Reproducibility of quantitative flow ratio: the QREP study

Background: Quantitative flow ratio (QFR) is a tool for physiological lesion assessment based on invasive coronary angiography. Aims: We aimed to assess the reproducibility of QFR computed from the same angiograms as assessed by multiple observers from different, international sites. Methods: We included 50 patients previously enrolled in dedicated QFR studies. QFR was computed twice, one month apart by five blinded observers. The main analysis was the coefficient of variation (CV) as a measure of infra- and inter-observer reproducibility. Key secondary analysis was the identification of clinical and procedural characteristics predicting reproducibility. Results: The intra-observer CV ranged from 2.3% (1.5-2.8) to 10.2% (6.6-12.0) among the observers. The inter-observer CV was 9.4% (8.0-10.5). The QFR observer, low angiographic quality, and low fractional flow reserve (FFR) were independent predictors of a large absolute difference between repeated QFR measurements defined as a difference larger than the median difference (>0.03). Conclusions: The inter- and intra-observer reproducibility for QFR computed from the same angiograms ranged from high to poor among multiple observers from different sites with an average agreement of 0.01 +/--0.08 for repeated measurements. The reproducibility was dependent on the observer, angiographic quality and the coronary artery stenosis severity as assessed by FFR.