DeFACTO Results Encouraging for 'Virtual' FFR Despite Missing Target

Reed Miller

August 26, 2012

August 26, 2012 (Munich, Germany) — Noninvasive fractional flow reserve calculated from computed tomography angiography (FFRCT) did not achieve the prespecified target for diagnostic accuracy in the Determination of Fractional Flow Reserve by Anatomic Computed Tomographic Angiography (DeFACTO) trial [1]. Nevertheless, investigators are optimistic that FFRCT can become an important tool for efficiently identifying high-grade stenoses and determining the hemodynamic significance of lesions.

As reported by heartwire, the 285-patient, 17-centerDeFACTO study--the first major trial of FFRCT--compared the novel modality with CT alone. Results were presented here today at the European Society of Cardiology 2012 Congress during a hot-line session by Dr James Min (Cedars-Sinai Medical Center, Los Angeles, CA). Results of the study are also published online in the Journal of the American Medical Association.

All patients in the study underwent CT, invasive coronary angiography (ICA), invasive FFR, and FFRCT to investigate suspected or known coronary disease. All scans were evaluated in a blinded fashion at independent core laboratories. The accuracy of FFRCT was compared with that of CT alone for the diagnosis of ischemia, defined by an FFR or FFRCT of <0.80. Anatomically obstructive coronary disease was defined as a stenosis of >50% on CT and ICA. A little over half of study subjects (n=137) had an abnormal invasive FFR. Primary study outcome was whether FFRCT plus CT improved the per-patient diagnostic accuracy such that the lower boundary of a one-sided 95% confidence interval of the estimate exceeded 70%.

Per Patient Diagnostic Accuracy of FFRCT in DeFACTO

Performance measure

Estimate, %

95% CI

Diagnostic accuracy









Positive predictive value



Negative predictive value



FFRCT improved the discrimination capability of CT. Compared with obstructive coronary disease diagnosed using CT alone (area under the receiver operating characteristic curve [AUC]: 0.68 vs 0.81; p<0.001).

FFRCT: Potential Advantages Over Existing Modalities

In the published paper, Min et al explain that FFRCT performed especially well compared with CT alone in patients with lesions of intermediate-stenosis severity, "who represent a particularly challenging clinical subset among whom angiographic severity is an often ambiguous metric for ischemia diagnosis."

The improvement in diagnostic accuracy provided by FFRCT in this group was seen in sensitivity rather than specificity. "These performance characteristics suggest a low false-negative rate if assessments by FFRCT were used to identify ischemia-causing intermediate lesions, with negligible effects on reductions of false-positive results," Min et al explain. "In this regard, the use of FFRCT may significantly advance clinical assessment of patients without conventional measures of anatomic high-grade coronary stenosis, largely by proper identification of a significantly greater proportion of patients with manifest ischemia rather than as a safeguard to further invasive evaluation."

A major potential advantage of FFRCT is its potential to do the job of two tests with only one scan. "Some have advocated for hybrid imaging with physiologic and anatomic evaluation of [coronary artery disease] CAD by stress testing and CT, respectively. However, this approach requires two tests and is associated with higher costs and greater per-patient radiation burden," Min et al write. "The addition of FFRCT to CT may allow for combined anatomic–physiologic assessment of CAD from performance of a single imaging test in a manner that may promote salutary outcomes. Future studies to assess the clinical and cost-effectiveness of such an approach now appear warranted."

In an accompanying editorial [2], Dr Manesh Patel (Duke University, Durham, NC) writes that FFRCT "represents a novel and important innovation, with the possibility not only to diagnose but also to help direct invasive treatment."

Because FFRCT demonstrated only "modest" specificity in DeFACTO, "at first glance, readers of the study may consider FFRCT technology to be limited based on the results presented," Patel acknowledges. "However, this would be a naive conclusion, likely based on the published diagnostic performance of noninvasive tests compared only with invasive angiography.

"If the existing noninvasive imaging technologies were compared with invasive angiography plus FFR, it is highly likely that the published diagnostic performance would be reduced," he points out.

The sole use of invasive angiography for lesion evaluation has decreased in clinical practice. In real-world practice, current noninvasive technologies for evaluating stable patients with suspected coronary disease prior to invasive angiography rarely perform at the levels seen in clinical trials, as demonstrated by the low rates of obstructive CAD found using elective catheterization, notes Patel.

"Hence, the current report describes an important noninvasive technology that may improve existing care and has the potential to outperform established noninvasive technologies," writes Patel.

Min reports research support from GE Healthcare and Philips Medical. Disclosures for the coauthors are listed in the paper. Patel reports consultancy for Bayer, Jansen, Baxter, and Otsuka, and grants from Johnson & Johnson, and AstraZeneca.