Accuracy of the Surface Contour of Three-Dimensional-Printed Canine Pelvic Replicas

The aim of this study was to determine the differences in surface contour between models of native pelvic bones and their corresponding three-dimensional (3D)-printed replicas.

Digital 3D models of five cadaveric hemipelves and five live dogs with contralateral pelvic fractures were generated based on computed tomographic images and 3D printed. The 3D-printed replicas underwent 3D scanning and digital 3D models of the replicas were created. The digital 3D model of each replica was superimposed onto the model of the native hemipelvis. Errors in the replicas were determined by comparing the distances of 120,000 corresponding surface points between models. The medial surface, lateral surface and dorsal surface of the acetabulum (DSA) of each hemipelvis were selected for further analysis. The root mean square error (RMSE) was compared between various selected areas using a one-way repeated measures analysis of variance, followed by a Bonferroni post-hoc test.

The RMSE of the hemipelvis was 0.25 ± 0.05 mm. The RMSE significantly decreased from the medial surface (0.28 ± 0.06mm), to the lateral surface (0.23 ± 0.06mm), to the DSA (0.04 ± 0.02mm) (p < 0.001).

The 3D-printed replicas were adequate in serving as a template for the pre-contouring of bone plates in fracture repair of pelvic fractures, particularly those that demand accurate reduction such as acetabular fractures.