ABSTRACT
<p><b>OBJECTIVE</b>This meta-analysis evaluated the accuracy of diagnosing coronary artery disease using 64-section spiral computed tomography, and compared the difference between Chinese studies and abroad studies.</p><p><b>METHODS</b>Relevant English and Chinese articles published from 1998 to 2009 were searched in Cochrane library, Medline, Embase database, OVID database and CNKI. Heterogeneity was tested, pooled weighted sensitivity and specificity and the corresponding 95%CI were calculated. Summary receiver operating characteristic (SROC) curve was drawn and the area under the curve was calculated, differences between studies from China and abroad were compared.</p><p><b>RESULTS</b>A total of 433 articles were searched and 108 articles were included (46 English articles and 62 Chinese articles) after excluding articles of research purposes or design does not match. Because of no gold standard, no blind, can not be calculated literature data, 7 and 20 (P > 0.05), 44 and 6 (P < 0.05), 3 and 1 (P < 0.05) Chinese studies and English articles respectively were excluded. Twenty-seven articles fulfilled all inclusion criteria (8 Chinese and 19 foreign studies) In 8 Chinese studies the pooled weighted sensitivity and specificity and area under SROC curve was 0.892 (95%CI: 0.868 - 0.913), 0.972 (95%CI: 0.966 - 0.977) and 0.983 (95%CI: 0.966 - 1.000) at segment-based analysis. In 19 foreign studies, the pooled weighted sensitivity and specificity and area under SROC curve was 0.971(95%CI: 0.957 - 0.982), 0.878 (95%CI: 0.852 - 0.902) and 0.973 (95%CI: 0.958 - 0.989) at patient-based analysis, 0.917 (95%CI: 0.895 - 0.936), 0.919 (95%CI: 0.909 - 0.928) and 0.974 (95%CI: 0.964 - 0.984) at vessel-based analysis, 0.882 (95%CI: 0.868 - 0.895), 0.959 (95%CI: 0.956 - 0.962) and 0.985 (95%CI: 0.978 - 0.992) at segment-based analysis. Pooled weighted pecificity of 64-section spiral CT angiography at segment-based analysis has significant different between home and abroad (P < 0.05).</p><p><b>CONCLUSIONS</b>Meta-analysis showed that noninvasive 64-section spiral computed tomography could correctly diagnose coronary artery disease with high sensitivity and specificity. Quality of related studies performed in abroad is significantly higher than those performed in China.</p>