Is CT-derived fractional flow reserve superior to ischemia testing?

KEYWORDS:

• Angina
• coronary artery disease
• coronary CT angiography
• fractional flow reserve
• myocardial perfusion imaging
• testing

1. Introduction

Guidelines for the diagnosis and management of stable coronary artery disease (CAD) recommend noninvasive testing as the first-line diagnostic approach [1,2]. Noninvasive testing can be functional or anatomically based.

2. Functional testing

Noninvasive detection and quantification of jeopardized myocardium relies on functional testing during exercise or pharmacologically induced stress. Noninvasive ischemia testing (NIT) modalities provide different surrogates of myocardial ischemia such as electrocardiographic ST segment changes, echocardiographic left ventricular wall motion abnormalities, or perfusion defects detected by single-photon emission computed tomography (SPECT), position emission tomography (PET), or cardiac magnetic resonance (CMR) imaging [1,2]. In patients with stable CAD, NIT has demonstrated overall good diagnostic accuracy when using invasive angiography (ICA) as the reference standard, and the absence of myocardial ischemia by NIT has been associated with a favorable prognosis [1,2].

3. Anatomic testing

The other principle of noninvasive testing is direct visualization of the coronaries by coronary CT angiography (CTA). CTA is increasingly used in the first-line diagnostic workup of patients with suspected stable CAD [1,2]. Recent trials demonstrate that first-line CTA clarified the diagnosis of CAD more accurately than NIT, leading to improvements in the use of evidence-based preventive therapies [3,4]. Moreover, anatomical information determined by CTA is a stronger predictor of outcomes than functional data by NIT [3–6]. Since coronary atherosclerosis is a main contributor to the ischemic burden, it is not surprising that the prognostic value of NIT is attenuated when adjusted for CAD [6]. On the other hand, CTA has modest diagnostic specificity due to, e.g., calcification-induced artifacts, and CTA is a poor discriminator of lesion-specific ischemia as assessed by fractional flow reserve (FFR), which is the gold standard for revascularization decision-making [1,2,7–9]. Accordingly, when compared to conventional NIT, CTA may be associated with increased downstream ICA, revascularization rates, and costs particularly in higher risk populations in whom anatomical stenosis is more likely to occur [1–3]. Diagnostic strategies to more accurately identify patients, who do not require any further testing, are needed to offset the potentially higher ICA utilization after CTA. Therefore, guidelines recommend second-line functional testing if CTA shows CAD of uncertain functional significance [1]. As an alternative to NIT in this setting, several CT derived functional assessment tools have been introduced [2,7–10].

4. CT-derived fractional flow reserve (FFR_CT)

Computed tomography perfusion (CTP) enables the assessment of stress myocardial perfusion during vasodilatation. CTP has shown promising diagnostic and prognostic value in patients with CAD determined by CTA; however the majority of existing evidence and clinical experience of CT-derived functional assessment are based on the HeartFlow CT-derived FFR (FFR_CT) tool, which is the only such method cleared for clinical use by the United States Food and Drug Administration. FFR_CT is based on standard acquired CTA data sets from which patient-specific 3-dimensional anatomical and physiological models are derived, and application of computational fluid dynamics principles (CFD) enables computation of a 3-dimensional pressure map informing the physiological consequences of CAD in each point of the coronary tree [7]. CTA images are transmitted to a central FFR_CT laboratory (HeartFlow Inc, Redwood City, US) for analysis by experienced personnel and dedicated computers using machine learning algorithms, which are continuously refined by the incoming data. The processing time varies according to the CTA image quality and disease severity (~1-3 hours). The fact that FFR_CT requires off-site analysis has driven renewed interest in past generations of reduced order CFD and other principles requiring less comprehensive anatomic modeling and less computational requirements enabling workstation-based on-site analysis with resultant reduced analysis times. Although seemingly providing additive diagnostic information to CTA, these onsite functional applications have not yet reached clinical use standards.

5. FFR_CT, diagnostic value relative to conventional functional testing modalities

Extensive validation of FFR_CT testing has demonstrated high correlation and diagnostic performance against measured FFR, providing additional diagnostic information to CTA alone [7,8]. These advantages include patients with intermediate range lesions and high calcification, representing the most challenging scenarios in CTA interpretation [7,8]. Accordingly, current scientific efforts are directed toward the clinical value of FFR_CT relative to NIT. Although FFR_CT is a strict second-line test, traditionally, it has been compared with the first-line diagnostic performance of NIT [1]. For example, in a meta-analysis of the diagnostic performance of first-line NIT and FFF_CT in stable CAD using FFR as the reference standard, CMR and FFR_CT provided the highest per-patient sensitivity (both, 0.90), and stress-echo and SPECT the lowest (0,77 and 0.70), whereas CMR had the highest specificity (0.94) followed by SPECT, stress-echo, and FFR_CT (0.78, 075, and 0.71) [11]. However, emerging data reveal that the diagnostic performance of NIT assessed in a first-line diagnostic setting ¹cannot be extrapolated to the CTA–NIT testing scenario. In the prospective PACIFIC trial including 208 patients (554 vessels) with stable CAD and intermediate range stenosis, adding the information of SPECT or PET to CTA alone improved the per-patient or per-vessel specificity when compared to FFR, however, with a large sacrifice in sensitivity (per-patient, from 90% to 50% and 74%; per-vessel, from 72% to 35% and 64%, respectively) [8]. Similarly, in the prospective DANICAD and ReASSESS studies including a total of 435 patients with intermediate range stenosis by CTA, the diagnostic sensitivities of SPECT and CMR against measured FFR ranged between 36% and 41% (Table 1) [10,12].

Table 1. Diagnostic performance of CTA, FFR_CT, and different noninvasive ischemia testing modalities in patients with stable CAD and intermediate range coronary artery stenosis determined by CTA. All prospective studies measured FFR as the reference standard

Study [Ref.]	Testing modality	Per-patient			Per-vessel
		Sens, %	Spec, %	Acc, %	Sens, %	Spec, %	Acc, %
NXT [7]	FFRCT	86	79	81	84	86	86
	CTA	94	34	53	83	60	65
DANICAD [12]	SPECT	36	93	70	33	92	80
	CMR	41	84	67	28	88	75
PACIFIC [9]	FFRCT	96	63	78	90	86	87
	CTA	87	67	76	68	83	79
PACIFIC [8]	SPECT	57	94	77	39	96	82
	PET	87	84	85	81	75	79
ReASSESS [10]	FFRCT	91	55	70	-	-	-
	SPECT	41	86	68	-	-	-

NXT [7], 254 patients, 484 vessels; DANICAD [12], 148 patients, 452 vessels; PACIFIC [9], 157 patients, 505 vessels; PACIFIC [8], 208 patients, 554 vessels; and ReASSESS [10], 143 patients, no per-vessel analyses.

CMR, cardiac magnetic resonance; CTA, coronary CT angiography; FFR_CT, CTA-derived FFR; SPECT, Single photon emission CT; PET, Positron emission tomography.

6. FFR_CT, clinical utility and prognosis

In the context of the higher focus on evidence-based clinical practice and the rising health care costs, new diagnostic technologies such as FFR_CT need to demonstrate beyond superior diagnostic performance that they inform clinical decision-making and preferably improve outcomes in a cost-efficient manner. The ability of FFR_CT testing to reduce the number of unnecessary ICAs following CTA testing has been documented in several observational, prospective, and randomized settings. In the prospective PLATFORM (n = 380) and randomized FORECAST (N = 1400) trials of patients with stable CAD, first-line CTA with selective FFR_CT testing compared with usual care was associated with a safe 61% and 22% reduction in downstream ICA utilization and a 83% and 52% lower prevalence of ICAs showing no obstructive disease with associated lower or neutral downstream patient management cost weights [13,14]. Although a strategy of selective FFR_CT when compared to usual care testing in the FORECAST trial was not associated with improved short-term (9 months) prognosis, several nonrandomized real-world and post hoc studies with longer-term follow-up and one meta-analysis support the value of FFR_CT to inform clinical outcomes [14,15]. These findings together with the high diagnostic negative predictive value of FFR_CT in patients with a normal test result support integration of FFR_CT in the diagnostic workup of patients with intermediate stenosis to safely mitigate the use of additional downstream ICA testing. On the other hand, opposite to NIT, which assesses the presence of stress-induced abnormalities in myocardial perfusion or contractility caused by epicardial stenosis, and/or diffuse CAD, and/or microvascular dysfunction, FFR_CT discriminates lesions causing impairment of coronary flow, which may be relevant for revascularization and physiologically nonfocal diffuse CAD, which is best managed by medication [7].

7. FFR_CT, unresolved questions

Despite support of FFR_CT testing in the clinical community, challenges remain. Similar to CTA, FFR_CT testing cannot be performed in all patients. The proportion of patients in whom FFR_CT could not be calculated due to impaired image CTA quality (e.g. presence of severe motion and/or calcium induced artifacts) varies between studies (3% to 25%) [7,9,10,13,14]. Importantly, FFR_CT should not be used as a bail-out diagnostic strategy in patients with poor CTA image quality. Studies on how to optimize CTA acquisition for FFR_CT analysis and standardized FFR_CT interpretation criteria are needed. The diagnostic value of FFR_CT in patients with previous myocardial infarction or coronary intervention is unresolved. Funding and reimbursement difficulties in countries where the costs cannot be passed on to patients or insurance companies may pose a barrier for the clinical adoption of FFR_CT. Ongoing large-scale observational and randomized trials will further delineate issues related to the optimal FFR_CT testing interpretation strategy, the optimal timing of testing in the diagnostic pathway and the potential added diagnostic and prognostic value of the coronary plaque phenotype and other CT-derived functional metrics.

8. Conclusion

In patients with stable CAD and intermediate-range stenosis, FFR_CT is efficient in discriminating patients in whom no further testing is necessary from patients in whom ICA and possible revascularization are needed. In this subset of patients, traditional NIT has limited diagnostic value with high rates of false-negative results. Based on the emergent evidence and many years of clinical experience with CTA, NIT, and FFR_CT testing, it is the belief of this author group, when good image quality CTA is present, that FFR_CT trumps NIT for clinical decision-making in the majority of patients with stable chest pain and moderate CAD. Moreover, FFR_CT has obvious logistic advances compared to NIT since it is derived from the original standard acquired CT data sets without the need of additional testing. Accordingly, in the recent 2021AHA/ACC/ASE/CHEST/SAEM/SCCT/SCMR Guideline for the evaluation and diagnosis of chest pain, it is stated that FFR_CT in intermediate-high risk patients with chest pain and stenosis, 40%-90% (mid-proximal segment on CTA) can be useful for the diagnosis of vessel-specific ischemia and to guide clinical decision-making on revascularization (evidence level, 2A) [2]. The optimal clinical use of FFR_CT will be better defined in upcoming trials.

Declaration of interest

BL Nørgaard has previously received unrestricted institutional grant funding from HeartFlow. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers of this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.

Notes

1. ”that” is not meaningfull here