An evaluation of the systematic accuracy of magnetic resonance (MR)-guided Gamma Knife surgery was performed. In two experiments, a cylinder phantom filled with dosimeter gel containing ferrous sulfate was fixed to a stereotactic frame. The gel phantom was irradiated with the Gamma Knife with a single shot using 4-mm collimators. The target point was set at the frame center of the stereotactic system giving coordinate values of X = 100, Y = 100, Z = 100. The maximum target dose was 15 Gy. MR imaging was undertaken immediately after the irradiation, using a superconductive 1.5-T MR scanner. Spin echo T1-weighted images, with transaxial, coronal, and sagittal views, were obtained. On the images, points with the highest signals were defined as the target point which received the maximum dose. Within the dose range of the experiment, this definition is based on a linear relationship between the dose to the gel and the T1 relaxation shortening after irradiation. The distances between the frame center and the target point defined on the MR images in the experiments were 0.12 mm (0.2375 pixels) and 0.43 mm (0.8515 pixels), respectively. Both are within the mechanical accuracy of the Gamma Knife. The imaging study confirms the accuracy of the Gamma Knife surgery used in the institution.