Purpose: To evaluate the actual state of quality control in Korea through an analysis of mammographic phantom images obtained from a multicenter, and to determine the proper exposure conditions required in order to obtain satisfactory phantom images. Materials and Methods: Between April and June, 2002, 193 phantom images were referred to the Korea Food and Drug Administration for evaluation. Two radiologists recorded the number of fibers, specks and masses they contained, and the "pass" criteria were as follows: checked number of fibers: four or more; specks, three or more; masses, three or more (a total of ten or more features). Images in which optical density was over 1.2 were classified as satisfactory. In addition, changes in the success ratio, and difference between the two groups (i.e. "pass" and "fail", with regard to exposure conditions and optical density) were evaluated. Results: Among the 193 images, 116 (60.1%) passed and 77 (39.9%) failed. Among those which passed, 73 /100 (73%) involved the use of a grid, 80/117 (68.3%) were obtained within the optimal kVp range, 50/111 (45.0%) involved the use of optimal mAs, and 79/112 (70.5%) were obtained within the optimal range of optical density. Among those which failed, the corresponding figures were 17/52 (32.6%), 33/66 (50.0%), 31/69 (44.9%), and 35/65 (53.8%). There were statistically significant differences between the pass and fail rates, and with regard to kVp, optical density, and the use of a grid, but with regard to mAs, statistical differences were not significant. If only phantom images with an optical density of over 1.2 [as per the rule of the Mammographic Quality Standard Act (MQSA)] was included, the success rate would fall from 60.1% to 43.0%. Conclusion: The pass rate for mammographic phantom images was 60.1%. If such images are to be satisfactory, they should be obtained within the optimal range of optical density, using optimal kVp and a grid.