FAST-FFR: FFR From Angiography Accuracy Similar to Pressure Wire

Neil Osterweil

September 27, 2018

SAN DIEGO — Fractional-flow reserve values derived from routine coronary angiography (FFRangio) are highly sensitive, specific, and accurate at predicting FFR derived via coronary pressure wire, results of the FAST-FFR trial show.

"FFRangio may provide an easier and potentially faster method for performing physiology-guided assessment of the overall coronary angiogram with similar accuracy to the reference standard, coronary pressure wire-based FFR," said investigator William F. Fearon, MD, Stanford University School of Medicine, Stanford, California.

"This may translate into a greater percentage of patients undergoing physiologic guidance for revascularization decisions and, ultimately, improve long-term outcomes," he said during a press briefing here at Transcatheter Cardiovascular Therapeutics 2018.

The trial results were published online September 24 in Circulation to coincide with Fearon's presentation of the data in a late-breaking clinical science session.

Asked by briefing moderator C. Michael Gibson, MS, MD, Beth Israel-Deaconess Medical Center, Boston, whether he was surprised by the results, Fearon confessed that he was.

"As a coronary physiologist, we don't believe the angiogram. That's why we measure with a pressure wire, so I was definitely impressed with how high the accuracy was," he replied.

Fearon, Gibson, and other panelists agreed that the findings are likely to be practice-changing, because the elimination of the pressure wire will remove some of the additional risk and technical challenges of FFR, as well as some of the expense and adverse effects associated with hyperemic agents used to measure it.

However, FFRangio, which involves 3-dimensional reconstruction of the coronary arterial system to estimate resistance and flow across a stenosis, currently adds 5 to 8 minutes to a percutaneous angiography procedure, Fearon said.

"Five to 8 minutes to an interventional cardiologist equals 5 to 8 hours, so it will be important to get the time down," Gibson said.


Although FFR-guided revascularization in the cardiac cath lab has been shown in multicenter randomized trials to improve clinical outcomes, the technique isn't used as often as it should be for a variety of reasons, Fearon said.

"Certainly the need for a wire, the handling characteristics of the wire, the pressure drift associated with the wire, the need for hyperemic agent, the extra time it takes and the cost all have limited its use," Fearon said. "And so for that reason, having a method for measuring FFR based on the angiogram alone and eliminating the need for a wire and hyperemic agent would be advantageous."

The FAST-FFR (FFRangio Accuracy vs Standard FFR) trial tested FFRangio and standard pressure-wire-guided FFR in 319 vessels of 301 patients from 10 centers in Europe, Israel, and the United States.

Patients with suspected coronary artery disease underwent routine coronary angiography with FFR measured with a coronary pressure wire and hyperemic stimulus in coronary lesions of various size and severity.

FFR was measured in standard fashion with intravenous or intracoronary adenosine or intracoronary papaverine. FFR tracings were sent to an FFR core laboratory for review, and reviewers were blinded to FFRangio values.

On-site operators blinded to FFR calculated FFRangio from angiograms of the arteries acquired in two or more different projections. The images were interpreted using proprietary software.

The trial met its coprimary end point of sensitivity and specificity of the dichotomously scored FFRangio for predicting FFR derived by pressure wire, using a cutoff value 0.80 or less.

The mean FFR was 0.81, and the mean FFRangio was 0.80. The respective median values were 0.83 and 0.82. In the FFR group, 43% of values were 0.80 or less, compared with 46% of values in the FFRangio group.

The sensitivity of FFRangio was 93.5% and the specificity was 91.2%.

The diagnostic accuracy was 92.2%, positive predictive value was 89.0%, and negative predictive value was 94.8%.

The diagnostic accuracy remained high (86.9%) even when the analysis was restricted to FFR values around the cutoff point (0.75 - 0.85), with 88.5% sensitivity and 85.1% specificity, Fearon noted.

FFRangio remained highly correlated with FFR over the entire range of FFR values, and was successfully measured in nearly all cases, he added.

In 24 cases, the FFRangio and standard FFR results were discordant, but the only parameter examined that differed significantly between concordant and discordant results was target vessel type, with significantly more lesions in the right coronary artery and significantly fewer lesions in the left anterior descending artery in the discordant group.

Fearon acknowledged that the investigators did not directly assess the time it takes to calculate FFRangio compared with pressure-wire delivered FFR, and that patients with left main disease, low ejection fraction, and in-stent restenosis were excluded from the trial and will need to be studied in the future.

"I think this technology is really a game-changer in terms of incorporating all the work that you've done in clinical practice," discussant Alexandra J. Lansky, MD, Yale School of Medicine, New Haven, Connecticut, told Fearon after his presentation.

She noted, however, that the trial represents a best-case scenario, with selected patients and highly trained professionals, and asked Fearon about the potential pitfalls related to the use of FFRangio.

"This was most centers' first time using this, so there is a bit of a learning curve, and I think we certainly saw that in our own site improvement as we gained experience with the system," Fearon said.

He added that "one of the main issues is obtaining [high] quality angiograms. It's not really a limitation, because we should all be doing this when we perform angiograms."

Discussant Stephen G. Ellis, MD, Cleveland Clinic Foundation, said that "this is such an important area because we need to bring physiology or at least estimated physiology to all or most of our patients in the cath lab."

Morton J. Kern, MD, University of California, Irvine, a discussant at the briefing, called FFRangio "a real advance in the field.

"The use of FFR now for better outcomes is not as great as it should be because of the limitations of mechanical wires and the technique and the time it takes. But if you can provide a better ischemic-derived marker for treatment, then you're going to get better outcomes. I think it's the hope of this process that it will improve the ability of the user to identify ischemic lesions," he commented.

In an editorial accompanying the publication in Circulation, Michail I. Papafaklis, MD, PhD, from Barts Heart Centre, St Bartholomew's Hospital, London, and Andreas Baumbach, MD, William Harvey Research Institute, Queen Mary University of London, ask whether this class of virtual functional tools could replace invasive FFR.

"To date, the results have shown high but not optimal numeric comparisons against the gold standard, and thus, there will always be some who will reply negatively to our question on the basis of a small but existent number of false positives and, in particular, false negative cases," they write. "However, no metric is probably perfect, and ultimately, success in predicting clinical outcomes is vital.

"Following the paradigm of invasive FFR itself which was established as a gold standard after the outcome-based FAME family studies, but also, the clinical validation of the instantaneous wave-free ratio (iFR) with non-inferior outcomes compared to FFR in randomized trials despite the variability in agreement in observational studies, angiography-based FFR could move one step forward by showing equivalence in the clinical laboratory," they conclude. "Only then could functional angiography find a niche as a combined anatomic and physiologic modality in the catheterization laboratory."

The study was sponsored by CathWorks, Inc. Fearon disclosed institutional grant/research support from CathWorks, Abbott, and Medtronic; consultant fees/honoraria from Boston Scientific; and minor stock options with HeartFlow. Gibson disclosed grant/research support and consultant fees/honoraria/speakers bureau participation for multiple companies. Lansky reported nothing to disclose. Kern reported fees/honoraria/speakers bureau participation for multiple companies. Papafaklis previously developed his own model of cardiovascular pathophysiology but has no commercial conflicts to declare.

Transcatheter Cardiovascular Therapeutics (TCT) 2018. Presented September 24, 2018.

Circulation. Published online September 23, 2018. Article, Editorial


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.