Factor analysis, a new method of functional imaging, has been applied to cardiovascular nuclear medicine. Because of the difficulty of its interpretation, it has not been popular as a method for detecting abnormal wall motion. The purpose of this study was to evaluate the usefulness of factor analysis in exercise gated blood-pool study in patients with ischemic heart disease. In our factor analysis, left ventricular region of interest (LVROI) was extracted to exclude the surrounding radioactivities. The new method was compared with 1) the conventional factor analysis using whole region (whole ROI method), and with 2) the other functional images, i.e. stroke volume, ejection fraction and phase images. At first we tried 3-factor analysis of the LVROI method, which resulted in many uninterpretable factors. Whereas in 2-factor analysis no uninterpretable factors were extracted. In comparison with cine-mode display, the LVROI method with 2-factor analysis showed the best sensitivity (85%) and specificity (100%). In exercise gated blood-pool study, it became easier to detect abnormal wall motion by comparing the factor image at exercise with resting image. In conclusion, the 2-factor analysis using the LVROI method greatly improved the limitation of conventional factor analysis, and will be useful in detecting wall motion abnormality in patients with ischemic heart disease.