Dosimetry for targeted radionuclide therapy is necessary for treatment planning and radiation protection. Currently, there are no standard methods either for performing dosimetry or to evaluate the uncertainties inherent in the dosimetric calculations. In this paper, we present an experimental method using polymer gel dosimeters, whereby absorbed-dose distributions resulting from nonuniform distributions of activity may be determined directly from T(2) magnetic resonance imaging (MRI) as well as from scintigraphic images. A phantom containing a nonuniform distribution of I-131 was prepared by mixing 58 MBq of activity within the gel as it was solidifying. The resulting absorbed-dose distribution was determined directly from the MRI and from sequential single-photon emission computed tomography (SPECT) images using the Medical Internal Radiation Dose (MIRD) schema. The MRI data were quantified using 12 calibration vials uniformly irradiated by 0-12 MBq of I-131. The agreement between the two absorbed-dose maps was verified by convolving the MRI-based absorbed-dose map with the SPECT system point spread function, which gave a correlation coefficient of 0.96. It was seen that the absorbed-dose distribution, as imaged by the MRI, was misrepresented by the SPECT owing to its relatively poor spatial resolution, which included a shift of the voxel containing the maximum absorbed dose. This technique could provide an independent benchmark for assessing patient-specific dosimetry and, therefore, could be used as a basis for quality control for dosimetry.