Fig 1.
Sites of bone formation and resorption during healing of a stable distal radius fracture.
3D HR-pQCT images of a stable, simple distal radius fracture at 9 days (baseline) and at follow-up 26, 44 and 87 days post-fracture showing regions of bone formation (green), resorption (red) and no change (blue). Images were obtained by superposition of the follow-up images over the baseline image. Cortical fracture locations (indicated by arrows) are bridged at 87 days post-fracture.
Fig 2.
Subvolumes and corresponding correlation coefficients.
Average correlation coefficients for registration of volumes with different sizes, using different step-sizes after different applied translations in x-, y- and z-direction and after different applied rotations around the x-, y- and z-axis.
Fig 3.
2D HR-pQCT slices with contours (left) and 3D model (right) of the fragments in two stable fractures (top and middle) and a fracture with secondary displacement (bottom) that were registered in this study. The volume [mm3] per fragment as well as of the complete fracture is shown between brackets.
Fig 4.
Overlay images of three fracture cases.
Overlay image after standard registration of the whole bone region in the follow-up image (3–4 weeks post-fracture) onto baseline (1–2 weeks post-fracture) with corresponding correlation coefficients (top). Overlapping regions (purple) are shown as well as regions belonging to baseline (red) and follow-up (green). To check whether the per-fragment registration improved the results compared to the standard registration, overlay images and corresponding correlation coefficients are presented for the same subregions obtained after standard and per-fragment registration. Arrows indicate locations of clear improvement after per-fragment registration. Correlation coefficients, which are reported below each fragment, are higher after per-fragment registration than after standard registration (Wilcoxon signed-rank test, p = 0.005).