The Novel Impact of Augmented Reality and 3D Printing in the Diagnosis of Complex Acetabular Fractures: A Comparative Randomized Study in Orthopedic Residents

Abstract

Augmented reality (AR) and 3D printing (3DP) are novel technologies in the orthopedic field. Over the past decade, enthusiasm for these new digital applications has driven new perspectives in improving diagnostic accuracy and sensitivity in the field of traumatology. Currently, however, it is still difficult to quantify their value and impact in the medical–scientific field, especially in the improvement of diagnostics in complex fractures. Acetabular fractures have always been a challenge in orthopedics, due to their volumetric complexity and low diagnostic reliability. Background/Objectives: The goal of this study was to determine whether these methods could improve the learning aspect and diagnostic accuracy of complex acetabular fractures compared to gold-standard CT (computed tomography). Methods: Orthopedic residents of our department were selected and divided into Junior (JUN) and Senior (SEN) groups. Associated fractures of acetabulum were included in the study, and details of these were provided as CT scans, 3DP models, and AR models displayed on a tablet screen. In a double-blind questionnaire, each resident classified every fracture. Diagnostic accuracy (DA), response time (RT), agreement (R), and confidence (C) were measured. Results: Twenty residents (JUN = 10, SEN = 10) classified five fractures. Overall DA was 26% (CT), 18% (3DP), and 29% (AR). AR-DA was superior to 3DP-DA (p = 0.048). DA means (JUN vs. SEN, respectively): CT-DA was 20% vs. 32% (p < 0.05), 3DP-DA was 12% vs. 24% (p = 0.08), and AR-DA was 28% vs. 30% (p = 0.80). Overall RT was 61.2 s (±24.6) for CT, 35.8 s (±20.1) for 3DP, and 46.7 s (±20.8) for AR. R was fairly poor between methods and groups. Overall, 3DPs had superior C (65%). Conclusions: AR had the same overall DA as CT, independent of experience, 3DP had minor differences in DA and R, but it was the fastest method and the one in which there was the most confidence. Intra- and inter-observer R between methods remained very poor in residents.