Abstract
BACKGROUND: This comparative observational study evaluated the reliability of 3-dimensional (3D) models versus radiograph imaging for distal radius fracture classification by measuring inter- and intraobserver agreement. METHODS: Twelve distal radius fractures were selected from our institution and 3D printed models were created using computed tomography (CT) data imaging. Each fracture model was printed in both resin and filament material. 3D models and radiograph images were randomized, and participants independently classified each based on the AO classification for distal radius fractures. Inter- and intraobserver agreement was measured by kappa coefficient and reliability was compared between radiograph and 3D model for each fracture using 95% confidence intervals. Results were analyzed by 3-level classification AO types (A, B, C) and 9-level classification using AO types and groups (A1-A3, B1-B3, C1-C3). RESULTS: Among 24 participants, 3D printed models showed improved inter-observer agreement when compared with radiographs using the 3-level AO classification system, although not statistically significant. Significant improvement in agreement was observed using the 3-level AO classification amongst the orthopedic providers in this study. Agreement in all above-mentioned groups improved from "none/slight" with radiograph images to "fair" or "moderate" agreement when using the 3D models. CONCLUSIONS: 3D-printing enhanced inter- and intrasurgeon reliability for distal radius fracture classification when compared with radiograph imaging alone. However, overall agreement with either method of classification was low. Statistical significance was achieved only with subgroup analysis. These findings suggest that while 3D-printed models offer some value, their impact on clinical practice may be limited in their current form.