Skip to main content
NCBI home page
Radiology: Cardiothoracic Imaging logoLink to Radiology: Cardiothoracic Imaging
. 2020 Jun 25;2(3):e190137. doi: 10.1148/ryct.2020190137

Value of Machine Learning–based Coronary CT Fractional Flow Reserve Applied to Triple-Rule-Out CT Angiography in Acute Chest Pain

Simon S Martin 1, Domenico Mastrodicasa 1, Marly van Assen 1, Carlo N De Cecco 1, Richard R Bayer 2nd 1, Christian Tesche 1, Akos Varga-Szemes 1, Andreas M Fischer 1, Brian E Jacobs 1, Pooyan Sahbaee 1, L Parkwood Griffith 1, Andrew J Matuskowitz 1, Thomas J Vogl 1, U Joseph Schoepf 1,✉
PMCID: PMC7978005  PMID: 33778579

Abstract

Purpose

To evaluate the additional value of noninvasive artificial intelligence (AI)–based CT-derived fractional flow reserve (CT FFR), derived from triple-rule-out coronary CT angiography for acute chest pain (ACP) in the emergency department (ED) setting.

Materials and Methods

AI-based CT FFR from triple-rule-out CT angiography data sets was retrospectively obtained in 159 of 271 eligible patients (102 men; mean age, 57.0 years ± 9.7 [standard deviation]) presenting to the ED with ACP. The agreement between CT FFR (≤ 0.80) and stenosis at triple-rule-out CT angiography (≥ 50%), as well as downstream cardiac diagnostic testing, was investigated. Furthermore, the predictive value of CT FFR for coronary revascularization and major adverse cardiac events (MACE) was assessed over a 1-year follow-up period.

Results

CT FFR and triple-rule-out CT angiography demonstrated agreement in severity of coronary artery disease (CAD) in 52% (82 of 159) of all cases. CT FFR of 0.80 and less served as a better predictor for coronary revascularization and MACE than stenosis of 50% and greater at triple-rule-out CT angiography (odds ratio, 3.4; 95% confidence interval: 1.4, 8.2 vs odds ratio, 2.2; 95% confidence interval: 0.9, 5.3) (P < .01). In the subgroup of patients with additional noninvasive cardiac testing (94 of 159), there was higher agreement as to the presence or absence of significant disease with CT FFR (55%) than with coronary triple-rule-out CT angiography (47%) (P = .23).

Conclusion

CT FFR derived from triple-rule-out CT angiography was a better predictor for coronary revascularization and MACE and showed better agreement with additional diagnostic testing than triple-rule-out CT angiography. Therefore, CT FFR may improve the specificity in identifying patients with ACP with significant CAD in the ED setting and reduce unnecessary downstream testing.

© RSNA, 2020

See also the commentary by Ihdayhid and Ben Zekry in this issue.

Summary

CT-based fractional flow reserve can improve the efficiency of patients assigned to triple-rule-out CT angiography in the emergency department.

Key Points

  • ■ Calculation of CT fractional flow reserve from triple-rule-out CT angiography was a better predictor for coronary revascularization and major adverse cardiac events than stenosis at triple-rule-out CT angiography (odds ratios, 3.4 vs 2.2).

  • ■ CT fractional flow reserve showed a better agreement with additional diagnostic testing than stenosis at triple-rule-out CT angiography (55% vs 47%, P = .23).

  • ■ Calculation of CT fractional flow reserve is feasible in the emergency setting and may improve the specificity in identifying patients with acute chest pain and significant coronary artery disease.

Introduction

Acute chest pain (ACP) is one of the most common reasons for presentation to the emergency department (ED). However, the differential diagnosis of ACP and the need to exclude significant coronary artery disease (CAD) as the cause of symptoms creates a complex scenario for the ED physician (1). Triple-rule-out CT angiography allows for simultaneous assessment of the coronary arteries, aorta, and pulmonary arteries in a single diagnostic test. This approach is most appropriate for patients with a low to intermediate risk of acute coronary syndrome and symptoms that may also be attributed to pulmonary embolism or aortic dissection (2,3). Since its early description (4), several studies have demonstrated good diagnostic accuracy of triple-rule-out CT angiography for evaluation of CAD (5–7). In general, the use of cardiac CT in the setting of ACP has been shown to enhance the effectiveness of evaluation in the ED (8–10), although some researchers have argued that the rate of downstream testing and overall radiation exposure may increase (11).

Advances in computational fluid dynamics and patient-specific three-dimensional image modeling allow for the noninvasive calculation of fractional flow reserve (FFR) derived from coronary CT angiography (CT FFR) data sets (12–14). CT FFR based on these physical principles has shown high diagnostic accuracy in the assessment of hemodynamic significance of individual coronary lesions and improved patient outcomes when basing further treatment on functional as opposed to anatomic data (13,15–18). Other approaches for CT FFR derivation use investigational artificial intelligence (AI) deep machine learning algorithm software (19), which has previously demonstrated good accuracy as well as potential time efficiency for patient management in an elective setting (19–22). This AI-based algorithm uses learned relationships to compute the functional severity of a lesion in a patient’s coronary CT angiographic study (19).

Overall, although evidence supports the application of CT FFR to guide patient management with regard to chronic stable angina (17), the use of CT FFR in a nonelective ACP setting, centered in the ED, is insufficiently studied. Therefore, the aim of this investigation was to evaluate AI-based CT FFR derived from triple-rule-out CT angiography for ACP in the ED setting.

Materials and Methods

Patient Selection

This retrospective single-center study was approved by the local institutional review board with a waiver for informed consent. Institutional databases were reviewed to identify patients who had undergone dedicated triple-rule-out CT angiographic examinations in the ACP-Center of our university hospital between March 2009 and April 2016. Patients identified as having at least one coronary stenosis, visually graded between 30% and 90% and thus appropriate for further evaluation with CT FFR were included. Patients with insufficient CT image quality for CT FFR analysis (n = 95), prior coronary stent placement (n = 8), or history of coronary artery bypass grafting (CABG) (n = 1) were excluded. Criteria for insufficient image quality were severe motion artifacts, severe image noise, or poor contrast enhancement that precluded the evaluation of the lumen (23). Another eight patients were lost to follow-up over the course of 1 year. A flowchart of the study enrollment is shown in Figure 1.

Figure 1:

Study flowchart. CABG = coronary artery bypass grafting, CAD = coronary artery disease, CT FFR = fractional flow reserve derived from coronary CT angiography.

Open in a new tab

Study flowchart. CABG = coronary artery bypass grafting, CAD = coronary artery disease, CT FFR = fractional flow reserve derived from coronary CT angiography.

Triple-Rule-Out Coronary CT Angiography

Triple-rule-out CT angiographic examinations were performed with either a second- or third-generation dual-source CT scanner (SOMATOM Definition Flash or SOMATOM Force; Siemens Healthineers, Forchheim, Germany). Prospective electrocardiographic triggering was used in patients with a regular rhythm, and retrospective electrocardiographic gating was used in patients with tachyarrhythmia. Sublingual nitroglycerin and β-blockers were used at the discretion of the supervising attending physician. Because of the different circumstances in the ED, only three patients received sublingual nitroglycerin and 13 patients received β-blockers prior to the CT examination.

An automated tube voltage selection algorithm was used (CARE kV; Siemens). The following parameters were used: collimation, 0.6 mm; tube voltage, 80–120 kV (second-generation dual-source CT) and 70–130 kV (third-generation dual-source CT); tube current–time product, 350–650 mAs (second-generation dual-source CT) and 200–650 mAs (third-generation dual-source CT); section thickness, 1 mm in a 0.5-mm increment. A triphasic contrast material injection protocol was used to administer 60 mL of contrast medium (iopromide, Ultravist; 370 mg of iodine per milliliter, Bayer, Whippany, NJ) at 5 mL/sec, followed by 15 mL of contrast material diluted with 35 mL of saline and a 50-mL saline bolus. The scan was initiated with a bolus-tracking technique 4 seconds after the bolus reached a threshold of 100 HU within a region of interest placed in the descending aorta. All images were acquired during inspiratory breath hold. The dose–length product and the effective dose (conversion coefficient of k = 0.017 mSv ∙ mGy-1 ∙ cm-1) were calculated for an estimation of the effective radiation dose equivalent (24).

The coronary CT angiography data sets derived from the triple-rule-out scan were postprocessed and analyzed on a dedicated workstation (Syngo.via, version VB20A; Siemens). A board-certified, subspecialty-trained cardiovascular imager (C.N.D.C.) with more than 5 years of experience interpreted all coronary CT angiographic scans. Coronary stenoses were graded as mild (<50%), moderate (50%–69%), or severe (≥70% or left main coronary artery >50%). Obstructive CAD was defined as stenosis of 50% or greater (25).

Computation of CT FFR from Coronary CT Angiography

The CT FFR analysis was performed using a previously described dedicated AI-based software prototype that is currently not commercially available (cFFR, version 3.0; Siemens Healthineers) (19,20). Compared with the standard CT FFR algorithm that is based on computational fluid dynamics, this prototype uses machine learning–based models for the assessment of the functional severity of a lesion (19). Therefore, the software automatically creates an initial model of the coronary anatomy with corresponding centerlines and luminal contours. After correction and acceptance of the coronary tree geometry by the user, the software uses a multilayer neural network architecture to compute the functional severity of the lesion based on the different features extracted from anatomic analysis and their corresponding hemodynamic conditions (26). This semiautomatic approach reduces the need for human input and increases the speed of CT FFR computations (19). All data sets were postprocessed by an observer (D.M.) with 4 years of experience in cardiovascular imaging. The spatial distribution of the CT FFR values were projected as a color-coded map onto the three-dimensional coronary model. A CT FFR measurement of 0.80 or less was used to indicate hemodynamically significant stenosis. The physician performing the CT FFR analysis (D.M.) was blinded to the results of the triple-rule-out CT angiography and clinical outcomes. The results of the CT FFR analysis were obtained post hoc, not available to care providers, and therefore were not used for clinical management.

Coronary Revascularization and MACE

Patients fulfilling the inclusion criteria were followed up for a period of 12 months using the electronic medical record system of our university hospital, data obtained from the United States Social Security Death Index, and patient phone calls. The primary endpoint was coronary revascularization, not driven by the index triple-rule-out CT angiography study, including percutaneous coronary intervention and CABG, or major adverse cardiovascular events (MACE), including death, nonfatal myocardial infarction, and hospitalization for unstable angina. Two blinded radiologists adjudicated all events (S.S.M., D.M.).

Additional Diagnostic Testing

We retrospectively evaluated patients for subsequent cardiac testing, including invasive coronary angiography (ICA), SPECT, stress echocardiography, and stress MRI. SPECT, stress echocardiography, and stress MRI were performed in accordance with society guidelines (27–29). FFR results derived from ICA were only available in 12 patients and therefore not included in further analysis.

In addition, we investigated the potential improvement in efficiency of referral to additional diagnostic testing by adding results of the CT FFR analysis. Therefore, we defined a CT FFR–guided decision rule that reserved downstream testing for patients with CT FFR of 0.80 and less (16). The projected numbers of study patients undergoing additional diagnostic testing, if that decision rule would have been followed, were described and compared with what was observed in terms of actual management.

Statistical Analysis

Absolute variables were reported as totals and percentages, and continuous variables were reported as means with standard deviations. Normal distribution was evaluated using the Kolmogorov-Smirnov test. Comparisons were performed with the Student t test for continuous variables and the χ2 test for categorical variables. Binary logistic regression analyses with odds ratios were conducted to evaluate the utility of a positive result at CT FFR (≤ 0.80) and a positive result at CT angiography (≥50% stenosis) as predictors of the occurrence of MACE or unplanned coronary revascularization within 1 year. To reduce the number of competing risk factors (age, body mass index, sex, race, diabetes mellitus, hypertension, dyslipidemia, smoking, family history of CAD), the least absolute shrinkage and selection operator was applied. A Fisher exact test was performed to compare patients with positive and negative CT FFR results in terms of the percentage manifesting the primary outcome. A McNemar test was performed to compare the agreement as to the presence or absence of significant disease with CT FFR and coronary triple-rule-out CT angiography. Statistical significance was assumed at a P value of less than .05. All statistical analyses were performed using MedCalc Statistical Software version 18 (MedCalc Software, Ostend, Belgium).

Results

Study Population

After exclusion criteria were applied, the study population consisted of 159 patients who had complete information from triple-rule-out coronary CT angiography and CT FFR. The average age was 57.0 years ± 9.7 (standard deviation), 64% were men (102 of 159), and 36% were women (57 of 159). Patient characteristics are summarized in Table 1.

Table 1:

Patient Characteristics and Cardiac Risk Factors

graphic file with name ryct.2020190137.tbl1.jpg

Open in a new tab

Eighty-seven patients (55%) had significant stenosis (≥50%) and eight patients (5%) showed severe stenosis (≥70% or left main coronary artery >50%) at triple-rule-out CT angiography. Mean per-patient CT FFR was 0.72 ± 0.21 with 50% of patients (80 of 159) having a CT FFR of 0.80 or less in a least one vessel. A subgroup of 18 patients (11%) had CT FFR values in the “gray zone” of CT FFR (ie, between 0.75 and 0.80). Overall, 26% (41 of 159) underwent ICA following triple-rule-out CT angiography and 11% (18 of 159) underwent revascularization (17 percutaneous coronary interventions and one CABG). During the 12-month follow-up, MACE occurred in 8% of patients (13 of 159), including one death, one nonfatal myocardial infarction, and 11 hospitalizations for unstable angina (Table 1). There was agreement between CT FFR and triple-rule-out CT angiography in 52% of patients (82 of 159). Among these, 55% (45 of 82) had CT FFR of 0.80 or less and a corresponding stenosis of 50% or greater at CT angiography, while 45% (37 of 82) had a negative result at CT FFR and less than 50% stenosis at triple-rule-out CT angiography. The mean duration for the CT FFR analysis was 17.8 minutes ± 8.1 (range, 9–60 minutes) per examination (Table 2). The mean dose–length product and effective dose during the triple-rule-out CT examination were 1321.6 mGy ∙cm ± 674.3 and 22.2 mSv ± 11.7, respectively. The average scan range was 71.4 cm ± 10.2, and the average scan time was 6.7 minutes ± 2.6.

Table 2:

Stenosis, Revascularization, and MACE

graphic file with name ryct.2020190137.tbl2.jpg

Open in a new tab

Association of CT FFR with Coronary Revascularization and MACE

During the follow-up of 12 months, the primary outcome of a composite of MACE or non–triple-rule-out CT angiography result-driven coronary revascularization occurred in 27.5% of patients (22 of 80) with positive result at CT FFR (≤ 0.80) and 10.1% (eight of 79) with negative CT FFR results. Patients with a CT FFR of 0.80 or less were significantly more likely to undergo coronary revascularization or experience a MACE than patients with CT FFR values greater than 0.80 (P < .01). After performing least absolute shrinkage and selection operator analysis, no competing risk factors were found to be significant predictors of the primary outcome. However, a CT FFR of 0.80 or less was a better predictor for coronary revascularization or MACE than obstructive stenosis (odds ratio, 3.4; 95% confidence interval: 1.4, 8.2 vs odds ratio, 2.2; 95% confidence interval: 0.9, 5.3) and contributed significantly to the prediction of the outcome (P < .01). Figures 2 and 3 show representative patient examples.

Figure 2a:

Images in a 57-year-old man presenting to the emergency department with acute chest pain. Triple-rule-out CT angiograms show (a, b) a moderate calcified stenosis of the proximal left anterior descending artery (arrows) and (c, d) 90% stenosis of the midportion of the left circumflex artery (arrows) due to noncalcified atherosclerotic plaque. (e) CT FFR analysis reveals lesion-specific ischemia (≤ 0.80) of the left anterior descending artery, left diagonal, and left circumflex arteries. The patient underwent invasive coronary angiography with revascularization of the left circumflex artery. FFR = fractional flow reserve.

Open in a new tab

Images in a 57-year-old man presenting to the emergency department with acute chest pain. Triple-rule-out CT angiograms show (a, b) a moderate calcified stenosis of the proximal left anterior descending artery (arrows) and (c, d) 90% stenosis of the midportion of the left circumflex artery (arrows) due to noncalcified atherosclerotic plaque. (e) CT FFR analysis reveals lesion-specific ischemia (≤ 0.80) of the left anterior descending artery, left diagonal, and left circumflex arteries. The patient underwent invasive coronary angiography with revascularization of the left circumflex artery. FFR = fractional flow reserve.

Figure 3a:

Images in a 69-year-old woman who presented to the emergency department with acute chest pain and was referred for triple-rule-out CT angiography. Coronary CT angiograms show (a, b) extensive atherosclerotic calcification in the right coronary artery and (c, d) left anterior descending coronary artery with at least moderate stenosis in both the right coronary and left anterior descending arteries (arrows). (e) However, CT FFR reveals no hemodynamically significant stenosis with absence of lesion-specific ischemia. (f) Gated SPECT images demonstrate a fixed perfusion defect in the basal septum. However, it is classified as an artifact, based on raw data and attenuation correction. FFR = fractional flow reserve.

Open in a new tab

Images in a 69-year-old woman who presented to the emergency department with acute chest pain and was referred for triple-rule-out CT angiography. Coronary CT angiograms show (a, b) extensive atherosclerotic calcification in the right coronary artery and (c, d) left anterior descending coronary artery with at least moderate stenosis in both the right coronary and left anterior descending arteries (arrows). (e) However, CT FFR reveals no hemodynamically significant stenosis with absence of lesion-specific ischemia. (f) Gated SPECT images demonstrate a fixed perfusion defect in the basal septum. However, it is classified as an artifact, based on raw data and attenuation correction. FFR = fractional flow reserve.

Figure 2b:

Images in a 57-year-old man presenting to the emergency department with acute chest pain. Triple-rule-out CT angiograms show (a, b) a moderate calcified stenosis of the proximal left anterior descending artery (arrows) and (c, d) 90% stenosis of the midportion of the left circumflex artery (arrows) due to noncalcified atherosclerotic plaque. (e) CT FFR analysis reveals lesion-specific ischemia (≤ 0.80) of the left anterior descending artery, left diagonal, and left circumflex arteries. The patient underwent invasive coronary angiography with revascularization of the left circumflex artery. FFR = fractional flow reserve.

Open in a new tab

Images in a 57-year-old man presenting to the emergency department with acute chest pain. Triple-rule-out CT angiograms show (a, b) a moderate calcified stenosis of the proximal left anterior descending artery (arrows) and (c, d) 90% stenosis of the midportion of the left circumflex artery (arrows) due to noncalcified atherosclerotic plaque. (e) CT FFR analysis reveals lesion-specific ischemia (≤ 0.80) of the left anterior descending artery, left diagonal, and left circumflex arteries. The patient underwent invasive coronary angiography with revascularization of the left circumflex artery. FFR = fractional flow reserve.

Figure 2c:

Images in a 57-year-old man presenting to the emergency department with acute chest pain. Triple-rule-out CT angiograms show (a, b) a moderate calcified stenosis of the proximal left anterior descending artery (arrows) and (c, d) 90% stenosis of the midportion of the left circumflex artery (arrows) due to noncalcified atherosclerotic plaque. (e) CT FFR analysis reveals lesion-specific ischemia (≤ 0.80) of the left anterior descending artery, left diagonal, and left circumflex arteries. The patient underwent invasive coronary angiography with revascularization of the left circumflex artery. FFR = fractional flow reserve.

Open in a new tab

Images in a 57-year-old man presenting to the emergency department with acute chest pain. Triple-rule-out CT angiograms show (a, b) a moderate calcified stenosis of the proximal left anterior descending artery (arrows) and (c, d) 90% stenosis of the midportion of the left circumflex artery (arrows) due to noncalcified atherosclerotic plaque. (e) CT FFR analysis reveals lesion-specific ischemia (≤ 0.80) of the left anterior descending artery, left diagonal, and left circumflex arteries. The patient underwent invasive coronary angiography with revascularization of the left circumflex artery. FFR = fractional flow reserve.

Figure 2d:

Images in a 57-year-old man presenting to the emergency department with acute chest pain. Triple-rule-out CT angiograms show (a, b) a moderate calcified stenosis of the proximal left anterior descending artery (arrows) and (c, d) 90% stenosis of the midportion of the left circumflex artery (arrows) due to noncalcified atherosclerotic plaque. (e) CT FFR analysis reveals lesion-specific ischemia (≤ 0.80) of the left anterior descending artery, left diagonal, and left circumflex arteries. The patient underwent invasive coronary angiography with revascularization of the left circumflex artery. FFR = fractional flow reserve.

Open in a new tab

Images in a 57-year-old man presenting to the emergency department with acute chest pain. Triple-rule-out CT angiograms show (a, b) a moderate calcified stenosis of the proximal left anterior descending artery (arrows) and (c, d) 90% stenosis of the midportion of the left circumflex artery (arrows) due to noncalcified atherosclerotic plaque. (e) CT FFR analysis reveals lesion-specific ischemia (≤ 0.80) of the left anterior descending artery, left diagonal, and left circumflex arteries. The patient underwent invasive coronary angiography with revascularization of the left circumflex artery. FFR = fractional flow reserve.

Figure 2e:

Images in a 57-year-old man presenting to the emergency department with acute chest pain. Triple-rule-out CT angiograms show (a, b) a moderate calcified stenosis of the proximal left anterior descending artery (arrows) and (c, d) 90% stenosis of the midportion of the left circumflex artery (arrows) due to noncalcified atherosclerotic plaque. (e) CT FFR analysis reveals lesion-specific ischemia (≤ 0.80) of the left anterior descending artery, left diagonal, and left circumflex arteries. The patient underwent invasive coronary angiography with revascularization of the left circumflex artery. FFR = fractional flow reserve.

Open in a new tab

Images in a 57-year-old man presenting to the emergency department with acute chest pain. Triple-rule-out CT angiograms show (a, b) a moderate calcified stenosis of the proximal left anterior descending artery (arrows) and (c, d) 90% stenosis of the midportion of the left circumflex artery (arrows) due to noncalcified atherosclerotic plaque. (e) CT FFR analysis reveals lesion-specific ischemia (≤ 0.80) of the left anterior descending artery, left diagonal, and left circumflex arteries. The patient underwent invasive coronary angiography with revascularization of the left circumflex artery. FFR = fractional flow reserve.

Figure 3b:

Images in a 69-year-old woman who presented to the emergency department with acute chest pain and was referred for triple-rule-out CT angiography. Coronary CT angiograms show (a, b) extensive atherosclerotic calcification in the right coronary artery and (c, d) left anterior descending coronary artery with at least moderate stenosis in both the right coronary and left anterior descending arteries (arrows). (e) However, CT FFR reveals no hemodynamically significant stenosis with absence of lesion-specific ischemia. (f) Gated SPECT images demonstrate a fixed perfusion defect in the basal septum. However, it is classified as an artifact, based on raw data and attenuation correction. FFR = fractional flow reserve.

Open in a new tab

Images in a 69-year-old woman who presented to the emergency department with acute chest pain and was referred for triple-rule-out CT angiography. Coronary CT angiograms show (a, b) extensive atherosclerotic calcification in the right coronary artery and (c, d) left anterior descending coronary artery with at least moderate stenosis in both the right coronary and left anterior descending arteries (arrows). (e) However, CT FFR reveals no hemodynamically significant stenosis with absence of lesion-specific ischemia. (f) Gated SPECT images demonstrate a fixed perfusion defect in the basal septum. However, it is classified as an artifact, based on raw data and attenuation correction. FFR = fractional flow reserve.

Figure 3c:

Images in a 69-year-old woman who presented to the emergency department with acute chest pain and was referred for triple-rule-out CT angiography. Coronary CT angiograms show (a, b) extensive atherosclerotic calcification in the right coronary artery and (c, d) left anterior descending coronary artery with at least moderate stenosis in both the right coronary and left anterior descending arteries (arrows). (e) However, CT FFR reveals no hemodynamically significant stenosis with absence of lesion-specific ischemia. (f) Gated SPECT images demonstrate a fixed perfusion defect in the basal septum. However, it is classified as an artifact, based on raw data and attenuation correction. FFR = fractional flow reserve.

Open in a new tab

Images in a 69-year-old woman who presented to the emergency department with acute chest pain and was referred for triple-rule-out CT angiography. Coronary CT angiograms show (a, b) extensive atherosclerotic calcification in the right coronary artery and (c, d) left anterior descending coronary artery with at least moderate stenosis in both the right coronary and left anterior descending arteries (arrows). (e) However, CT FFR reveals no hemodynamically significant stenosis with absence of lesion-specific ischemia. (f) Gated SPECT images demonstrate a fixed perfusion defect in the basal septum. However, it is classified as an artifact, based on raw data and attenuation correction. FFR = fractional flow reserve.

Figure 3d:

Images in a 69-year-old woman who presented to the emergency department with acute chest pain and was referred for triple-rule-out CT angiography. Coronary CT angiograms show (a, b) extensive atherosclerotic calcification in the right coronary artery and (c, d) left anterior descending coronary artery with at least moderate stenosis in both the right coronary and left anterior descending arteries (arrows). (e) However, CT FFR reveals no hemodynamically significant stenosis with absence of lesion-specific ischemia. (f) Gated SPECT images demonstrate a fixed perfusion defect in the basal septum. However, it is classified as an artifact, based on raw data and attenuation correction. FFR = fractional flow reserve.

Open in a new tab

Images in a 69-year-old woman who presented to the emergency department with acute chest pain and was referred for triple-rule-out CT angiography. Coronary CT angiograms show (a, b) extensive atherosclerotic calcification in the right coronary artery and (c, d) left anterior descending coronary artery with at least moderate stenosis in both the right coronary and left anterior descending arteries (arrows). (e) However, CT FFR reveals no hemodynamically significant stenosis with absence of lesion-specific ischemia. (f) Gated SPECT images demonstrate a fixed perfusion defect in the basal septum. However, it is classified as an artifact, based on raw data and attenuation correction. FFR = fractional flow reserve.

Figure 3e:

Images in a 69-year-old woman who presented to the emergency department with acute chest pain and was referred for triple-rule-out CT angiography. Coronary CT angiograms show (a, b) extensive atherosclerotic calcification in the right coronary artery and (c, d) left anterior descending coronary artery with at least moderate stenosis in both the right coronary and left anterior descending arteries (arrows). (e) However, CT FFR reveals no hemodynamically significant stenosis with absence of lesion-specific ischemia. (f) Gated SPECT images demonstrate a fixed perfusion defect in the basal septum. However, it is classified as an artifact, based on raw data and attenuation correction. FFR = fractional flow reserve.

Open in a new tab

Images in a 69-year-old woman who presented to the emergency department with acute chest pain and was referred for triple-rule-out CT angiography. Coronary CT angiograms show (a, b) extensive atherosclerotic calcification in the right coronary artery and (c, d) left anterior descending coronary artery with at least moderate stenosis in both the right coronary and left anterior descending arteries (arrows). (e) However, CT FFR reveals no hemodynamically significant stenosis with absence of lesion-specific ischemia. (f) Gated SPECT images demonstrate a fixed perfusion defect in the basal septum. However, it is classified as an artifact, based on raw data and attenuation correction. FFR = fractional flow reserve.

Figure 3f:

Images in a 69-year-old woman who presented to the emergency department with acute chest pain and was referred for triple-rule-out CT angiography. Coronary CT angiograms show (a, b) extensive atherosclerotic calcification in the right coronary artery and (c, d) left anterior descending coronary artery with at least moderate stenosis in both the right coronary and left anterior descending arteries (arrows). (e) However, CT FFR reveals no hemodynamically significant stenosis with absence of lesion-specific ischemia. (f) Gated SPECT images demonstrate a fixed perfusion defect in the basal septum. However, it is classified as an artifact, based on raw data and attenuation correction. FFR = fractional flow reserve.

Open in a new tab

Images in a 69-year-old woman who presented to the emergency department with acute chest pain and was referred for triple-rule-out CT angiography. Coronary CT angiograms show (a, b) extensive atherosclerotic calcification in the right coronary artery and (c, d) left anterior descending coronary artery with at least moderate stenosis in both the right coronary and left anterior descending arteries (arrows). (e) However, CT FFR reveals no hemodynamically significant stenosis with absence of lesion-specific ischemia. (f) Gated SPECT images demonstrate a fixed perfusion defect in the basal septum. However, it is classified as an artifact, based on raw data and attenuation correction. FFR = fractional flow reserve.

Association of CT FFR with Additional Diagnostic Testing

Overall, 67% of patients (106 of 159) underwent a total of 135 cardiac downstream tests, including 26 patients (16%) with more than one additional examination. ICA was performed in 41 patients, of which 54% (22 of 41) had a CT FFR of 0.80 or less. In the subgroup of patients with additional SPECT imaging (n = 62), the percentage of patients with a CT FFR of 0.80 or less and signs of ischemia as determined with SPECT was 5% (three of 62); furthermore, the percentage with a CT FFR of greater than 0.80 and no signs of ischemia was 47% (29 of 62). In 48% of patients (30 of 62), the results of CT FFR and SPECT were discordant. Stress echocardiography was performed in 19% of patients (31 of 159), demonstrating concordant results compared with CT FFR in 61% of all patients (19 of 31). Triple-rule-out CT angiography evaluation was discordant in 54% of patients (34 of 62) compared with SPECT, and in 48% of patients (15 of 31) compared with stress echocardiography. One patient underwent downstream stress MRI with no signs of ischemia at perfusion imaging nor CT FFR (> 0.80) analysis (Table 3). In the subgroup of patients with additional noninvasive cardiac testing (84 patients undergoing 94 cardiac tests), there was higher agreement as to the presence or absence of significant disease with CT FFR (55%; 52 of 94) than with coronary triple-rule-out CT angiography (47%; 44 of 94) (P = .23).

Table 3:

Additional Diagnostic Testing

graphic file with name ryct.2020190137.tbl3.jpg

Open in a new tab

Reserving downstream testing for patients with a CT FFR of 0.80 or less would have reduced the number of subsequent cardiac examinations (n = 135) after triple-rule-out CT angiography by 47% (64 of 135), including 19 patients who were referred for ICA. Conversely, in this scenario (ie, following a CT FFR–driven patient management decision rule), four patients with CT FFR values of 0.84, 0.96, 0.89, and 0.85 would not have undergone coronary revascularization, and three patients with a CT FFR of 0.80 or greater showed a positive result at SPECT. The impact from this exclusion cannot be assessed in this study. Notably, none of these patients experienced a MACE during the 12 months of follow-up and had mild (n = 2) to moderate (n = 2) stenosis at ICA.

Discussion

The results of our study demonstrated that the addition of an AI-based CT FFR prototype to triple-rule-out CT angiographic examinations improved the specificity for obstructive CAD in patients presenting to the ED with ACP. A CT FFR threshold of 0.80 or less yielded a significantly better predictive value for patient outcome than stenosis of 50% or greater at triple-rule-out CT angiography during the follow-up period of 1 year. Furthermore, the agreement of CT FFR with the results of additional functional testing was higher than triple-rule-out CT angiography alone, although the differences did not reach statistical significance. Thus, CT FFR may improve the efficiency and cost-effectiveness of managing treatment of patients with ACP in the ED by lowering the number of additional downstream tests.

Many EDs have incorporated coronary CT angiography into an early evaluation strategy for patients with ACP concerning for a cardiac origin (10,30). This strategy has shown to improve the efficiency of clinical decision making in comparison with standard evaluation in the ED by reducing the time to diagnosis and the average length of stay (11). However, the evaluation of chest pain remains a challenging task in the ED. Triple-rule-out CT angiography is most appropriate for patients in the category of low-to-moderate risk for acute coronary syndrome with symptoms that could also be concerning for other acute pathologic conditions of the aorta or pulmonary arteries (2,31). Moreover, with contemporary imaging equipment, triple-rule-out CT angiography can provide comparable coronary artery visualization and image quality compared with standard, dedicated coronary CT angiography (5,31).

Several multicenter trials have argued that the use of coronary CT angiography increases the number of subsequent invasive catheterizations in patients with symptoms suggestive of CAD compared with alternative testing (11,32). A recent study by Lu et al revealed that the addition of CT FFR to coronary CT angiography may reduce the referral to invasive catheterization by up to 28% in a nonemergent elective setting (16). Furthermore, it was demonstrated that the addition of CT FFR to coronary CT angiography could increase the percentage of invasive catheterizations leading to revascularization from 49% to 61% (16), enhancing the effectiveness of catheterization laboratory resource use. In line with these observations, our study results revealed that CT FFR might reduce the number of subsequent downstream tests and therefore improve the efficiency of patient management, also in an ACP ED setting.

In alignment with previous studies, we found a substantial disagreement between CT FFR and triple-rule-out CT angiography evaluation (16,33). In addition, more patients with a positive CT FFR analysis underwent coronary revascularization or experienced MACE in comparison with patients with obstructive stenosis at triple-rule-out CT angiography alone. This is notable because results of the triple-rule-out CT angiography guided clinical decision making, whereas CT FFR was not available to the treating physician at the time of patient disposition in this retrospective study. Moreover, our data showed that coronary revascularization was more likely to be performed in patients with hemodynamically significant lesions as determined with CT FFR.

Although CT FFR analysis achieved a higher agreement with additional diagnostic testing compared with triple-rule-out CT angiography, a high number of patients with positive CT FFR findings did not show any evidence of ischemia at SPECT or stress echocardiography. Specifically, 43% of the negative SPECT and 35% of the negative stress echocardiography studies demonstrated a positive result at CT FFR. This was remarkable, considering prior studies have demonstrated comparably high diagnostic accuracies of both SPECT and CT FFR in assessing hemodynamically significant stenoses (34). In general, SPECT examinations are susceptible to artifacts, which may result in false-positive findings. Conversely, the limited spatial resolution may result in the underdetection of small subendocardial ischemic lesions, especially in obese patients. In addition, the dependence on the operator for echocardiography image acquisition may result in more subjective interpretation, making accurate, objective assessment of ischemia more difficult in some cases.

Of note, we obtained a mean duration of 17.8 minutes for the CT FFR evaluation, and these results were in line with previous studies focused on AI-based CT FFR computation (35). These short analysis times may be of particular suitability and usefulness in the emergency setting because it allows for faster patient disposition times. Moreover, CT FFR calculation could also be performed by an experienced CT technologist to relieve the physician in the ED.

The results of our study should be considered in the context of the retrospective design and its inherent limitations. This was also a single-center study in a somewhat limited population, which may prevent results from being more widely applicable. Only patients with coronary artery stenosis between 30% and 90% on the triple-rule-out CT angiography scan were included, which may result in selection bias. Moreover, the results of CT FFR were not available to the treating physicians in real time and as such, all clinical decision making was performed in a post hoc “virtual” scenario. Eight patients with a CT FFR of greater than 0.80 had coronary revascularization or experienced MACE, and we cannot assess the consequences for these patients if they had been deferred from this treatment. Further investigation is needed in larger prospective multicenter trials to determine the diagnostic accuracy and utility of this on-site technique. In addition, a large number of patients were excluded because the image quality was not adequate for calculation of CT FFR. This might be attributable to the more complex setting of patients with ACP in the ED. Finally, invasive FFR results were only available in a small portion of patients and were ultimately not included in the analysis. Although the results of invasive FFR and CT FFR showed good agreement, the number of patients was too small to perform a statistical analysis.

In conclusion, our data suggested that AI-based CT FFR derived from triple-rule-out CT angiography data sets provided additional diagnostic and prognostic value in the evaluation of patients presenting to the ED with chest pain and adding CT FFR in the ED may reduce subsequent downstream testing.

The concepts and information presented are based on research and are not commercially available.

Disclosures of Conflicts of Interest: S.S.M. disclosed no relevant relationships. D.M. disclosed no relevant relationships. M.v.A. disclosed no relevant relationships. C.N.D.C. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: institution receives research grant from Siemens; author receives royalties from Springer; author has received honoraria for consulting and speaking from Guerbet and Siemens. Other relationships: disclosed no relevant relationships. R.R.B. Activities related to the present article: institution receives research support from Siemens Healthineers. Activities not related to the present article: author receives reimbursement for travel related to ongoing research initiatives from Siemens. Other relationships: disclosed no relevant relationships. C.T. disclosed no relevant relationships. A.V.S. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: author is consultant for Elucid BioImaging; institution receives grant from Siemens Healthcare; author has stock/stock options in Elucid BioImaging; author receives travel accommodations from Siemens Healthcare; support from Guerbet. Other relationships: disclosed no relevant relationships. A.M.F. disclosed no relevant relationships. B.E.J. disclosed no relevant relationships. P.S. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: Siemens employee. Other relationships: research collaboration. L.P.G. disclosed no relevant relationships. A.J.M. disclosed no relevant relationships. T.J.V. disclosed no relevant relationships. U.J.S. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: institutional research support from Astellas, Bayer, Bracco, and Siemens and has received honoraria for consulting and speaking from Bayer, Elucid BioImaging, GE, Guerbet, HeartFlow, and Siemens. Other relationships: disclosed no relevant relationships.

Abbreviations:

ACP
acute chest pain
AI
artificial intelligence
CABG
coronary artery bypass grafting
CAD
coronary artery disease
CT FFR
CT-derived FFR
ED
emergency department
FFR
fractional flow reserve
ICA
invasive coronary angiography
MACE
major adverse cardiovascular event

References

  • 1.Mahler SA, Riley RF, Hiestand BC, et al. The HEART Pathway randomized trial: identifying emergency department patients with acute chest pain for early discharge. Circ Cardiovasc Qual Outcomes 2015;8(2):195–203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Halpern EJ. Triple-rule-out CT angiography for evaluation of acute chest pain and possible acute coronary syndrome. Radiology 2009;252(2):332–345. [DOI] [PubMed] [Google Scholar]
  • 3.Burris AC 2nd, Boura JA, Raff GL, Chinnaiyan KM. Triple rule out versus coronary CT angiography in patients with acute chest pain: results from the ACIC consortium. JACC Cardiovasc Imaging 2015;8(7):817–825. [DOI] [PubMed] [Google Scholar]
  • 4.Savino G, Herzog C, Costello P, Schoepf UJ. 64 slice cardiovascular CT in the emergency department: concepts and first experiences. Radiol Med (Torino) 2006;111(4):481–496. [DOI] [PubMed] [Google Scholar]
  • 5.Halpern EJ, Levin DC, Zhang S, Takakuwa KM. Comparison of image quality and arterial enhancement with a dedicated coronary CTA protocol versus a triple rule-out coronary CTA protocol. Acad Radiol 2009;16(9):1039–1048. [DOI] [PubMed] [Google Scholar]
  • 6.Madder RD, Raff GL, Hickman L, et al. Comparative diagnostic yield and 3-month outcomes of “triple rule-out” and standard protocol coronary CT angiography in the evaluation of acute chest pain. J Cardiovasc Comput Tomogr 2011;5(3):165–171. [DOI] [PubMed] [Google Scholar]
  • 7.Rogers IS, Banerji D, Siegel EL, et al. Usefulness of comprehensive cardiothoracic computed tomography in the evaluation of acute undifferentiated chest discomfort in the emergency department (CAPTURE). Am J Cardiol 2011;107(5):643–650. [DOI] [PubMed] [Google Scholar]
  • 8.Goldstein JA, Chinnaiyan KM, Abidov A, et al. The CT-STAT (Coronary computed tomographic angiography for systematic triage of acute chest pain patients to treatment) trial. J Am Coll Cardiol 2011;58(14):1414–1422. [DOI] [PubMed] [Google Scholar]
  • 9.Hoffmann U, Pena AJ, Cury RC, et al. Cardiac CT in emergency department patients with acute chest pain. RadioGraphics 2006;26(4):963–978; discussion 979–980. [DOI] [PubMed] [Google Scholar]
  • 10.Miller MM, Ridge CA, Litmanovich DE. Computed tomography angiographic assessment of acute chest pain. J Thorac Imaging 2017;32(3):137–150. [DOI] [PubMed] [Google Scholar]
  • 11.Hoffmann U, Truong QA, Schoenfeld DA, et al. Coronary CT angiography versus standard evaluation in acute chest pain. N Engl J Med 2012;367(4):299–308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Koo BK, Erglis A, Doh JH, et al. Diagnosis of ischemia-causing coronary stenoses by noninvasive fractional flow reserve computed from coronary computed tomographic angiograms. Results from the prospective multicenter DISCOVER-FLOW (Diagnosis of ischemia-causing stenoses obtained via noninvasive fractional flow reserve) study. J Am Coll Cardiol 2011;58(19):1989–1997. [DOI] [PubMed] [Google Scholar]
  • 13.Nørgaard BL, Leipsic J, Gaur S, et al. Diagnostic performance of noninvasive fractional flow reserve derived from coronary computed tomography angiography in suspected coronary artery disease: the NXT trial (Analysis of coronary blood flow using CT angiography: next steps). J Am Coll Cardiol 2014;63(12):1145–1155. [DOI] [PubMed] [Google Scholar]
  • 14.Benton SM Jr, Tesche C, De Cecco CN, Duguay TM, Schoepf UJ, Bayer RR 2nd. Noninvasive derivation of fractional flow reserve from coronary computed tomographic angiography: a review. J Thorac Imaging 2018;33(2):88–96. [DOI] [PubMed] [Google Scholar]
  • 15.Nørgaard BL, Hjort J, Gaur S, et al. Clinical use of coronary CTA-derived FFR for decision-making in stable CAD. JACC Cardiovasc Imaging 2017;10(5):541–550. [DOI] [PubMed] [Google Scholar]
  • 16.Lu MT, Ferencik M, Roberts RS, et al. Noninvasive FFR derived from coronary CT angiography: management and outcomes in the PROMISE trial. JACC Cardiovasc Imaging 2017;10(11):1350–1358. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Douglas PS, De Bruyne B, Pontone G, et al. 1-year outcomes of FFRCT-guided care in patients with suspected coronary disease: the PLATFORM study. J Am Coll Cardiol 2016;68(5):435–445. [DOI] [PubMed] [Google Scholar]
  • 18.Douglas PS, Pontone G, Hlatky MA, et al. Clinical outcomes of fractional flow reserve by computed tomographic angiography-guided diagnostic strategies vs. usual care in patients with suspected coronary artery disease: the prospective longitudinal trial of FFR(CT): outcome and resource impacts study. Eur Heart J 2015;36(47):3359–3367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Tesche C, De Cecco CN, Baumann S, et al. Coronary CT angiography-derived fractional flow reserve: machine learning algorithm versus computational fluid dynamics modeling. Radiology 2018;288(1):64–72. [DOI] [PubMed] [Google Scholar]
  • 20.Coenen A, Lubbers MM, Kurata A, et al. Fractional flow reserve computed from noninvasive CT angiography data: diagnostic performance of an on-site clinician-operated computational fluid dynamics algorithm. Radiology 2015;274(3):674–683. [DOI] [PubMed] [Google Scholar]
  • 21.Kruk M, Wardziak Ł, Demkow M, et al. Workstation-based calculation of CTA-based FFR for intermediate stenosis. JACC Cardiovasc Imaging 2016;9(6):690–699. [DOI] [PubMed] [Google Scholar]
  • 22.Schwartz FR, Koweek LM, Nørgaard BL. Current evidence in cardiothoracic imaging: computed tomography-derived fractional flow reserve in stable chest pain. J Thorac Imaging 2019;34(1):12–17. [DOI] [PubMed] [Google Scholar]
  • 23.Min JK, Koo BK, Erglis A, et al. Effect of image quality on diagnostic accuracy of noninvasive fractional flow reserve: results from the prospective multicenter international DISCOVER-FLOW study. J Cardiovasc Comput Tomogr 2012;6(3):191–199. [DOI] [PubMed] [Google Scholar]
  • 24.Takakuwa KM, Halpern EJ, Gingold EL, Levin DC, Shofer FS. Radiation dose in a “triple rule-out” coronary CT angiography protocol of emergency department patients using 64-MDCT: the impact of ECG-based tube current modulation on age, sex, and body mass index. AJR Am J Roentgenol 2009;192(4):866–872. [DOI] [PubMed] [Google Scholar]
  • 25.Litt HI, Gatsonis C, Snyder B, et al. CT angiography for safe discharge of patients with possible acute coronary syndromes. N Engl J Med 2012;366(15):1393–1403. [DOI] [PubMed] [Google Scholar]
  • 26.Itu L, Sharma P, Kamen A, Suciu C, Comaniciu D. A novel coupling algorithm for computing blood flow in viscoelastic arterial models. Conf Proc IEEE Eng Med Biol Soc 2013;2013:727–730. [DOI] [PubMed] [Google Scholar]
  • 27.Verberne HJ, Acampa W, Anagnostopoulos C, et al. EANM procedural guidelines for radionuclide myocardial perfusion imaging with SPECT and SPECT/CT: 2015 revision. Eur J Nucl Med Mol Imaging 2015;42(12):1929–1940. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Cheitlin MD, Armstrong WF, Aurigemma GP, et al. ACC/AHA/ASE 2003 guideline update for the clinical application of echocardiography--summary article: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/ASE Committee to Update the 1997 Guidelines for the Clinical Application of Echocardiography). J Am Coll Cardiol 2003;42(5):954–970. [DOI] [PubMed] [Google Scholar]
  • 29.Nagel E, Lehmkuhl HB, Bocksch W, et al. Noninvasive diagnosis of ischemia-induced wall motion abnormalities with the use of high-dose dobutamine stress MRI: comparison with dobutamine stress echocardiography. Circulation 1999;99(6):763–770. [DOI] [PubMed] [Google Scholar]
  • 30.Yoo SM, Chun EJ, Lee HY, Min D, White CS. Computed tomography diagnosis of nonspecific acute chest pain in the emergency department: from typical acute coronary syndrome to various unusual mimics. J Thorac Imaging 2017;32(1):26–35. [DOI] [PubMed] [Google Scholar]
  • 31.Takakuwa KM, Halpern EJ. Evaluation of a “triple rule-out” coronary CT angiography protocol: use of 64-Section CT in low-to-moderate risk emergency department patients suspected of having acute coronary syndrome. Radiology 2008;248(2):438–446. [DOI] [PubMed] [Google Scholar]
  • 32.Douglas PS, Hoffmann U, Patel MR, et al. Outcomes of anatomical versus functional testing for coronary artery disease. N Engl J Med 2015;372(14):1291–1300. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Tonino PA, Fearon WF, De Bruyne B, et al. Angiographic versus functional severity of coronary artery stenoses in the FAME study fractional flow reserve versus angiography in multivessel evaluation. J Am Coll Cardiol 2010;55(25):2816–2821. [DOI] [PubMed] [Google Scholar]
  • 34.Sand NPR, Veien KT, Nielsen SS, et al. Prospective Comparison of FFR Derived From Coronary CT Angiography With SPECT Perfusion Imaging in Stable Coronary Artery Disease: The ReASSESS Study. JACC Cardiovasc Imaging 2018;11(11):1640–1650. [DOI] [PubMed] [Google Scholar]
  • 35.Mastrodicasa D, Albrecht MH, Schoepf UJ, et al. Artificial intelligence machine learning-based coronary CT fractional flow reserve (CT-FFRML): impact of iterative and filtered back projection reconstruction techniques. J Cardiovasc Comput Tomogr 2019;13(6):331–335. [DOI] [PubMed] [Google Scholar]

Articles from Radiology: Cardiothoracic Imaging are provided here courtesy of Radiological Society of North America

RESOURCES