VERDICT/F1RST Confirms Modest Correlation Between IVUS, MLA and FFR

MIAMI BEACH, FLA.—Minimum luminal area (MLA) correlates with fractional flow reserve (FFR) values indicative of functionally ischemic lesions, but at varying cut-offs in different vessels, according to pooled results of the VERDICT/F1RST trial presented at TCT 2012.

Both VERDICT and F1RST were prospective, multicenter evaluations studying the correlation between IVUS measurements and FFR in intermediate lesions, according to TCT Course Director Gregg W. Stone, MD, of Columbia University Medical Center in New York.

VERDICT

The VERDICT pilot was a prospective, multicenter, non-randomized study in up to 300 intermediate coronary lesions or lesions with a visual diameter stenosis of 40% to 80% and a reference vessel diameter of 2.75 mm to 4.0 mm.

FFR and virtual histology IVUS (VH-IVUS) were used to assess all lesions.

Pooled data

Underway at the same time as the VERDICT trial was the F1RST study, also a multicenter, prospective study. It was conducted at 10 sites in four countries and included 350 patients (369 lesions) with the goal of studying the correlation between IVUS and FFR.

“The major differences between the VERDICT pilot and F1RST trial is that F1RST went down to a 2.5-mm immediate lesion instead of 2.75 mm in VERDICT, and they did not exclude patients with wall motion abnormalities,” Stone said.

Therefore, Stone said he teamed up with Ron Waksman, MD, of the Medstar Washington Hospital Center, Washington, DC, who led the F1RST trial, to pool the data. Using that pooled data, a MLA vs. FFR regression plot indicated a rough correlation between FFR and IVUS, but it was not significant. The r value was 0.34, C-statistic was 0.66, and the overall diagnostic accuracy was 66%.

Looking at false positives and true negatives on the MLA vs. FFR regression plot, the positive predictive value was 46.7% and the negative predictive value was lower than previous studies at 81%.

Stone said that the optimal MLA cutoff varied with vessel characteristics, and the MLA value with the strongest and most accurate correlation to an FFR of 0.8 was an MLA of less than 2.9 mm2.

 “Left anterior descending (LAD) arteries were much more likely to have a positive FFR than were non LADs. In fact, circumflex arteries were extremely unlikely to have a positive FFR,” Stone said.

Based on logistic regression analysis, several lesion types correlated with a positive FFR at different MLA cutoff levels (see Figure). Younger age was also predictive of a positive FFR, Stone added.

The overall correlation between MLA and FFR ≤ 0.80 was 41.2% in LADs, 23.9% in right coronary arteries and 12.7% in circumflex arteries. “This is striking,” Stone said. “Only one in eight circumflex arteries were positive for an FFR, despite the fact that the QCA DS was identical at 48% between the three different vessels.”

Study details

The VERDICT pilot was led by the Cardiovascular Research Foundation and enrolled 291 patients (312 lesions) at 15 hospitals in six countries. The trial included patients who only had one lesion per major vessel, and multiple vessels in the same patient could be enrolled.