The accuracy of myocardial perfusion imaging using quantitative analysis for patients with coronary artery disease: a Chinese hospital experience

Objective To evaluate the accuracy of automated quantification of myocardial perfusion imaging (MPI) using a method based on a Western normal database for the detection of coronary artery disease (CAD) in a group of Chinese patients. Methods Seventy-two Chinese patients who underwent coronary angiography (CAG) and MPI within 3 months were recruited into this study. Eighty selected from 140 Chinese patients with low probability of CAD ( ＜ 5% ) were enrolled into local normal database of 99Tcm-methoxyisobutylisonitrile (MIBI) MPI using Cedars quantitative perfusion SPECT (QPS) database. Two Western MPI normal databases (CSMC MibiMbiAuto and Mibimibi) were used for processing the Chinese CAD patients recruited in this study, and the results were compared with those using local normal database and visual interpretation. T-test and z-test were used for statistical analysis. Results The extent (EXT)measurement obtained from Mibimibi and local database was ( 10.73 ± 14.54)% and ( 14.22 ± 16.51 )%,respectively ( t = 7.87, P ＜ 0.001 ); the severity (SEV) was 1.07 ± 0.93 and 1.34 ± 1.20, respectively ( t =7.45, P＜0.001). The area under curve(AUC) by using EXT measurement for local database (0.85 ±0.05) was larger than that for CSMC MibiMbiAuto ( AUC = 0.72 ± 0.06, z = 2.50, P ＜ 0.01 ) and Mibimibi ( AUC = 0.77 ± 0.06, z = 2.47, P = 0.014). The AUC of local database showed no significant difference from that of visual interpretation (AUC=0.83 ±0.05, z=0.05, P＞0.05). Conclusion Quantification of MPI of our Chinese patients using Western normal database would decrease the accuracy for the detection of CAD. Key words: Coronary disease;  China;  Tomography, emission-computed, single-photon;  MIBI