Selection of animal bone surrogate samples for orthopaedic screw testing based on human radius CT-derived bone morphology

• Published in: Medical engineering & physics Vol. 103; p. 103786
• Main Authors: Silva-Henao, J.D., Synek, A., Pahr, D.H., Reisinger, A.G.
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01-05-2022
• Subjects: Animals; Bone; Bone and Bones - diagnostic imaging; Bone and Bones - surgery; Bone morphology; Bone Screws; Computer tomography; Data visualization; Humans; Orthopaedic implants; Orthopedics; Radius - diagnostic imaging; Sample selection; Sheep; Swine; Tomography, X-Ray Computed; Bone; Computer tomography; Sample selection; Bone morphology; Data visualization; Orthopaedic implants
• ISSN: 1350-4533; 1873-4030
• DOI: 10.1016/j.medengphy.2022.103786

•Present a new methodology for bone sample selection for orthopaedic screw testing.•The method is based on local, CT derived, morphology from real implantation sites.•Regions with human-like bone morphologies were identified in in 2 different animal models, 12 bones in total.•Regions with human-like morphologies coincide among bones from different provenance.
Animal bones are commonly used to test the mechanical competence of bone screws since they are easier to obtain compared to human bones. Nevertheless, selecting an appropriate animal sample that correctly represents the human bone architecture where the screw is implanted is frequently overlooked. This study presents a protocol for bone sample selection for screw mechanical testing based on a characterization of the local CT-derived bone morphology. For this, 36 human radii were used to quantify the local peri-implant bone morphology of 360 osteosynthesis screws, 10 per bone, whose implantation site and depth were fully known. A cylindrical volume of interest was created along the screw path and used to measure the local morphology. With this, 10 average peri-implant bone morphologies were defined. Additionally, two animal models, pig, and sheep, were selected and used as potential sample sources. From each model, six bones were selected for analysis. Based on a surface mesh of each bone a computational algorithm was created to automatically extract cylindrical probes in several locations from which the local bone morphometry was calculated. A multi-parametric bone similarity score was developed and used to compare the local morphology of each animal bone to that of the human average peri-implant bone morphology. The score was then mapped to the surface of the bone thus allowing to visually identify regions on the animal bone with human-like bone morphology. By using this methodology, the use of human bones can be avoided since samples with human-like bone morphologies can be found on animal bones. This is not only useful in cases where strict ethical constrains must be fulfilled, but also in studies where the relationship between morphology and screw competence is to be studied, something that is hard to replicate with commercially available synthetic alternatives.