BACKGROUND: Myocardial perfusion imaging (MPI) has been used to estimate cardiac event risk. The aim of the present study is to achieve stable risk estimation based on perfusion scoring and a multi-center prognostic database. METHODS AND RESULTS: Multivariate logistic regression analysis was performed to estimate cardiac event risk based on a J-ACCESS study. A stress-MPI was performed in 45 patients with coronary artery disease (CAD) and in 25 non-CAD patients. Perfusion defect scoring of summed stress score (SSS) was performed by 5 methods: (1) visual scoring; (2) automatic scoring of 3 short-axis and 1 vertical long-axis slices; (3) visual modification of Method 2; (4) automatic polar map scoring based on a Japanese multi-center database; and (5) visual modification of Method 4. Agreement of SSS between 2 observers was good (r=0.87-0.97). Agreement of estimated cardiac event risk between observers and among 5 methods was very good (r=0.99-1.00). Regarding diagnostic accuracy for CAD, Method 5 showed optimal diagnostic yields (sensitivity 84%, accuracy 77%). CONCLUSIONS: Estimation of cardiac event risk in conjunction with polar map segmentation and common normal databases resulted in stable risk values, and might be used for risk stratification in patients suspected of having CAD.
BACKGROUND: Myocardial perfusion imaging (MPI) has been used to estimate cardiac event risk. The aim of the present study is to achieve stable risk estimation based on perfusion scoring and a multi-center prognostic database. METHODS AND RESULTS: Multivariate logistic regression analysis was performed to estimate cardiac event risk based on a J-ACCESS study. A stress-MPI was performed in 45 patients with coronary artery disease (CAD) and in 25 non-CAD patients. Perfusion defect scoring of summed stress score (SSS) was performed by 5 methods: (1) visual scoring; (2) automatic scoring of 3 short-axis and 1 vertical long-axis slices; (3) visual modification of Method 2; (4) automatic polar map scoring based on a Japanese multi-center database; and (5) visual modification of Method 4. Agreement of SSS between 2 observers was good (r=0.87-0.97). Agreement of estimated cardiac event risk between observers and among 5 methods was very good (r=0.99-1.00). Regarding diagnostic accuracy for CAD, Method 5 showed optimal diagnostic yields (sensitivity 84%, accuracy 77%). CONCLUSIONS: Estimation of cardiac event risk in conjunction with polar map segmentation and common normal databases resulted in stable risk values, and might be used for risk stratification in patients suspected of having CAD.