From the Freezer to Implant: Does Cadaver Freezing Affect the Accuracy of 3D-Printed Scaphoid Prosthesis Models?

BACKGROUND: Cadaveric specimens used in research are commonly frozen when stored and may undergo multiple freeze-thaw cycles during experimental procedures. To streamline workflows and minimize potential tissue degradation from repeated thawing, this study evaluated whether a frozen cadaver wrist can be reliably used to model a scaphoid for the development of a patient-specific prosthesis design method in a preclinical setting. Additionally, we assessed the impact of segmentation smoothing techniques on the accuracy of the prosthesis model. METHODS: High-resolution computed tomography (CT) scans were performed on a cadaver wrist in both frozen and thawed states. Scaphoid bones were segmented using two approaches: a tight (native surface) segmentation and a smoothened version optimized for prosthetic articulation. The resulting 3D models were registered, and volume and shape differences between frozen and thawed states, as well as between segmentation methods, were quantified. RESULTS: No statistically significant volume differences were observed between scaphoid models segmented from frozen and thawed conditions (p = 0.46). The average difference was 0.14% (SD 0.55). Furthermore, smoothing segmentation had minimal impact on the overall dimensions of the scaphoid model. CONCLUSIONS: Frozen cadaver wrists can be used to accurately model scaphoid bones for patient-specific prosthesis design without introducing significant volumetric deviations. Segmentation smoothing, which is necessary for prosthesis fabrication, does not compromise anatomical accuracy, supporting the feasibility of using frozen specimens for preclinical modeling.