We showed scanner dependence of brain (18)F-FDG and (11)C-PiB images by using phantom examination with nine kinds of positron emission tomography (PET) scanners. We used two types of phantoms, cylindrical phantom with 15 cm inside diameter and three-dimensional (3D) brain phantom, and we set the body phantom on the bed to examine the effect of scatter and random coefficients from outside of the axial field of view (AFOV). Radioactivity and distance of the two phantoms were determined by a pilot study to obtain a condition similar to the clinical study. Axial uniformity was evaluated by circular region of interest (ROI) of 12 cm diameter, set in the center of the reconstruction image of the cylindrical phantom. As a result, the standardized uptake value (SUV) was lower than the true value in some scanners, and there was a scanner in which the axial uniformity was deteriorated by high radioactivity outside the AFOV. In the cylindrical phantom, the axial uniformity of the scanner was improved using the new dead-time correction method; however, it was not improved in the 3D brain phantom. Quality-controlled PET scanners are important to maintain constant levels for multicenter studies.