Highlights
- The REPEAT-QFR study found a modest diagnostic agreement (72%) between in-procedure and core laboratory QFR analyses within the FAVOR III Europe trial.
- Suboptimal angiographic quality and poor adherence to Standard Operating Procedures (SOP) were identified as the primary drivers of measurement variability.
- Complex clinical and anatomical factors, including high SYNTAX scores and diabetes mellitus, independently predict increased discrepancy in physiological assessments.
- The findings suggest that the clinical outcomes of the FAVOR III Europe trial may have been influenced by operator-dependent analytical quality rather than software limitations alone.
Introduction: The Evolution of Functional Lesion Assessment
The transition from purely anatomical assessment of coronary artery disease to functional evaluation has been a cornerstone of modern interventional cardiology. Fractional Flow Reserve (FFR) remains the clinical gold standard; however, its requirement for invasive pressure wires, hyperemic agents, and the associated costs and procedural time have limited its universal adoption. In response, angiography-based alternatives like the Quantitative Flow Ratio (QFR) have emerged, offering a non-invasive estimation of FFR through computational fluid dynamics and 3D vessel reconstruction.
While initial validation studies showed high accuracy for QFR, the FAVOR III Europe trial recently raised critical questions regarding its safety and efficacy compared to traditional FFR. The REPEAT-QFR study was designed to investigate whether the unexpected clinical outcomes in FAVOR III Europe were attributable to the QFR technology itself or to variability in how the analysis was performed at the bedside during the procedure.
Study Design and Methodology
The REPEAT-QFR study analyzed data from 1,008 patients (comprising 1,233 vessels) who were randomized to the QFR-guided arm of the FAVOR III Europe trial. The study’s primary objective was to perform a head-to-head comparison between “in-procedure” QFR (performed by clinicians in the catheterization lab) and “core laboratory” QFR (performed by independent, blinded observers in a controlled setting).
To evaluate the quality of the in-procedure analyses, researchers developed a scoring system from 1 (very poor) to 5 (very good). This score was based on strict adherence to the Standard Operating Procedure (SOP), which includes criteria such as optimal vessel selection, accurate frame count for flow velocity estimation, and precise anatomical contouring. Two blinded core lab observers performed independent assessments to ensure the highest degree of reproducibility for the reference standard.
Key Findings: The Reality of Bedside Analysis
The results of the REPEAT-QFR study provide a sobering look at the challenges of implementing computational physiology in a real-world clinical environment. The median in-procedure QFR was recorded at 0.81, whereas the core laboratory QFR was 0.84.
Diagnostic Agreement and Correlation
While the mean difference between the two groups was small (0.02), the 95% limits of agreement were notably wide, ranging from -0.26 to 0.29. This indicates that in individual cases, the discrepancy could be large enough to change the clinical management strategy. The Spearman’s rank correlation coefficient was 0.58, and the overall diagnostic agreement—the frequency with which both methods agreed on whether a lesion was hemodynamically significant (QFR ≤0.80)—was 72%.
Predictors of Variability
The study identified that 64% of in-procedure analyses were rated as “good” or “very good” in terms of SOP adherence. However, 28% were only deemed “acceptable,” and 8% were classified as “poor” or “very poor.” The researchers found several key predictors that significantly increased the variability of the QFR measurements:
- Suboptimal angiographic quality: Issues such as poor contrast opacification, vessel overlap, or foreshortening hindered accurate 3D reconstruction.
- Poor in-procedure SOP adherence: Deviations from the standardized analysis steps led to less reliable results.
- High SYNTAX score: More complex coronary anatomy and higher atherosclerotic burden increased the difficulty of the analysis.
- Diabetes Mellitus: This clinical factor was independently associated with higher measurement discrepancy, potentially due to the presence of more diffuse microvascular disease or calcification.
Expert Commentary: Clinical Implications and Translational Insight
The findings of REPEAT-QFR offer a critical perspective for interventional cardiologists. The modest agreement observed between bedside clinicians and core lab specialists suggests that QFR is highly operator-dependent. Unlike pressure-wire FFR, which relies on a direct physical measurement, QFR is a software-based interpretation of visual data. Therefore, the quality of the input—the angiographic image—and the precision of the operator’s interaction with the software are paramount.
The Importance of Standardized Training
One of the most significant takeaways is the need for rigorous, standardized training for any clinician performing QFR. The 8% of cases rated as poor or very poor quality represent a clear target for improvement. In these cases, the functional assessment may lead to inappropriate revascularization or, conversely, the failure to treat a hemodynamically significant lesion. The REPEAT-QFR study demonstrates that simply having the software available is not enough; centers must implement ongoing quality assurance programs to maintain diagnostic integrity.
Technical Limitations vs. Human Error
It is important to distinguish between software limitations and human error. The fact that nearly 97% of the vessels were successfully analyzable in the core lab suggests that the QFR software is robust. The variability stems largely from how the software is utilized in the high-pressure environment of the catheterization lab. Future iterations of QFR technology may benefit from AI-driven automated contouring and real-time quality feedback to reduce this human-operator variability.
Conclusion and Future Directions
The REPEAT-QFR study provides essential context for the FAVOR III Europe trial results. It highlights that the success of angiography-based physiology depends on a “chain of quality” that starts with the acquisition of high-quality angiographic images and ends with a meticulous adherence to analytical protocols. To bridge the gap between in-procedure and core laboratory results, the medical community must prioritize clinician training and potentially integrate more automated features into the software to safeguard against common pitfalls.
As we move toward a more digital and less invasive era of cardiology, studies like REPEAT-QFR serve as a vital reminder that technology is a tool whose efficacy is inextricably linked to the expertise and diligence of the user.
Funding and ClinicalTrials.gov
The FAVOR III Europe trial and the REPEAT-QFR sub-study were supported by academic grants and institutional funding. The trial is registered at ClinicalTrials.gov with the identifier NCT03729518.
References
1. Kristensen SK, Holm MB, Maillard L, et al. Repeatability and quality assessment of QFR in the FAVOR III Europe trial: the REPEAT-QFR study. EuroIntervention. 2026;22(1):e53-e65.
2. Westra J, Andersen BK, Campo G, et al. Diagnostic Performance of Angiography-Derived Fractional Flow Reserve: A Systematic Review and Meta-Analysis. J Am Heart Assoc. 2018;7(14):e009453.
3. Tu S, Westra J, Yang J, et al. Diagnostic Accuracy of Angiography-Based Quantitative Flow Ratio for Online Assessment of Coronary Stenosis. JACC Cardiovasc Interv. 2016;9(19):2020-2030.