Value of CT-derived Fractional Flow Reserve in the Context of Outpatient Cardiac CT in Germany: A Propensity Score Matched Analysis

Keywords

cardiac - angiography - coronary angiography

Introduction

Coronary artery disease (CAD) remains the leading cause of death in Germany, accounting for approximately one third of all deaths [1]. Accurate diagnostics therefore play a crucial role in medical care and healthcare policy [2]. Coronary computed tomography angiography (CCTA) has established itself as a non-invasive procedure that allows for reliable visualization of the coronary arteries through high-resolution cross-sectional imaging. Its high negative predictive value makes it especially well suited for ruling out CAD and avoiding invasive procedures [3]. In fact, studies show that CCTA can improve prognosis and reduce the need for invasive procedures [4] [5]. In the SCOT-HEART study, the procedure led to a reduction in myocardial infarction and cardiovascular death compared to conventional routine tests [6]. With the decision by the Federal Joint Committee (G-BA) of January 18, 2024, CCTA is available in Germany as an outpatient service covered by statutory health insurance. The basis for conducting the test should be a pre-test probability of at least 15% [7].

For the large group of patients with intermediate stenoses, it is important to assess hemodynamic relevance. To do so, the CT-derived fractional flow reserve (FFR-CT) method is available, which enables virtual analysis based on computed tomography (CT) data [8]. A study by the NHS in England with over 90,000 patients shows that FFR-CT reduces the need for additional ischemia tests and cardiac catheterizations [9]. Despite high diagnostic accuracy, cost-effectiveness is currently a point of debate [10] [11]. There is also increasing focus on FFR-CT in Germany. While current recommendations for intermediate stenoses still include additional functional diagnostics such as cardiac stress MRI, stress echocardiography, or myocardial scintigraphy [2], FFR-CT could replace these procedures in the future and further reduce cardiac catheterizations [12].

The primary endpoint of this study was to compare the diagnostic accuracy of CCTA with FFR-CT and without FFR-CT, based on the positive predictive values compared to invasive coronary angiography. As a secondary endpoint, the study examined whether FFR-CT can reduce invasive procedures in outpatient settings. In addition, referral structures, indications, radiation exposure, and the need for further invasive and non-invasive diagnostics were studied following the introduction of outpatient reimbursement for CCTA.

Materials & Methods

Patient population

This monocentric, retrospective study included a total of 640 patients who had undergone outpatient CCTA and had evidence of at least mild coronary stenosis (>25%). The examinations were carried out consecutively between February 1, 2024 and March 1, 2025. Patients with prior coronary intervention, known CAD, acute coronary syndrome, and exclusion of coronary stenoses in CCTA were excluded. The data were made available in pseudonymized form. FFR-CT was performed on 107 patients. Since FFR-CT is not covered by health insurance in the context of stable CAD, the patients were exclusively private or self-paying. The 533 patients without FFR-CT served as a control group. The severity of the stenoses was divided into five categories: 0 = no stenosis (<25%); 1 = mild stenosis (25–50%); 2 = moderate stenosis (50–75%); 3 = severe stenosis (75–90% or >90%); and 4 = complete occlusion (100%).

Computed tomography

The CCTA examinations were performed using routine protocols on a state-of-the-art multi-layer device (GE Revolution Apex, GE Healthcare, USA).

Patients were pretreated with sublingual isosorbide dinitrate (Isoket 10 mg, Merus Labs Luxco II S.à r.l., Luxembourg) 3–5 minutes before the CCTA examination. If the initial heart rate was above 85 beats per minute (bpm), the patients received 5 mg of bisoprolol orally (Ratiopharm GmbH, Ulm, Germany) one hour before the CT scan.

With a heart rate between 65 and 85 bpm, up to 200 mg of esmolol (Brevibloc 10 mg/ml, Baxter Deutschland GmbH, Germany) was administered intravenously immediately before image acquisition to lower the heart rate and reduce motion blur. An intravenous injection of 50 ml of Iohexal (Accupaque 350 mg iodine/ml, GE Healthcare, Vienna, Austria) was followed by 50 ml of saline solution. The standard flow rate was 6.5 ml per second via an 18-gauge catheter in the elbow. Slight deviations occurred in cases of limited venous access, using a 16-gauge catheter with a flow rate of at least 5 ml/second.

To achieve the target signal-to-noise ratio (SNR) of 30, tube voltage and tube current were adjusted using the automatic SNR modulation of the CT software. A collimation of 512 × 0.625 mm was used with a z-detector coverage of 14–16 cm and a field of view of 32 cm. A z-detector coverage of 16 cm represents the maximum coverage. All scans were performed in high-resolution mode with a matrix size of 512 × 512. The layer thickness was 0.625 mm. The device's rotation time was 0.23 seconds, with the entire image being captured during a single rotation. For each patient, an additional non-contrast-enhanced low-dose cardiac CT scan was performed to calculate the calcium score, using the same CT parameter settings from the CCTA.

FFR determination

The fractional flow reserve was calculated using the artificial intelligence-supported software HeartFlow (HeartFlow Inc., Redwood City, USA). The Heartflow FFR-CT analysis uses deep learning algorithms and computational fluid dynamics (CFD) to create a personalized, digital 3D model of each patient's coronary arteries based on previously acquired CCTA image data. The patient's CCTA image data were uploaded to the Heartflow platform via a secure, cloud-based server. The incoming CCTA data were checked for quality by analysts to ensure the data’s suitability for analysis. Computer-based algorithms identified and extracted anatomical structures from the CCTA images for segmentation and creation of the patient's personal coronary artery model. A physiological model was then created based on the anatomical model. Maximal hyperemia was simulated to replicate the conditions of an invasive FFR measurement. Millions of complex equations were solved using CFD, resulting in a 3D model of coronary blood flow. The resulting model provided the calculated FFR-CT values along the modeled coronary arteries. A clinically relevant threshold was defined as FFR ≤ 0.80. The Heartflow FFRCT analysis was transmitted via a secure web portal. The image data underwent rigorous quality control. The rejection rate was 0.9% during the observation period (2 out of 109 patients). In one case, due to a change in breathing maneuver, the cranial boundary of the LAD wall was not captured fully; in a second case, there was excessive image noise and motion blur of the RCA in an obese and tachycardic patient. These patients were excluded from the analysis.

Coronary angiography

The primary endpoint of this study was the positive predictive value (PPV) per patient of CCTAs performed with FFR-CT and without FFR-CT, compared to the gold standard of invasive coronary angiography. The coronary angiographies were evaluated by the interventional cardiologist based on the clinical assessment of the presence of a relevant coronary stenosis. In selected cases, invasive FFR or intravascular ultrasound (IVUS) were used for additional assessment, if considered necessary by the treating cardiologist. An invasive FFR of ≤0.8 was considered hemodynamically relevant. A cardiac catheterization without relevant findings was defined as a coronary angiography in which no hemodynamically relevant stenosis or myocardial bridging was found, therefore no interventional or surgical therapy was performed, and – if performed – an invasive FFR showed a non-significant finding (FFR>0.80).

Statistical analysis

To analyze the data in our study, we used the software IBM SPSS Statistics version 29.0.2. Categorical variables were expressed as absolute and relative frequencies. The Mann-Whitney-Wilcoxon test was used to compare independent groups. Relationships between ordinal and metric data were calculated using the chi-square test. Normally distributed variables were represented as mean ± standard deviation (SD) and compared using the t-test for independent samples. Propensity score (PS) matching was performed to equalize differences in the initial characteristics of the patients. The aim was to create comparable cohorts of patients with FFR-CT and without FFR-CT. The following variables were included when calculating the propensity score: gender, age, diabetes mellitus, arterial hypertension, hyperlipidemia, smoking, positive family history, carotid artery plaques, and high cardiovascular risk profile (≥3 risk factors), as well as a history of ventricular arrhythmia. Missing values were replaced using fivefold multiple imputation. Matching was performed using the nearest-neighbor method at a ratio of 1:1 without replacement, using a caliper of 0.2. To test the balance between the groups, standardized mean differences were calculated before and after matching; values <0.2 were considered an indication of sufficient comparability. Additionally, correlations between metric variables were calculated based on Pearson’s correlation coefficient. The Sankey diagram was created online using SankeyMATIC. A correlation analysis between FFR-CT and invasive FFR measured during cardiac catheterization was performed and Spearman’s rank correlation coefficient was determined. Previously, the Shapiro-Wilk test was used to test for normal distribution. The p-values were 0.006 (FFR-CT) and 0.001 (invasive FFR), indicating a non-normal distribution. Since a high correlation does not necessarily reflect good agreement between two diagnostic methods, an analysis of the intraclass correlation coefficient (ICC) was also performed. A value of p<0.05 was considered hemodynamically relevant.

Results

A total of 640 consecutive patients with a first diagnosis of coronary artery disease (mild stenosis ≥25%) were included in the outpatient CCTA study. Of these, 255 were female (39.8%) and the mean age was 65.4±9.6 years. The vast majority of patients were referred for CCTA by cardiologists (89.7%), followed by internists (6.7%), and general practitioners (2.5%; [Fig. 1] A). The indications for performing CCTA were mostly symptoms such as typical angina pectoris (17.8%), atypical angina pectoris (29.8%), and dyspnea (40.3%) paired with a high cardiovascular risk profile (≥3 risk factors; 69.2%; [Fig. 1] B). Less frequently, pathological examination findings such as ergometry (14.2%), ventricular arrhythmias (6.7%), or atrial fibrillation (9.2%) led to outpatient CCTA ([Fig. 1] B). Routine diagnostics prior to CCTA included transthoracic echocardiography (96.6%), while ergometry (39.9%) and stress echocardiography (6.3%) were performed significantly less frequently ([Fig. 1] C). Plaques in the carotid artery were also frequently detected beforehand using duplex sonography. The average pre-test probability in the presence of typical angina pectoris was 59.2±22.9% and in the presence of atypical angina pectoris up to 44.6±16.8% ([Fig. 1] D).

Fig. 1 Referral structure, indication, diagnostics, and pre-test probability for outpatient coronary CT angiography. A Percentage distribution of referring physicians for 640 outpatient coronary CT angiographies. B Indications for cardiac CT in percent. A high cardiovascular risk was present in addition to the patient's symptoms or pathological examination findings. C Cardiological diagnostics prior to cardiac CT in percent and absolute values. D Percentage distribution of typical and atypical angina pectoris and pre-test probability.

To evaluate the benefit of using FFR-CT in outpatient CCTA, patients were divided according to FFR-CT (n=107) and no FFR-CT (n=533). [Table 1] lists the patient characteristics of both groups. Differences were found between the groups with regard to age (67.3±7.8 versus 65.0±9.8 years; p=0.024), sex (female 16.8 versus 44.5%; p<0.001), and cardiovascular risk factors such as hyperlipidemia (57.9 versus 48.6%; p=0.078), smoking (23.3 versus 41.8%; p=0.027), and plaques in the carotid artery (47.7 versus 30.8%; p<0.001). A high cardiovascular risk profile was found in 55.1% of the FFR-CT group and in 72.1% of the no FFR-CT group (p<0.001; [Table 1]). Furthermore, the group without FFR-CT had more symptomatic patients with regard to atypical angina pectoris (19.3% versus 31.9%; p<0.001) and dyspnea (20.6% versus 44.7%; p=0.002).

Table 1 Characteristics of patients with outpatient cardiac CT depending on whether FFR-CT analysis was performed.

	FFR-CT		no FFR-CT
	n = 107		n = 533		p-value
	n	% or mean ± SD	n	% or mean ± SD
BMI = body mass index, PAD = peripheral artery disease, CCS = Canadian Cardiology Society (classification of angina pectoris), NYHA = New York Heart Association (classification of dyspnea), LVEF = left ventricular ejection fraction, TTE = transthoracic echocardiography; eGFR = estimated glomerular filtration rate, TSH = thyroid-stimulating hormone; SD = standard deviation. Chi-square test, Mann-Whitney-Wilcoxon test, or independent samples t-test.
Age (years)	107	67.3 ± 7.8	533	65.0 ± 9.8	0.024
Female gender	18	16.8	237	44.5	<0.001
Cardiovascular risk factors
Arterial hypertension	69	64.5	341	64.0	0.86
Diabetes mellitus	14	13.1	115	21.6	0.15
Hyperlipidemia	62	57.9	259	48.6	0.078
Smoking	25	23.3	223	41.8	0.027
Positive family history	34	31.8	165	31.0	0.80
Carotid artery plaques	51	47.7	164	30.8	<0.001
High cardiovascular risk	59	55.1	384	72.1	<0.001
Pathology
Angina pectoris	22	20.6	93	17.4	0.20
CCS I	2	1.9	17	3.2
CCS II	18	16.8	71	13.3
CCS III	2	1.9	5	0.9
Atypical angina pectoris	21	19.3	170	31.9	<0.001
Dyspnea	22	20.6	238	44.7	0.002
NYHA I	4	3.7	2	0.4
NYHA II	18	16.9	221	41.5
NYHA III	0	0.0	14	2.6
Atrial fibrillation	6	6.5	53	9.9	0.19
Ventricular arrhythmia	12	11.2	40	7.5	<0.001
Cardiac diagnostics
Pre-test probability (%)	41	64.3 ± 18.0	298	48.5 ± 20.6	<0.001
LVEF (TTE, %)	85	57.5 ± 3.6	85	58.7 ± 5.5	0.18
Wall motion abnormalities (at rest)	0	0.0	16	3.0	0.10
Pathological ergometry	19	17.8	72	13.5	0.16
Pathological stress echocardiography	3	2.8	8	1.5	0.34
Laboratory diagnostics
Creatinine	107	1.0 ± 0.2	505	0.9 ± 0.2	0.18
eGFR	73	75.8 ± 12.3	359	75.1 ± 15.1	0.63
TSH	107	1.7 ± 1.1	508	1.7 ± 1.5	0.88
Cardiac CT
Radiation dose (mGy*cm)	107	144.2 ± 61.0	533	158.5 ± 82.6	0.074
Contrast agent (ml)	107	71.5 ± 3.8	533	71.7 ± 3.9	0.68

The Agatston score was not significantly different between the group without and with FFR-CT (425.6±634.1 versus 474.0±580.4; p=0.466). In the group without FFR-CT, 279 people (52.4%) showed significant coronary stenosis (luminal narrowing CT angiography ≥50%; [Fig. 2]). Of these patients, 146 (27.4%) subsequently underwent invasive coronary angiography, while 15 (2.8%) underwent ischemia diagnostics and 118 (22.1%) were treated with medication alone. Eighty-eight patients (60.3%) underwent percutaneous coronary intervention (PCI), and six (4.1%) underwent coronary artery bypass grafting (CABG). Two cases (1.4%) were invasively classified as myocardial bridging. In 48 cases (32.9%), the cardiac catheterization revealed no relevant findings ([Fig. 2]). The positive predictive value of CCTA without FFR-CT was 67.1% ([Fig. 2]).

Fig. 2 Diagnostic and therapeutic course after CCTA with FFR-CT and without FFR-CT. Description of the further course for a total of 640 patients who underwent coronary CT angiography (CCTA). Listed here are the diagnostic and therapeutic measures depending on the result of the CCTA (positive/negative) and – if performed – the FFR-CT. The data show frequencies and percentage distributions for invasive diagnostics, supplementary ischemia diagnostics, drug therapy, and revascularization (PCI, CABG surgery). PCI=percutaneous coronary intervention; CABG=coronary artery bypass grafting; FFR=fractional flow reserve.

FFR-CT was performed on 107 patients. In 80 cases (74.8%), normal values were found (FFR≥0.80), and 27 cases (25.2%) had pathological findings (FFR<0.80). Of the latter, 26 (24.3%) underwent an invasive cardiac catheterization. Of these, 19 cases (73.1%) underwent PCI, three patients (11.5%) underwent CABG surgery, and one case (3.9%) was diagnosed with hemodynamically relevant myocardial bridging. In three cases (11.5%), the cardiac catheterization produced no relevant findings. An invasive FFR measurement was performed in nine cases (34.6%). The positive predictive value of CCTA with FFR-CT was 88.5% ([Fig. 2]). A significant difference was found between the positive predictive values with FFR-CT and without FFR-CT (p=0.028).

To align the baseline characteristics, propensity score matching was performed to create two comparable cohorts with FFR-CT and without FFR-CT. The calculation included basic clinical variables (including age, gender, cardiovascular risk factors, carotid plaques, ventricular arrhythmias). After the matching step, two groups of 105 patients each were available for analysis ([Table 2]). Propensity score matching revealed no significant difference in the Agatston score between the two matched groups (CCTA: 527.5±762.4 versus FFR-CT: 441.8±840.7; p=0.331). In the FFR-CT group, 79 individuals (75.2%) showed normal results (FFR ≥0.80), while 26 patients (24.8%) had pathological findings (FFR <0.80) ([Fig. 3]). Of the latter, 25 (23.8%) underwent an invasive coronary angiography. Of these, 18 cases (72.0%) underwent PCI, three patients (12.0%) underwent CABG surgery, and one case (4.0%) was diagnosed with hemodynamically relevant myocardial bridging. Invasive FFR measurement was performed in nine patients (34.6%). In the FFR-CT-negative cases, no cardiac catheterization was performed, however, in two cases (1.9%) additional functional ischemia diagnostics were performed ([Fig. 3]). The positive predictive value of CCTA with FFR-CT after propensity score matching was 88.0% ([Fig. 3]).

Table 2 Characteristics of patients with outpatient cardiac CT depending on whether FFR-CT analysis was performed after propensity score matching.

	FFR-CT		no FFR-CT
	n = 105		n = 105		p-value
	n	% or mean ± SD	n	% or mean ± SD
CCS = Canadian Cardiology Society (classification of angina pectoris); NYHA = New York Heart Association (classification of dyspnea); LVEF = left ventricular ejection fraction; TTE = transthoracic echocardiography; eGFR = estimated glomerular filtration rate; TSH = thyroid-stimulating hormone; SD = standard deviation. Chi-square test, Mann-Whitney-Wilcoxon test, or independent samples t-test.
Age (years)	105	67.1 ± 7.7	105	66.3 ± 8.6	0.47
Female gender	18	17.1	19	18.1	0.81
Cardiovascular risk factors
Arterial hypertension	68	64.8	67	63.8	0.69
Diabetes mellitus	14	13.1	16	15.2	0.80
Hyperlipidemia	62	59.0	64	60.9	0.76
Smoking	24	22.9	22	21.0	0.55
Positive family history	33	31.4	30	28.6	0.41
Carotid artery plaques	49	46.7	48	45.7	0.88
High cardiovascular risk	59	56.2	57	54.3	0.59
Pathology
Angina pectoris	21	20.0	20	17.3	0.45
CCS I	2	1.9	1	1.0
CCS II	17	16.2	17	16.2
CCS III	2	1.9	2	1.9
Atypical angina pectoris	21	20.0	23	21.9	0.63
Dyspnea	20	19.1	23	21.9	0.44
NYHA I	4	3.7	2	1.9
NYHA II	18	16.9	20	17.3
NYHA III	0	0.0	1	1.9
Atrial fibrillation	6	5.7	5	4.8	0.80
Ventricular arrhythmia	11	10.5	13	12.4	0.96
Cardiac diagnostics
Pre-test probability (%)	40	63.7 ± 17.7	53	66.4 ± 14.9	0.44
LVEF (TTE, %)	84	56.9 ± 7.2	104	58.4 ± 5.7	0.18
Wall motion abnormalities (at rest)	0	0.0	2	1.9	0.10
Pathological ergometry	19	17.8	72	13.5	0.16
Pathological stress echocardiography	3	2.8	8	1.5	0.34
Laboratory diagnostics
Creatinine	105	1.0 ± 0.2	101	1.0 ± 0.2	0.67
eGFR	72	76.1 ± 12.2	64	71.9 ± 14.7	0.17
TSH	105	1.7 ± 1.1	102	1.7 ± 1.1	0.31
Cardiac CT
Radiation dose (mGy*cm)	105	144.2 ± 61.6	105	159.9 ± 92.4	0.11
Contrast agent (ml)	105	71.5 ± 3.8	105	72.1 ± 4.1	0.30

Fig. 3 Analysis of patients with FFR-CT and without FFR-CT after propensity score matching. Sankey plot of the diagnostic and therapeutic course of patients with FFR-CT and without FFR-CT (n=105 for each) after propensity score matching. PCI=percutaneous coronary intervention; CABG=coronary artery bypass grafting; FFR=fractional flow reserve; CCTA=coronary CT angiography.

In the group without FFR-CT, CCTA examinations showed no relevant stenosis in 39 cases (37.1%) ([Fig. 3]). No patient underwent coronary angiography here; however, further ischemia diagnostics were performed in two cases (1.9%). In 66 patients (62.9%), CCTA revealed a relevant stenosis (≥50%). Of the latter, 30 patients (28.6%) underwent an invasive coronary angiography. In 15 cases (50.0%), a PCI was performed and for six patients (20.0%), a CABG operation ([Fig. 3]). Eight examinations (26.7%) yielded no relevant findings in the cardiac catheter; in one case (3.3%), purely drug-based therapy was used. A total of 33 patients (31.4%) received conservative therapy without undergoing cardiac catheterization ([Fig. 3]). Without predictive value of CCTA without FFR-CT after propensity score matching was 73.3% ([Fig. 3]). Statistically, compared to the FFR-CT group, there was no statistical difference between the positive predictive values in the smaller group size (n=105 in each case) (p=0.176).

In nine patients with FFR-CT, the FFR was measured invasively during cardiac catheterization. A good correlation was found between the two measurement methods, with a Spearman rank correlation coefficient of 0.87 ([Fig. 4]) and an intraclass correlation coefficient (ICC) of 0.95.

Fig. 4 Comparison of FFR-CT and invasive FFR. A Left: As an example, an FFR-CT measurement yields a value of 0.89 in the RCx. Center: Guided catheter with pressure wire inserted in the RCx for invasive FFR measurement (fluoroscopy during cardiac catheterization). Right: Invasive FFR measurement during a cardiac catheterization. The FFR measurement value in yellow also yields 0.89. B Left: Measurement table with FFR-CT and corresponding invasive FFR during cardiac catheterization (CCT). Right: Correlation of FFR-CT and invasive FFR during cardiac catheterization (CCT). Pearson’s correlation coefficient.

Discussion

The decision of the Federal Joint Committee dated January 18, 2024, establishes CCTA for the first time as an outpatient health insurance benefit in Germany [7]. This step represents an important milestone for the non-invasive diagnosis of CAD, and it follows international guidelines that give a central role to CCTA when ruling out CAD due to its high negative predictive value [3]. This will improve access to accurate diagnostics across the broader healthcare system and should reduce the need for invasive procedures.

Our study provides, for the first time, data on referral structures and indications in a German metropolitan area following implementation of the Federal Joint Committee’s decision. The majority of patients were referred by cardiologists. In addition to symptoms such as angina pectoris or dyspnea, a risk profile with ≥3 risk factors was often decisive, and carotid plaques were also frequently found. In the majority of cases, the basic cardiological diagnostics required by the Federal Joint Committee have been carried out and a suspicion of chronic coronary syndrome has been established [2]. However, the referrals did not indicate whether pre-test probability had been established, so it remains unclear whether this was taken into account when determining the indication.

In contrast to CCTA, FFR-CT has not yet been included by the Federal Joint Committee, which is due mainly due to its unclear cost-benefit assessment [2]. The software used in this study (HeartFlow) is based on computational fluid dynamics and third-party, high-performance computers. Numerous studies have demonstrated high diagnostic accuracy compared to invasive FFR [13] [14] [15], with significantly higher specificity than CCTA at similar sensitivity [15]. Our study showed a very high correlation between FFR-CT and invasive FFR (Spearman r = 0.87; ICC 0.95), although only in a small subgroup. In addition to the US software HeartFlow, European systems are now also available, such as CorEx (Spimed AI, France) and the cFFR approach from Siemens Healthineers, all of which have performed well clinically in multiple studies [16] [17] [18] [19] [20].

After propensity score matching, the positive predictive value of FFR-CT was 88%, which is significantly higher than that of CCTA without FFR-CT (73%). This led to a reduction in cardiac catheterizations, as patients with unremarkable FFR-CT findings were treated conservatively. However, the significance of false negative results could not be determined on the basis of the data available. Similar results have been reported in large registries such as the NHS analysis with >90,000 patients [9].

Another advantage of FFR-CT is its predictive power: In the PROMISE substudy, an FFR-CT value ≤0.80 was a better predictor for revascularization or MACE than CCTA stenosis assessment [21]. ADVANCE and ADVANCE-DK also confirmed the increased risk with low FFR-CT, independent of coronary calcium [22] [23]. A meta-analysis of almost 5,700 patients further showed an increased risk of infarctions, revascularizations, and MACE in patients with FFR-CT ≤0.80 as well as a continual increase in risk as this value decreases [24].

The cost issue is frequently cited as a counterargument to FFR-CT. While some analyses describe higher initial costs, data from systems with widespread implementation (e.g. NHS England) show that savings are possible in the medium term [9]. Although costs are currently high and cost-effectiveness compared to other non-invasive ischemia tests has not been demonstrated [11], direct FFR-CT analyses can reduce follow-up visits and circumvent regional bottlenecks for alternative tests such as non-invasive ischemia diagnostics.

This could have the following implications for Germany: The inclusion of FFR-CT examinations in outpatient reimbursement should be reviewed again, and scientific support should be provided as part of this review process. It would also make sense for the Federal Joint Committee to reassess its decision, particularly with regard to providing a patient-centric, risk-adjusted, and efficient diagnostics.

Limitations

Our study has several limitations. This is a retrospective analysis, which means that biases in data collection and interpretation are possible. The number of patients with FFR-CT was comparatively small at 107, which limits the significance, especially for subgroups. In addition, the analysis was conducted monocentrically in one metropolitan region, so its transferability to other healthcare structures is limited. The comparison with invasive FFR was performed only in a small subcohort and must be interpreted with caution. On top of that, only a few patients received an invasive reference examination with FFR measurement, which limits the significance of the results regarding test quality. Since, for ethical reasons, only pathological FFR-CT findings were verified invasively, there is a potential verification bias. The influence of coronary calcifications on diagnostic accuracy was not examined systematically and may have particularly affected the results of the CCTA examinations. Additionally, during the study period, FFR-CT was available only to privately insured individuals or self-payers, which may lead to social or economic selection bias. Despite propensity score matching, not all confounders can be excluded. Larger, prospective, multicenter studies are needed to confirm our results. Since the sample size was limited after propensity score matching, the study may be underpowered to detect significant differences in positive predictive values.

Conclusions

Outpatient treatment using CCTA examinations was implemented successfully, with a high referral rate from cardiologists and a risk-oriented approach to indications. FFR-CT has the potential to increase the predictive value of CCTA and further reduce invasive procedures. The integration of FFR-CT in routine practice offers the opportunity to combine anatomical and functional information, thereby improving diagnostics, prognosis, and clinical outcomes over the long term.

Clinical relevance

• FFR-CT could increase diagnostic accuracy in outpatient care and reduce invasive procedures.

• By better differentiating hemodynamically relevant stenoses, FFR-CT can support targeted therapy decisions in stable CAD.

• FFR-CT appears to be a useful addition to CCTA and could be included in outpatient standard care in the future.

• The close correlation with invasive FFR confirms the practical benefit of the procedure in everyday clinical practice.

• Literatur

• 1 DESTATIS Todesursachenstatistik 2023. Zugriff am 26. August 2025 unter: https://www.destatis.de/DE/Themen/Gesellschaft-Umwelt/Gesundheit/Todesursachen/_inhalt.html
• 2 Korosoglou G, Thiele H, Silber S. et al. Bedarfs- und leitliniengerechte Diagnostik bei symptomatischer obstruktiver koronarer Herzkrankheit mittels Kardio-CT und MRT. Kardiologie 2023; 17: 406-417
• 3 Knuuti J, Ballo H, Juarez-Orozco LE. et al. The performance of non-invasive tests to rule-in and rule-out significant coronary artery stenosis in patients with stable angina. A meta analysis focused on post-test disease probability. Eur Heart J 2018; 39: 3322-3330
• 4 Williams MC, Wereski R, Tuck C. et al. Coronary CT angiography-guided management of patients with stable chest pain: 10-year outcomes from the SCOT-HEART randomised controlled trial in Scotland. Lancet 2025; 405: 329-337
• 5 Marwan M, Achenbach S, Korosoglou G. et al. German cardiac CT registry: Indications, procedural data and clinical consequences in 7061 patients undergoing cardiac computed tomography. Int J Cardiovasc Imaging 2018; 34: 807-819
• 6 Newby DE, Adamson PD, Berry C. et al. Coronary CT angiography and 5-year risk of myocardial infarction. N Engl J Med 2018; 379: 924-933
• 7 Gemeinsamer Bundesausschuss 2024. Zugriff am 16. Mai 2025 unter: https://www.g-ba.de/beschluesse/6418/
• 8 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: 433-445
• 9 Fairbairn TA, Mullen L, Nicol E. et al. Implementation of a national AI technology program on cardiovascular outcomes and the health system. Nat Med 2025; 31: 1903-1910
• 10 Karády J, Mayrhofer T, Ivanov A. et al. Cost-effectiveness Analysis of Anatomic vs Functional Index Testing in Patients With Low-Risk Stable Chest Pain. JAMA Netw Open 2020; 3: e2028312
• 11 Mittal TK, Hothi SS, Venugopal V. et al. The Use and Efficacy of FFR-CT: Real-World Multicenter Audit of Clinical Data With Cost Analysis. JACC Cardiovasc Imaging 2023; 16: 1056-1065
• 12 Kloth C, Brendel JM, Kübler J. et al. CT-FFR: How a new technology could transform cardiovascular diagnostic imaging. Rofo 2025;
• 13 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: 1989-1997
• 14 Min JK, Leipsic J, Pencina MJ. et al. Diagnostic accuracy of fractional flow reserve from anatomic CT angiography. JAMA 2012; 308: 1237-1245
• 15 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: 1145-1155
• 16 Glessgen CG, Boulougouri M, Vallée JP. et al. Artificial intelligence-based opportunistic detection of coronary artery stenosis on aortic computed tomography angiography in emergency department patients with acute chest pain. Eur Heart J Open 2023; 3: oead088
• 17 Brendel JM, Walterspiel J, Hagen F. et al. Coronary artery disease detection using deep learning and ultrahigh-resolution photon-counting coronary CT angiography. Diagn Interv Imaging 2025; 106: 68-75
• 18 Kübler J, Brendel JM, Küstner T. et al. Artificial intelligence-enhanced detection of subclinical coronary artery disease in athletes: Diagnostic performance and limitations. Int J Cardiovasc Imaging 2024; 40: 2503-2511
• 19 Brendel JM, Walterspiel J, Hagen F. et al. Coronary artery disease evaluation during transcatheter aortic valve replacement work-up using photon-counting CT and artificial intelligence. Diagn Interv Imaging 2024; 105: 273-280
• 20 Renker M, Schoepf UJ, Wang R. et al. Comparison of diagnostic value of a novel noninvasive coronary computed tomography angiography method versus standard coronary angiography for assessing fractional flow reserve. Am J Cardiol 2014; 114: 1303-1308
• 21 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: 1350-1358
• 22 Patel MR, Nørgaard BL, Fairbairn TA. et al. 1-Year impact on medical practice and clinical outcomes of FFR(CT): The ADVANCE registry. JACC Cardiovasc Imaging 2020; 13: 97-105
• 23 Madsen KT, Noergaard BL, Oevrehus KA. et al. Prognostic value of FFRCT in patients with stable chest pain – a 3-year follow-up of the ADVANCE-DK registry. Eur Heart J 2022; 43
• 24 Nørgaard BL, Gaur S, Fairbairn TA. et al. Prognostic value of coronary computed tomography angiographic derived fractional flow reserve: A systematic review and meta-analysis. Heart 2022; 108: 194-202