| Objective To compare the differences between the K software of D company and the S software of S company in measuring CT coronary flow reserve fraction(CT-FFR),and to explore the difference and correlation influencing factors.between the two softwares in measuring CT coronary flow reserve fraction(CT-FFR).Methods Seventy-eight patients(234 vessels)who completed coronary CT angiography(CCTA)were retrospectively analyzed and transferred to professional CT-FFR measurement software in DICOM format.Bland-Altman diagram and paired sample t test were used to analyze the two The consistency of software measurement was analyzed,and the influence of each factor on the difference was analyzed by Pearson,Spearman correlation test and independent sample t test,and the factors related to the difference of CT-FFR value were analyzed by Logistic univariate and multivariate regression.By drawing receiver operating characteristic(ROC)curve,the size and optimal cut off point of measurement consistency under different influencing factors of the two softwares were inferred.Results Among the 78 patients,6 patients successfully completed the S software analysis.During the K software analysis process,the group could not be calculated due to the existence of serious artifacts in the images.Finally,72 patients(216 vessels)were included.Consistency test between K and S software,the Bland-Altman difference plot and ratio plot both show that an average of 6%(14/216)points are outside the 95% concordance limit.The paired sample t test results show that K and S There was a significant difference between software measurements(P<0.05);univariate analysis of differences between different software showed that heart rate variability,vascular dislocation offset distance,subjective image quality and lumen diameter were the two software-measured CT-FFR.The influencing factors of difference(P<0.05);ROC curve showed that heart rate variability,vascular dislocation offset distance,subjective image quality score and lumen diameter optimal cut off points were 20 beats/min,0.85 mm,3.5 minute and2.95 mm respectively,and the area under the curve(AUG)were 0.718,0.899,0.843 and 0.568 respectively;univariate and multivariate logistic analysis showed that vascular dislocation offset distance,subjective image quality score and heart rate variability,were the predictors of the difference in CT-FFR measured by the two software(P < 0.01);The Kappa consistency test showed that in the measurement of CT-FFR,the three phases were all P<0.01,and the optimal diastolic phase and the75% phase showed excellent agreement(Kappa value 0.794);the remaining two groups were consistent The performance was better(Kappa values were 0.682 and0.675,respectively).Conclusion(1)There are differences in CT-FFR measurement of different manufacturers and different software,and even qualitative diagnosis errors occur.(2)The heart rate variability,the subjective image quality score and the degree of vascular dislocation distance are the independent influencing factors of the difference between the two CT-FFR software.When the heart rate variability is less than 20beats/min,the subjective image quality score is greater than or equal to 3.5points,and the vascular dislocation offset distance is less than 0.85 mm,it is more conducive to ensure the stability of CT-FFR measurement with different software,and to provide reliable information for clinical intervention decisions. |
PDF Full Text Request