Skeletal dosimetry based on microCT images of trabecular bone has recently been introduced to calculate the red bone marrow (RBM) and the bone surface cell (BSC) equivalent doses in human phantoms for external exposure to photons. In order to use the microCT images for skeletal dosimetry, spongiosa voxels in the skeletons were replaced at run time by so-called micromatrices, which have exactly the size of a spongiosa voxel and contain segmented trabecular bone and marrow micro-voxels. A cluster (=parallelepiped) of 2 x 2 x 2 = 8 micromatrices was used systematically and periodically throughout the spongiosa volume during the radiation transport calculation. Systematic means that when a particle leaves a spongiosa voxel to enter into a neighboring spongiosa voxel, then the next micromatrix in the cluster will be used. Periodical means that if the particle travels through more than two spongiosa voxels in a row, then the cluster will be repeated. Based on the bone samples available at the time, clusters of up to 3 x 3 x 3 = 27 micromatrices were studied. While for a given trabecular bone volume fraction the whole-body RBM equivalent dose showed converging results for cluster sizes between 8 and 27 micromatrices, this was not the case for the BSC equivalent dose. The BSC equivalent dose seemed to be very sensitive to the number, form, and thickness of the trabeculae. In addition, the cluster size and/or the microvoxel resolution were considered to be possible causes for the differences observed. In order to resolve this problem, this study used a bone sample large enough to extract clusters containing up to 8 x 8 x 8 = 512 micro-matrices and which was scanned with two different voxel resolutions. Taking into account a recent proposal, this investigation also calculated the BSC equivalent dose on medullary surfaces of cortical bone in the arm and leg bones. The results showed (1) that different voxel resolutions have no effect on the RBM equivalent dose but do influence the BSC equivalent dose due to voxel effects by up to 5% for incident photon energies up to 200 keV, (2) that the whole-body BSC equivalent dose calculated with a cluster with 2 x 2 x 2 = 8 micromatrices is consistent with results received with clusters of up to 8 x 8 x 8 = 512 micromatrices, and (3) that for external whole-body exposure the inclusion of the BSC on medullary surfaces of cortical bone has a negligible effect on the whole-body BSC equivalent dose.