to lesions with less than half-circle calcification. A greater calcification arc is associated with false positive FFR-CT results, regardless of patient- or vessel-based CACS. Thus, the assessment of calcium morphology could be valuable in the diagnostic process.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
