The Clinical Research Of Dual-source Ct Coronary Artery Imaging

Part I Diagnostic Accuracy of Dual-source Computed Tomography Coronary Angiography for the Detection of Coronary Stenosis SeverityPurpose:To determine the diagnostic accuracy of dual-source computed tomography coronary angiography (DSCT-CA) to detect various stenosis degrees of coronary arteries in comparison with conventional coronary angiography (CCA).Materials and Methods:Eighty patients with suspected coronary artery disease (CAD) who had undergone DSCT-CA and CCA from February2008and December2008were enrolled. All of DSCT-CA examinations were earlier than CCA within an interval of less than30days. Stenosis (percentage) was obtained from detected lumen contours at the minimal lumen level. Coronary artery stenosis of more than50%was defined as significant obstruction and more than70%stenosis was considered to need a therapy of implanting a stent in the corresponding segment. Sensitivities, specificities, positive predict value (PPV), negative predict value (NPV), accuracy with patient-based, vessel-based and segment-based methods were calculated respectively using stenosis of50%and70%as cutoff points.Results:The prevalence of CAD was78.7%(63/80). DSCT-CA detected65significant obstruction patients with three false positive (FP) cases and one false negative (FN) case. In the patient-based analysis, the sensitivity and specificity for detection of significant CAD were98.4%and82.4%. In the vessel-based analysis,327vessels were included into our study. One hundred and three vessels with at least stenosis more than50%were diagnosed by CCA, while DSCT-CA detected99significant obstructive vessels with10FP vessels and14FN vessels. The sensitivity and specificity of DSCT-CA for diagnose of stenosis of more than50%were86.4% and92.5%. In the segment-based analysis, a total of149among926coronary artery segments of more than50%were quantified stenosis by CCA. The sensitivity, specificity, PPV, NPV and accuracy of DSCT-CA were80.5%,98.2%,89.6%,96.3%, and95.4%respectively. The paired χ2tests revealed significant difference between DSCT-CA and CCA at the level of segment-based analysis (P=0.03). Fifty-two patients with at least one segment stenosis more than70%were diagnosed by CCA, while forty-eight patients were diagonsed by DSCT-CA with4TP cases and4TN cases. The sensitivity and specificity of DSCT-CA using70%as cutoff point with the patient-based method were92.3%and85.7%. The sensitivity and specificity with the vessel-based method were73.0%and93.9%. The sensitivity, specificity, PPV, NPV and accuracy with the segment-based method were69.1%,98.4%,83.3%,96.6%, and95.5%respectively. The paired χ2tests revealed significant difference between DSCT-CA and CCA at the level of segment-based analysis (P=0.02) using70%as a cutoff point.Conclusions:DSCT-CA can accurately determine the severity of different coronary artery stenosis at the level of patient-based and vessel-based analysis. However, the significant differences in the paired χ2tests at the level of segment-based analysis indicate that DSCT-CA cannot be used as a total alternative technique of CCA at present. Part Ⅱ Causes of Inconsistence between Dual-source CT Coronary Angiography and Conventional Coronary Angiography on Evaluation of Coronary Stenosis Purpose:To analyze the main causes of inconsistence between dual-source CT angiography (DSCT-CA) and convention coronary angiography (CCA) on evaluation of coronary stenosis.Materials and Methods:Head to head comparison was applied in80patients who were enrolled in Part I. Coronary stenosis was classified into three groups:mild stenosis (<50%), moderate stenosis (≥50%and<70%), and severe stenosis (≥70%). Segments classified into different groups between DSCT-CA and CCA were defined as inconsistence. Image quality of DSCT-CA for each segment was assessed on a three-points ranking scale:3, excellent (no artifacts);2, good (minor artifacts) and1, adequate (moderate artifacts) by two experienced radiologists.Results:Inconsistence was observed in sixty-three of926segments, including22overestimated segments and41underestimated segments. The main reasons of inconsistence included motion artifact due to fast heart rate (5segments), step-stair artifact due to respiratory motion or variability of heart rate (5segments), vein contamination (3segments), omitting of non-calcified plaques (12segments), severe calcification plaques (20segments), inappropriate measurement position (10segments), metal beam hardening due to pacemaker (2segments), contrast medial (2segments), small vessels (3segments), and large patient size (1segment). There are724segments from major coronary arteries (RCA, LCA and LCX) and202segments from secondary branches were compared. Fifty-five inconsistent segments in major coronary arteries and8inconsistent segments in secondary branches were observed separately. Further statistical analysis using χ2test did not show significant difference among the major coronary arteries of inconsistent segments (P=0.7). No significant difference was found among the different image qualities of inconsistent segments (P=0.4).Conclusions:Knowledge about the technical artifacts, reviewing the areas where lesions could be easy to be missed, and selecting appropriate measurement positions could be helpful to improve the consistence between DSCT-CA and CCA. Purpose:To investigate the role of diagnostic experience in diagnostic performance of Dual-source CT coronary angiography (DSCT-CA) for detection of coronary artery disease (CAD) and to determine the impact of plaque composition on quantification of coronary stenosis degree.Materials and Methods:A total of ninety-four patients underwent both DSCT-CA and conventional coronary angiography (CCA) from January2009to December2009. Sensitivity, specificity, positive predict value (PPV), negative predict value (NPV) and accuracy of DSCT-CA were calculated in comparison with CCA using50%stenosis as a cutoff point. Chi-square test was used to compare the segment-based sensitivity, specificity, PPV, NPV and accuracy between Part I and Part III. Plaques detected by DSCT-CA were classified as calcified, mixed and non-calcified plaques by visualization. Correlations of quantification of coronary stenosis between DSCT-CA and CCA among different plaque compositions were evaluated with linear regression. Altman-bland analysis was used to compare the difference between DSCT-CA and CCA among different plaque compositions.Results:A total of92.6%(87/94) patients was diagnosed CAD by CCA. The patient-based sensitivity and specificity of DSCT-CA for detection CAD were98.9%and57.1%. The vessel-based sensitivity and specificity of DSCT-CT were97.4%and91.1%. The segment-based sensitivity, specificity, PPV, NPV and accuracy were95.3%,97.2%,91.3%,98.5%, and96.7%respectively. There were no significant differences of sensitivity, specificity, PPV, NPV, and accuracy between Part I and Part III with the segment-based analysis. Three hundred and thirty one plaques were detected by DSCT-CA including170non-calcified plaques,66calcified plaques and95mixed plaques. The correlation coefficients between DSCT-CA and CCA in the lesions with calcified plaques were0.65(P<0.05), while the corresponding values in the lesions with non-calcified plaques were0.74(P<0.05). On Bland-Altman analysis for comparison of DSCT-CA and CCA results, the mean differences were2.7±18.3%in lesions with calcified plaques and-5.3±17.8%in lesions with non-calcified plaques.Conclusions:Although there is no significant difference due to rather small sample size, increasing experience with DSCT-CA improved the diagnostic performance of inexperienced radiologists in evaluation of coronary stenosis. The quantification of coronary stenosis by DSCT-CA had a poorer correlation in the lesions with calcified plaques than non-calcified plaques.