Purpose To investigate whether the dose-scoring process of Monte Carlo (MC) simulations of Gold nanoparticles (GNPs) in radiation therapy affects the results. Methods The GATE MC toolkit was used to simulate the irradiation of a water phantom containing a single solid or hollow GNP with 250 kVp and 6 MV photons. The dose was scored in 20 nm × 20 nm × 50 μm, 100 nm × 100 nm × 50 μm and 200 nm × 200 nm × 50 μm volumes using dose-scoring voxels of size 1 nm × 1 nm × 50 μm, 10 nm × 10 nm × 50 μm, 50 nm × 50 nm × 50 μm and 100 nm × 100 nm × 50 μm Εxcess dose depth-dose (EDDD) curves and lateral beam profiles were used to compare the dose-scoring voxels. Results In a given volume, neither the EDDD curves nor the lateral beam profiles are affected by the size of the dose-scoring voxels, subject to noise and uncertainty. Certain features of the EDDD curves are clearly seen in larger volumes, but hidden within the uncertainty and noise levels in smaller volumes. For the lateral beam profiles, it is the larger volumes that result in misleading results and the smaller ones that give the expected results. However, the limited statistics result in asymmetries and skewness in the profiles. Conclusion For a given volume, the dose curves are not affected by the size of the dose-scoring voxels. However, the voxel size may hide or reveal the finer structure of the dose curves and/or may result in misleading curves.