MUNICH, Germany—A noninvasive method of measuring fractional flow reserve (FFR) based on data from computed tomography (CT) angiography is superior to CT angiography alone for diagnosing ischemic coronary artery disease (CAD). The results,  released during a late breaking session at the European Society of Cardiology (ESC) Congress on August 26, 2012, came despite the fact that the novel technology failed the trial’s primary endpoint.

The prospective study, presented by James K. Min, MD, of Cedars-Sinai Medical Center (Los Angeles, CA), was simultaneously published online ahead of print in the Journal of the American Medical Association.

The FFR-CT technology (CT-Flow; HeartFlow; Redwood City, CA) applies computational fluid dynamics to standard CT data to determine coronary pressure and flow and thus the hemodynamic significance of lesions.

Invasive FFR Used as Gold Standard 

For the DeFACTO (Diagnostic accuracy of Fractional flow reserve from Anatomic Computed TOmographic angiography) trial, Dr. Min and colleagues evaluated the performance of FFR-CT in diagnosing physiologically significant stenoses in 252 stable patients (406 vessels) with suspected or known CAD at 17 centers in 5 countries. Invasive FFR was set as the gold standard. All patients first underwent CT angiography with at least a 64-row scanner. Conventional FFR was performed during invasive angiography in vessels determined to have between 30% and 90% stenosis. FFR values of 0.80 or less were considered diagnostic ischemia.

Binary stenosis was found in 46.5% of patients by invasive angiography and 53.2% by CT angiography. Invasive FFR identified ischemia in 37.1%, while FFR-CT identified it in 53.3%. FFR-CT’s diagnostic accuracy of 73% did not meet the primary endpoint, with the lower boundary of the 95% confidence interval (67%) falling below the pre-specified lower limit of 70%, which represents a 15% increase in diagnostic accuracy over myocardial perfusion imaging and stress echocardiography in relation to invasive FFR.

FFR-CT Accuracy Driven by High Sensitivity 

Nevertheless, the diagnostic accuracy, sensitivity, and specificity of FFR-CT were superior to those of CT alone (table 1).

 Table 1. Per-Patient Diagnostic Performance 

When compared with CT detection of binary stenosis alone, FFR-CT showed greater discriminatory power to identify or exclude ischemia on a per-patient (difference in AUC of 0.13) and per-vessel (difference in AUC of 0.06) basis (both P < 0.001).

When analysis was restricted to patients with intermediate stenosis (30%-70%), for whom anatomic findings are more ambiguous with regard to ischemia, the diagnostic accuracy and sensitivity of FFR-CT were higher than with CT alone, and specificity was similar (table 2).

 Table 2. Per-Patient Diagnostic Performance for Intermediate Stenoses 

   
Direct per-vessel correlation of FFR-CT with FFR was generally good, with underestimation of ischemia by FFR-CT (mean difference 0.058; 95% CI 0.05-0.07).

Dr. Min said the study was limited by the fact that invasive angiography was performed based on the results of CT angiography, opening the possibility of referral bias. In addition, not every vessel was interrogated with invasive FFR, and patients with prior CABG and suspected ACS or in-stent restenosis were excluded. Most importantly, the study did not address the clinical question of whether FFR-CT-guided revascularization reduces ischemic events.

These data provide “proof of feasibility of FFR-CT for the diagnosis of lesion-specific ischemia,” Dr. Min said, adding that a plus for the technology is that it does not require subjecting patients to additional radiation.

 “Our findings also suggest that FFR-CT could be particularly useful for evaluating patients with arterial blockages of an intermediate severity, which are often the most difficult to assess noninvasively. This represents a large group of patients who unfortunately are often prone to frequent misdiagnosis,” Dr. Min noted in a press statement.

In the JAMA paper, Dr. Min and colleagues observe that the combined anatomic and physiologic information imparted by noninvasive FFR-CT after CT “may allow for more refined patient-physician discussions regarding treatment options in a manner not possible with either stress imaging or CT testing alone.”

 FFR-CT as Gatekeeper Against Unnecessary Angiography 

In particular, the high sensitivity and negative predictive value of FFR-CT means that “clinicians may be confident in not proceeding to invasive angiography in patients with stenoses on CT when FFR-CT results are normal,” Dr. Min and colleagues say.

However, they add, the relatively low specificity of FFR-CT suggests that “universal application of FFR-CT to guide invasive assessment may result in referral of a non-negligible number of patients without ischemia, and future studies will be needed to determine the potential clinico-economic effectiveness of FFR-CT for this particular indication.”

In an editorial accompanying the JAMA paper, Manesh R. Patel, MD, of Duke University Medical Center (Durham, NC), said FFR-CT “represents a novel and important innovation, with the possibility not only to diagnose but also to help direct invasive treatment.” He cautioned that judging the value of FFR-CT based on the diagnostic performance of noninvasive imaging tests compared with invasive angiography is “naive” since the latter should actually be compared with invasive angiography paired with invasive FFR.

“In real-world practice, the current noninvasive technologies used for diagnosis and risk stratification in stable elective patients prior to invasive angiography do not perform at the published diagnostic levels,” he writes, concluding that the DeFACTO results suggest that FFR-CT “may improve existing care and outperform established noninvasive technologies.”

 Noninvasive FFR Still a Dream 

However, commenting on the ESC presentation, Jean-Pierre Bassand, MD, of University Hospital Jean-Minjoz (Besançon Cedex, France), was disappointed by FFR-CT in its current state. He underlined that invasive FFR has been a “major breakthrough in interventional cardiology” and its use is strongly recommended by the guidelines because it has been shown in several studies including FAME to provide “what really matters to clinicians and patients alike: improved outcomes after revascularization.

“Noninvasive FFR is the dream of every interventional cardiologist,” but one that has yet to come true, Dr. Bassand said. Although the correlation between FFR-CT and invasive FFR in the current study is statistically significant, it is not meaningful to clinicians, he commented, adding, “I’m really questioning the diagnostic accuracy of this [new] technique.” In fact, FFR-CT is still in the early stages of development and widespread use would be “premature,” he concluded.

Note: Study coauthor Martin B. Leon, MD, is a faculty member of the Cardiovascular Research Foundation, which owns and operates TCTMD.

1. Diagnostic accuracy of Fractional flow reserve from Anatomic Computed TOmographic angiography: The De FACTO study. Presented at: European Society of Cardiology Congress; August 26, 2012; Munich, Germany.
2. Min JK, Leipsic J, Pencina MJ, et al. Diagnostic accuracy of fractional flow reserve from anatomic CT angiography. JAMA. 2012;Epub ahead of print.
3. Patel MR. Detecting obstructive coronary disease with CT angiography and noninvasive fractional flow reserve. JAMA. 2012;Epub ahead of print.

 Disclosures: 

• The trial was sponsored by HeartFlow.
• Dr. Min reports research support from GE Healthcare and Philips Medical.
• Dr. Patel reports serving as a consultant for Baxter, Bayer, Jansen, and Otsuka and receiving grants from AstraZeneca and Johnson and Johnson.
• Dr. Bassand reports no conflicts of interest.

 Related Stories: 

•  DISCOVER-FLOW Published: Noninvasive FFR Shows High Diagnostic Accuracy 
•  CT-Derived FFR Enables Accurate ‘Virtual Stenting’ 
•  Noninvasive MR Perfusion Testing Matches FFR in Identifying Ischemic Lesions