Purpose <p>In voxel-based dosimetry, dose-voxel kernels (DVKs) represent the absorbed dose in a target voxel per decay in a source voxel. In this work, we aimed to validate DVKs generated using Particle and Heavy Ion Transport Code System (PHITS) as a first step in establishing a quantitative SPECT/CT-based dosimetry framework for selective internal radiation therapy (SIRT).</p> Methods <p>A 1-MBq source of ⁹⁰Y, ¹⁵³Sm, ¹⁷⁷Lu, ¹⁸⁸Re, and ¹⁶⁶Ho located in central cubic voxels of 1&#xa0;mm, 3&#xa0;mm, and 6&#xa0;mm dimension within soft tissue (1.04&#xa0;g⋅cm⁻³), was simulated using 2.5 × 10⁷ histories in PHITS. The DVK results were compared to published data (Lanconelli et al., 2012). In a pilot study, 17 MBq of ¹⁵³Sm microspheres were injected into a liver tumor-bearing rat and imaged 24&#xa0;h later using SPECT/CT. SPECT/CT images were then converted to cumulated activity and convolved with a ¹⁵³Sm DVK using Fast Fourier Transform in MATLAB to create an absorbed dose map.</p> Results <p>As expected, DVKs for the radionuclides showed a steep decrease from the source voxel to the maximum continuous slowing down approximation (CSDA) of beta components. Beyond the CSDA, gamma and bremsstrahlung contributed minimally (10⁻⁶ to 10⁻⁴) to voxel doses. Observed DVK changes as voxel size increased were in good agreement with published data, with differences in source voxel ranging from 1 to 24%. Rat imaging revealed ¹⁵³Sm microspheres localized at injection sites, with mean and maximum doses of 45&#xa0;Gy and 194&#xa0;Gy, respectively.</p> Conclusion <p>The DVKs calculated in PHITS were successfully validated. ¹⁵³Sm dose maps were produced from SPECT/CT images in a rat model. Further studies on the application of DVKs in heterogeneous media will be conducted.</p>