High-resolution peripheral quantitative computed tomography measures 3-dimensional bone micro-architecture at the distal radius and tibia, and provides new insight issues of bone quality in patients. The typical measurement region of interest (ROI) with an axial dimension of 9.02 mm is positioned at a fixed distance from landmarks at the distal end of the radius or tibia, but little is known about the variation of bone architecture in the vicinity of this region and related sensitivity to measurement location. Variation in ROI position can lead to differences in quantitative outcomes, and this is compounded when reference landmarks are difficult for an operator to discern. Therefore, the purpose of this study was to quantify the site-specific variation in bone micro-architecture at the human distal radius and tibia, and the impact of variation of the ROI position. Using cadavers, a 27.06-mm axial ROI including the typical clinical measurement site was scanned, and the variation of morphological (apparent density, trabecular density, trabecular number, and cortical thickness) and biomechanical (polar moment of inertia, cross-sectional area) indices were assessed along the 27.06 mm length. It was found that the radius has more site-specific variation than the tibia. Furthermore, the shifting of the 9.02 mm clinical ROI position by ±0.5 mm resulted in approximately a 2% and 6% error in the tibia and radius density, respectively. For longitudinal assessment of bone architecture, site-specific variation is not as critical because comparisons are made within an individual;​ however, site-specific variation is an important consideration for cross-section studies and assessment of individuals against a normative population database.