Abstract

Open reduction and internal fixation of pelvic acetabular fractures are challenging due to the limited surgical exposure from surrounding abdominal tissue. There have been a number of recent trials using metallic 3D-printed pelvic fracture plates to simplify and improve various elements of these fracture fixation surgeries; however, the amount of time and accuracy involved in the design and implantation of customised plates have not been well characterised. This study recorded the amount of time related to the design, manufacture and implantation of six customised fracture plates for five cadaveric pelvic specimens with acetabular fracture, while manufacturing, and surgical accuracy was calculated from computed tomography imaging. Five of the fracture plates were designed within 9.5 h, while the plate for a pelvis with a pre-existing fracture plate took considerably longer (20.2 h). Manufacturing comprised 3D-printing the plates in Ti6Al4V with a sintered laser melting (SLM) 3D-printer and post-processing (heat treatment, smoothing, tapping threads). The manufacturing times varied from 27.0 to 32.5 h, with longer times related to machining a thread for locking-head screws with a multi-axis computer numerical control (CNC) mill. For the surface of the plate in contact with the bone, the root-mean-square errors of the print varied from 0.10 to 0.49 mm. The upper range of these errors was likely the result of plate designs that were relatively long with thin cross-sections, a combination that gives rise to high thermal stresses when using a SLM 3D-printer. A number of approaches were explored to control the trajectories of locking or non-locking head screws including guides, printed threads or hand-taps; however, the plate with CNC-machined threads was clearly the most accurate with screw angulation errors of 2.77° (range 1.05–6.34°). The implanted position of the plates was determined visually; however, the limited surgical exposure and lack of intra-operative fluoroscopy in the laboratory led to high inaccuracies (translational errors of 1.74–13.00 mm). Plate mal-positioning would lead to increased risk of surgical injury due to misplaced screws; hence, it is recommended that technologies that can control plate positioning such as fluoroscopy or alignment guides need to be implemented into customised plate design and implantation workflow. Due to the plate misalignment and the severe nature of some acetabular fractures comprising numerous small bone fragments, the acetabular reduction exceeded the clinical limit of 2 mm for three pelvises. Although our results indicate that customised plates are unsuitable for acetabular fractures comprising six or more fragments, confirmation of this finding with a greater number of specimens is recommended. The times, accuracy and suggested improvements in the current study may be used to guide future workflows aimed at producing customised pelvic fracture plates for greater numbers of patients.