DISCOVER-FLOW: CT-Derived FFR Measurements Show High Diagnostic Accuracy

PARIS, France—A noninvasive method that calculates fractional flow reserve (FFR) based on coronary computed tomographic angiography (CTA) offers a way to physiologically assess coronary artery disease without some of the drawbacks inherent in both conventional imaging techniques, according to new data presented Wednesday, May 18 at EuroPCR 2011.

Presenter Bon-Kwon Koo, MD, PhD, of Seoul National University Hospital (Seoul, South Korea), noted that while coronary CTA provides accurate anatomical information, the imaging technology cannot predict the functional significance of lesions and its high rate of false positive diagnoses may lead to unnecessary invasive procedures. Conventional FFR, while the 'gold standard' for identifying ischemia-causing lesions, also is an invasive test.

Diagnostic Accuracy Evaluated

In the DISCOVER-FLOW trial, Dr. Koo and colleagues tested an experimental method of noninvasive FFR based on standard CTA images that does not require any additional radiation, imaging procedures, or medications. The method creates a 3-dimensional coronary model “and then—with the help of some rather complex modeling procedures and a computer—coronary blood flow, elasticity, or pressure can be simulated” and used to calculate FFR, he explained. The calculations mimic the hyperemia that occurs after adenosine administration in conventional FFR.

The prospective, multicenter study included 103 patients (n = 159 vessels) who had stenosis in a major epicardial coronary artery as well as diagnostic quality coronary CTA obtained from a 64-detector row or higher CT scanner. Most lesions were located in the left anterior descending artery (54.7%), with the remainder in the right coronary (19.5%) and left circumflex arteries (25.8%). Obstructive CAD was defined as the presence of at least 50% diameter stenosis on CTA, while lesion-specific ischemia was defined as measurements of 0.80 or below on both conventional and noninvasive FFR.

CT-derived FFR results correlated well with those of invasive FFR (R = 0.72; P < 0.001), though there was some scattering. Compared with the invasive approach, CTA had a false positive rate of 38% and noninvasive FFR had a much lower false positive rate of 11%.

Moreover, on a per-vessel analysis, the diagnostic accuracy of noninvasive FFR exceeded that of CTA. Although the latter method offered excellent sensitivity and negative predictive value, the newer modality provided better specificity and positive predictive value (table 1). Similar results were found on a per-patient analysis.

Table 1. Diagnostic Performance of Noninvasive FFR and Coronary CTA: Per-Vessel Analysis

^a Primary endpoint.

Based on the results, “[t]his noninvasive 'all-in-one' technology may reduce unnecessary invasive coronary angiography and revascularization procedures,” Dr. Koo concluded.

Meant to Supplement, Not Replace Other Imaging

Following the presentation, Ron Waksman, MD, of Washington Hospital Center (Washington, DC), asked 2 questions: whether noninvasive FFR could be derived from conventional angiography and why there was some lack of correlation between the 2 FFR methods.

“Theoretically, it's possible” to calculate FFR based on angiography, Dr. Koo replied. “But the problem with coronary angiography is that it [produces] 2-dimensional images, so unless it can give very detailed 3-dimensional anatomical structures, I think the results will not be as good as with the CT images.”

Dr. Koo agreed there were discrepancies between the noninvasive and invasive FFR methods, though he added that this was most apparent at lower FFR values. “At the range of 0.75 to 1, scatter was very narrow, but the scatter was [increased] at lower levels. I think that's natural, because in this simulation we cannot account for collateral flow,” he noted. “From a clinical standpoint, we don't care if an FFR is 0.3 or 0.4. Once the FFR is 0.75 or 0.8, that's the point when we make a clinical decision to intervene or not.”

In addition, Dr. Koo pointed out that CT measurements, which form the basis of noninvasive FFR, are imperfect and therefore still can lead to false positives and negatives. “Although the range is not huge, there [also] is the possibility of discrepancies between hypothesized simulations and real human coronary situations, so I don't want to say that this technology will replace FFR and will replace all the invasive and noninvasive technologies,” he concluded. “The most important thing is that the adoption of this technology will improve the diagnostic accuracy of CT scans. . . . That's what I want to stress.”

 

---

Source:
Koo B-K. Diagnosis of ischemia-causing stenoses obtained via noninvasive fractional flow reserve (DISCOVER-FLOW): A prospective multicentre first-in-man study. Presented at: EuroPCR; May 18, 2011; Paris, France.

 

 

Related Stories:

• FFR Gains Popularity, but Not All Clinicians Are Converts
• FFR Guidance for Questionable Lesions Dramatically Cuts Stenting vs. IVUS
• IVUS Cannot Replace FFR But Can Help Rule Out Ischemia-Producing Stenoses