Lattice structures have become an innovative solution for the improvement of part design, as they are able to substitute solid regions, maintain mechanical capabilities, and reduce material usage; however, dimensional quality control of these geometries is challenging. X-ray computed tomography (XCT) is the most suitable non-destructive metrological technique as it is capable of characterizing internal features and hidden elements. Uncertainty estimation of XCT is still in development, and studies typically use high-resolution calibrated devices such as focal variation microscopes (FVMs) as a reference, focusing on certain parts of the lattice but not the whole structure. In this paper, an estimation of the accuracy of XCT evaluation of a complete lattice structure in comparison to a higher-resolution reference device (FVM) is presented. Experimental measurements are taken on ad hoc designed test objects manufactured in polyamide 12 (PA12) using selective laser sintering (SLS), optimized for the evaluation on both instruments using different cubic-based lattice typologies. The results confirm higher precision on XCT evaluation in both qualitative and quantitative analysis. Even with a lower resolution, XCT is able to characterize details of the surface such as re-entrant features; as well, standard deviations and uncertainties in strut diameter evaluation remain more stable in all cells in XCT, identifying on the other hand reconstruction problems on FVM measurements. Moreover, it is shown that, using XCT, no additional evaluation errors were found in inner cells, suggesting that the measurement of external elements could be representative of the whole structure for metrological purposes.