Open AccessCoronary CT Angiography–derived Fractional Flow Reserve: Machine Learning Algorithm versus Computational Fluid Dynamics Modeling
Author(s) -
Christian Tesche,
Carlo N. De Cecco,
Stefan Baumann,
Matthias Renker,
Tindal W. McLaurin,
Taylor M. Duguay,
Richard R. Bayer,
Daniel Steinberg,
Katharine L. Grant,
Christian Canstein,
Chris Schwemmer,
Max Schöebinger,
Lucian Itu,
Saikiran Rapaka,
Puneet Sharma,
U. Joseph Schoepf
Publication year - 2018
Publication title -
radiology
Language(s) - Uncategorized
Resource type - Journals
SCImago Journal Rank - 3.118
H-Index - 295
eISSN - 1527-1315
pISSN - 0033-8419
DOI - 10.1148/radiol.2018171291
Subject(s) - fractional flow reserve , medicine , angiography , receiver operating characteristic , lesion , coronary angiography , radiology , stenosis , coronary artery disease , myocardial infarction , surgery
Purpose To compare two technical approaches for determination of coronary computed tomography (CT) angiography-derived fractional flow reserve (FFR)-FFR derived from coronary CT angiography based on computational fluid dynamics (hereafter, FFR CFD ) and FFR derived from coronary CT angiography based on machine learning algorithm (hereafter, FFR ML )-against coronary CT angiography and quantitative coronary angiography (QCA). Materials and Methods A total of 85 patients (mean age, 62 years ± 11 [standard deviation]; 62% men) who had undergone coronary CT angiography followed by invasive FFR were included in this single-center retrospective study. FFR values were derived on-site from coronary CT angiography data sets by using both FFR CFD and FFR ML . The performance of both techniques for detecting lesion-specific ischemia was compared against visual stenosis grading at coronary CT angiography, QCA, and invasive FFR as the reference standard. Results On a per-lesion and per-patient level, FFR ML showed a sensitivity of 79% and 90% and a specificity of 94% and 95%, respectively, for detecting lesion-specific ischemia. Meanwhile, FFR CFD resulted in a sensitivity of 79% and 89% and a specificity of 93% and 93%, respectively, on a per-lesion and per-patient basis (P = .86 and P = .92). On a per-lesion level, the area under the receiver operating characteristics curve (AUC) of 0.89 for FFR ML and 0.89 for FFR CFD showed significantly higher discriminatory power for detecting lesion-specific ischemia compared with that of coronary CT angiography (AUC, 0.61) and QCA (AUC, 0.69) (all P < .0001). Also, on a per-patient level, FFR ML (AUC, 0.91) and FFR CFD (AUC, 0.91) performed significantly better than did coronary CT angiography (AUC, 0.65) and QCA (AUC, 0.68) (all P < .0001). Processing time for FFR ML was significantly shorter compared with that of FFR CFD (40.5 minutes ± 6.3 vs 43.4 minutes ± 7.1; P = .042). Conclusion The FFR ML algorithm performs equally in detecting lesion-specific ischemia when compared with the FFR CFD approach. Both methods outperform accuracy of coronary CT angiography and QCA in the detection of flow-limiting stenosis.