1Universitätsklinikum Leipzig Klinik und Poliklinik für Kardiologie Leipzig, Deutschland
Abstract
Background: Angiography-based functional assessment remains the state of the art to assess the hemodynamic relevance of intermediate coronary stenoses. In addition to fractional flow reserve (FFR), adenosin free indices like instantaneous wave-free ratio (iFR) or resting full-cycle ratio (RFR) are widely used for the invasive evaluation. Although these methods are widely accepted, they are underused in clinical routine due to costs, their invasive nature, longer procedural times, and lack of availability. Vessel-fractional flow reserve (vFFR) and quantitative flow ratio (QFR) are new software-based methods for the evaluation of coronary-physiology. The aim of this study is to compare vFFR and QFR with the established invasive methods iFR or RFR.
Methods: We prospectively analyzed 95 Patients in our QFR-Registry vFFR-a nd QFR-computation was performed three-times by two certified, blinded, and independent experts using the QAngio XA 3D 3.2- and CAAS-software, respectively.
Results: 101 main epicardial coronary arteries (52 LAD; 19 LCx; 30 RCA) were analyzed using vFFR, QFR and iFR or RFR. vFFR showed a good correlation with adenosine-free indices in all three measurements (AUC 0.883 with r=0.672, p<0,001; AUC 0.841 with r=0.618 p<0,001; AUC 0.840 with r=0.592, p<0,001 respectively). The diagnostic accuracy between the three vFFR-measurements and iFR/RFR was 83%; 81% and 79% respectively. QFR-computations showed a similar correlation with iFR/RFR (AUC 0.899 with r=0.720, p<0,001; AUC 0.919 with r=0,649, p<0,001; AUC 0.930 with r=0,703, p<0,001 respectively) with a diagnostic accuracy 88%, 89% and 88% respectively. The intra- and inter-observer reliability for vFFR (intra: mean difference 0.002±0.0568; inter: 0.024±0.054) and QFR (intra: mean difference 0.018±0.071; inter: 0.016±0.075) analyzed in a Bland-Altman Blot showed a good correlation between virtual assessments and invasive measurements with iFR/RFR.
|
|
n |
Sensitivity |
Specificity |
PPV |
NPV |
Diagnostic accuracy |
|
vFFR1 Total |
101 |
0.8696 |
0.8 |
0.7843 |
0.88 |
0.8317 |
|
LAD |
52 |
0.8387 |
0.7619 |
0.8387 |
0.7619 |
0.8077 |
|
LCX |
19 |
0.8333 |
1 |
1 |
0.9286 |
0.9474 |
|
RCA |
30 |
1 |
0.7143 |
0.6 |
1 |
0.8 |
|
vFFR2 Total |
101 |
0.8913 |
0.7455 |
0.7455 |
0.8913 |
0.8119 |
|
LAD |
52 |
0.871 |
0.7143 |
0.8182 |
0.7895 |
0.8077 |
|
LCX |
19 |
0.8333 |
0.9231 |
0.8333 |
0.9231 |
0.8947 |
|
RCA |
30 |
1 |
0.6667 |
0.5625 |
1 |
0.7667 |
|
vFFR3 Total |
101 |
0.8696 |
0.7273 |
0.7273 |
0.8696 |
0.7921 |
|
LAD |
52 |
0.8387 |
0.6190 |
0.7647 |
0.7222 |
0.75 |
|
LCX |
19 |
0.8333 |
0.9231 |
0.8333 |
0.9231 |
0.8947 |
|
RCA |
30 |
1 |
0.7143 |
0.6 |
1 |
0.8 |
|
Tab.1: Diagnostic performance of vFFR. | ||||||
|
|
n |
Sensitivity |
Specificity |
PPV |
NPV |
Diagnostic agreement |
|
QFR1 Total |
101 |
0.9130 |
0.8727 |
0.8571 |
0.9231 |
0.8911 |
|
LAD |
52 |
0.9355 |
0.9048 |
0.9355 |
0.9048 |
0.9231 |
|
LCX |
19 |
0.6667 |
1 |
1 |
0.8667 |
0.8947 |
|
RCA |
30 |
1 |
0.7619 |
0.6429 |
1 |
0.8333 |
|
QFR2 Total |
101 |
0.9348 |
0.8364 |
0.8269 |
0.9388 |
0.8812 |
|
LAD |
52 |
0.9355 |
0.8571 |
0.9062 |
0.9 |
0.9038 |
|
LCX |
19 |
0.8333 |
1 |
1 |
0.9286 |
0.9474 |
|
RCA |
30 |
1 |
0.7143 |
0.6 |
1 |
0.8 |
|
QFR3 Total |
101 |
0.9348 |
0.8364 |
0.8269 |
0.9388 |
0.8812 |
|
LAD |
52 |
0.9355 |
0.8571 |
0.9062 |
0.9 |
0.9038 |
|
LCX |
19 |
0.8333 |
1 |
1 |
0.9286 |
0.9474 |
|
RCA |
30 |
1 |
0.7143 |
0.6 |
1 |
0.8 |
|
Tab.2: Diagnostic performance of QFR. | ||||||
Fig.1:
Inter- and intra-observer-reliability vFFR and QFR.
Fig.2:
ROC-curves of vFFR and QFR.
Conclusion: The data show a good diagnostic agreement between vFFR and QFR and adenosine free invasive assessments. vFFR and QFR appear reasonable alternative approaches for the evaluation of intermediate stenoses. These data provide the rational to prospectively test for a clinical non-inferiority of the non-invasive flow measurements.