Author + information
- Nils P. Johnson, MD, MS∗ (, )@wfearonmd,
- Mitsuaki Matsumura, BS,
- Stephan Achenbach, MD,
- Thomas Engstrom, MD, PhD,
- Abid Assali, MD,
- Allen Jeremias, MD,
- Stephane Fournier, MD,
- Bernard De Bruyne, MD, PhD,
- Martin B. Leon, MD and
- William F. Fearon, MD
- ↵∗Weatherhead PET Center, McGovern Medical School at UTHealth, 6431 Fannin Street, Room MSB 4.256, Houston, Texas 77030
Invasive coronary physiology to measure fractional flow reserve (FFR) or nonhyperemic pressure ratios (NHPRs) requires vessel instrumentation using a pressure wire. Although modest, the pressure wire and its insertion carry some cost and risk. Deriving pressure loss from anatomy by simulating flow began over 40 years ago (1) and has progressed to routine clinical application via several technologies. In parallel, numerous NHPR algorithms have sought to avoid the small cost, increase in procedure time, and risk of vasodilator hyperemia necessary to measure FFR invasively (2).
Choosing between invasive NHPRs and noninvasive FFR derived from the angiogram (FFRangio) involves a tradeoff. On the one hand, NHPRs measure coronary pressure directly but assumes a fixed relationship between rest and hyperemia to predict FFR ≤0.80 (2). On the other hand, FFRangio customizes its analysis to the specific patient but only computes coronary pressure indirectly. It is uncertain if one approach agrees better with the reference standard, invasive FFR (3).
To address this unresolved question, we performed a post hoc (not pre-specified) analysis of the previously published FAST-FFR (FFRangio Accuracy vs. Standard FFR) trial (4). In brief, 10 international centers prospectively enrolled patients undergoing routine coronary angiography and invasive FFR assessment. Pressure tracings were saved and analyzed centrally at a blinded core laboratory for FFR and 3 NHPRs as previously described (2) averaged over 5 consecutive beats: instantaneous wave-free ratio (iFR), diastolic pressure ratio (dPR), and whole-cycle coronary or aortic pressure ratio (Pd/Pa). After routine invasive coronary angiography, each site processed at least 3 digital angiograms to compute FFRangio locally by an operator blinded to the invasive FFR value (4). Our primary endpoint was accuracy against invasive FFR ≤0.80, compared between FFRangio ≤0.80 and iFR ≤0.89, dPR ≤0.89, or Pd/Pa ≤0.91 using a McNemar test in R version 3.5.1 (R Foundation for Statistical Computing, Vienna, Austria). The primary FAST-FFR study was sponsored by CathWorks (Kfar-Saba, Israel), which provided additional funding to the physiology core laboratory for this investigator-initiated substudy.
From the 301 enrolled subjects and 319 lesions, a total of 265 (88%) subjects and 278 (87%) lesions remained after removing tracings rejected by the core lab for inadequate baseline recordings or inability to compute all 3 NHPR values. The median for each physiologic index was: FFR 0.84 (interquartile range [IQR]: 0.74 to 0.90), FFRangio 0.83 (IQR: 0.74 to 0.89), iFR 0.92 (IQR: 0.85 to 0.96), dPR 0.92 (IQR: 0.86 to 0.96), and Pd/Pa 0.93 (IQR: 0.89 to 0.97).
Binary accuracy against invasive FFR ≤0.80 was 92.4% for FFRangio, 85.3% for Pd/Pa, and 82.7% for iFR and dPR. Figure 1 provides an example showing an angiogram plus all physiology indexes. In all pairwise comparisons FFRangio demonstrated superior agreement: ratio of disagreements favoring FFRangio to disagreements favoring Pd/Pa was 2.4 (in 223 cases both FFRangio and Pd/Pa agreed with FFR, in 7 cases both FFRangio and Pd/Pa disagreed with FFR, in 34 cases FFRangio agreed with FFR but Pd/Pa disagreed with FFR, and in 14 cases Pd/Pa agreed with FFR but FFRangio disagreed with FFR, so the ratio equals 34/14 = 2.4; p = 0.006); and ratio of 2.8 (p < 0.001) versus both iFR and dPR. No difference existed among the different NHPRs for predicting invasive FFR ≤0.80: ratio 2.4 (p = 0.14) for Pd/Pa versus iFR; ratio 2.8 (p = 0.12) for Pd/Pa versus dPR; and ratio 1.0 (p = 1.0) for iFR versus dPR. No significant heterogeneity of FFRangio superiority existed for lesion location (left anterior descending vs. other) or clinical presentation (stable vs. acute). Additionally, these findings did not change materially after randomly selecting a single lesion from subjects who underwent multivessel assessment.
In conclusion, FFRangio agrees more often with invasive FFR than any of 3 NHPRs (2). Physiologically, our results imply that “simulated hyperemia” outperforms “assumed hyperemia” even when disadvantaged by an indirect computation of coronary pressure. Practically, our findings suggest that centers wishing to avoid invasive FFR for whatever logistical or clinical reasons would be better served by abandoning invasive pressure wires completely and instead employing noninvasive FFR derived from the angiogram, although clinical outcome data will be important to validate this approach.
Please note: Dr. Johnson has received internal funding from the Weatherhead PET Center for Preventing and Reversing Atherosclerosis; has an institutional licensing and consulting agreement with Boston Scientific for the smart minimum fractional flow reserve algorithm; has received significant institutional research support from St. Jude Medical (CONTRAST; NCT02184117) and Philips Volcano Corporation (DEFINE-FLOW, NCT02328820) for studies using intracoronary pressure and flow sensors; and has a patent pending on diagnostic methods for quantifying aortic stenosis and TAVI physiology. Dr. Engstrom has served as a consultant for Boston Scientific, Abbott, Bayer, Novo Nordisc, and AstraZeneca. Dr. Jeremias has received institutional funding (unrestricted education grant) from Philips/Volcano and Abbott Vascular; and has served as a consultant for Philips/Volcano, Abbott Vascular, Opsens Medical, and Boston Scientific. Dr. Fournier has served as a consultant for CathWorks. Dr. Fearon has received institutional research support from Abbott Vascular, Medtronic, and CathWorks; and owns minor stock options in HeartFlow.
- 2019 American College of Cardiology Foundation
- Brown B.G.,
- Bolson E.,
- Frimer M.,
- Dodge H.T.
- van 't Veer M.,
- Pijls N.H.J.,
- Hennigan B.,
- et al.
- ↵(2018) Correction. J Am Coll Cardiol 71:2279–2280. Correction. J Am Coll Cardiol 2018;71:2279-2280..
- Fearon W.F.,
- Achenbach S.,
- Engstrom T.,
- et al.