Coronary CT Angiography-derived Fractional Flow Reserve In Myocardial Bridging

PART I Diagnostic Performance of Coronary CT Angiography-derived Fractional Flow Reserve in Detecting Lesion-specific Ischemia in Myocardial BridgingPurpose:To investigate the diagnostic performance of coronary CT ang:iography-derived fractional flow reserve(CT-FFR)to detect lesion-specific ischemia in myocardial bridging(MB)using invasive FFR as reference standard.Methods:Patients with CCTA-confirmed MB of the left anterior descending coronary artery were retrospectively included from eight centers.All patients underwent CCTA and were subsequently referred to invasive coronary angiography(ICA)and FFR within 60 days of CCTA.MB location,length,depth,muscle index,and stenosis rate were measured on CCTA.MB was classified as either superficial(?2mm)or deep(>2mm)according to the depth of MB,short(<30cm)or long(>30cm)according to the length of MB.All MB vessels were further divided into<50%,50%-69%and ?70%groups according to proximal lumen stenosis on invasive coronary angiography(ICA).MB vessels with proximal plaques were divided into no calcification,low calcification and high calcification groups according to the degree of coronary calcification.CT-FFR values were measured at 2-4 cm distal to the stenosis in all MB vessels by cFFR software.An invasive FFR value<0.80 was considered to be ischemia-specific.Diagnostic performance and receiver operating characteristics(ROC)of CT-FFR to detect lesion-specific ischemia was assessed on a per-vessel level,using invasive FFR as reference standard.Pearson's correlation analysis and Bland-Altman plots were used for agreement measurement.Results:Of 104 MBs,89 were classified as superficial and 15 were classified as deep.Fity-six were classified as short and 48 were classified as long.Forty-eight MB vessels(46.2%)and 55 MB vessels(52.9%)showed ischemia by invasive FFR and CT-FFR,respectively.Sensitivity,specificity,accuracy,positive predictive value(PPV)and negative predictive value(NPV)of CT-FFR to detect lesion-specific ischemia were 0.96(0.85-0.99),0.84(0.71-0.92),and 0.89(0.81-0.94),0.84(0.71-0.92),and 0.96(0.85-0.99),respectively,in all MB vessels.There were no differences in diagnostic performance between superficial and deep MB groups,or between short and long MB groups(all P>0.05).The PPV of CT-FFR in 50%-69%stenosis group is lower than that in>70%stenosis group[0.59(0.33-0.81)vs.0.97(0.84-1.00),P<0.001],respectively.Bland-Altman analysis showed a mild systematic underestimation of CT-FFR when compared to invasive FFR(mean difference=0.014,95%LoA:-0.117-0.145).The correlation coefficient was r=0.798(P<0.001).Conclusion:CT-FFR demonstrated high diagnostic performance for identifying lesion-specific ischemia in vessels with MB when compared to invasive FFR.Part ? The Role of Coronary CT Angiography-Derived Fractional Flow Reserve in Assessing the Hemodynamic Relevance of Myocardial BridgingPurpose:To evaluate the feasibility of fractional flow reserve-derived from coronary CT angiography(CT-FFR)in patients with myocardial bridging(MB),its relationship with MB anatomical features,and clinical relevance.Methods:Patients without obstructive coronary artery disease but with CCTA-confirmed MB of the left anterior descending coronary artery(LAD)and patients with negative CCTA findings as control group were retrospectively included in this study.Patients with MB were divided into 2 groups(superficial and deep MB group)according to the depth of MB.Age and sex were matched among three groups.MB location,length,depth,muscle index,stenosis rate and systolic compression index were measured.CT-FFR values were measured at three points(segments 1-2 cm proximal to a MB,mid-tunneled segment and segments 1-2cm distal to the MB)by cFFR software.Patients with CT-FFR<0.80 were deemed to have hemodynamic relevance(abnormal group).x2 test,ANOVA test,Mann-Whitney U test,Kruskal-Wallis H test and logistic regression model were used for statistical analysis.Factors associated with abnormal CT-FFR values(?0.80)were analyzed.Results:A total of 120 patients with LAD MB and 41 patients as control group were included in this study.Of 120 MBs,59 were classified as superficial and 61 were classified as deep.The CT-FFR values decreased from diastolic phase to systolic phase in deep MB group[0.91(0.83-0.95)vs.0.94(0.91-0.96),P=0.003].Compared to control group,the CT-FFR values decreased in both diastolic phase and systolic phase in superficial MB group as well as deep MB group[systole0.92(0.90-0.94)control vs.0.82(0.71-0.89)superficial vs 0.72(0.60-0.89)deep,P<0.001;diastole 0.93(0,90-0.94)control vs.0.84(0.69-0.92)superficial vs 0.82(0.68-0.90)deep,P<0.001].A significant CT-FFR difference was only found in the MB segment during systole between superficial(0.94,0.90-0.96)and deep MB(0.91,0.83-0.95)(p=0.018).Abnormal CT-FFR values were found in 69(57.5%)MB patients(29[49.2%]superficial vs.40[65.6%]deep;p=0.069).MB length(OR=1.06,95%Cl:1.03-1.10;p=0.001)and systolic stenosis(OR=1.04,95%Cl:1.01-1.07;p=0.021)were the main predictors for abnormal CT-FFR,with an area under the curve of0.774(95%CI:0.689-0.858;p<0.001),MB patients with abnormal CT-FFR reported more typical angina(18.8%vs3.9%,p=0.023)than patients with normal values.Conclusion:MB patients showed lower CT-FFR values than controls.MB length and systolic stenosis demonstrate moderate predictive value for an abnormal CT-FFR value.Abnormal CT-FFR values have a positive association with symptoms of typical angina.Part ? Machine Learning Using Coronary CT Angiography-Derived Fractional Flow Reserve Predicts Proximal Atherosclerotic Plaque Formation Associated with Myocardial BridgingPurpose:To investigate the role of fractional flow reserve derived from coronary CT angiography(CT-FFR)for predicting proximal atherosclerotic plaque formation associated with myocardial bridging(MB)in the left anterior descending artery(LAD)using machine learning(ML)approaches.Methods:This retrospective study included patients with LAD MB without LAD plaque at baseline who underwent repeated coronary CT angiography(CCTA)studies with at least three months interval.Patient demographics morphological and CT-FFR features of LAD MB were recorded and analyzed.MB location,length,depth,muscle index and stenosis rate were measured.CT-FFR values were measured at segments 1 cm proximal to MB and segments 2-4 cm distal to the MB by cFFR software,the difference between the two is recorded as ACT-FFR.Proximal atherosclerotic plaque formation of LAD MB was the clinical endpoint.A ML-based prediction model was implemented in the least absolute shrinkage and selection operator(LASSO)algorithms.Receiver operating characteristic(ROC)curve analysis was performed to analyze the role of features in predicting proximal plaque formation of LAD MB.Results:This retrospective study included 188 patients with LAD MB.Of 188 patients,49(26.1%)had proximal atherosclerotic plaque formation associated with LAD MB at a median follow-up of 3.2 years(range 0.3-9.4 years).At baseline,lower CT-FFR values[0.84(0.65-0.92)vs.0.90(0.81-0.94)](p=0.005)and higher ACT-FFR[0.12(0.07-0.34)vs.0.08(0.04-0.18)](p=0.005)were found in LAD MB groups with plaque formation than LAD MB groups without plaque formation.The three highest LASSO coefficient's absolution values for predicting plaque formation were for CT-FFR(1.987),ACT-FFR(0.991)and MB depth(0.236).Features selected by ML showed a higher area under the curve(AUC)(0.7510.04)compared to clinical features(0.53±0.09,p<0.0001),morphological features(0.59±0.06,p=0.0025),and CT-FFR features(0.62±0.06,p=0.0127)alone.Conclusion:CT-FFR is a strong predictor for coronary plaque formation associated with LAD MB.A combination of all features selected by ML algorithms improves prediction of plaque formation.