Inter- and intra-observer reliability of injury diagnosis for peri-knee fractures: A comparison between two- and three-dimensional CT imaging

Abstract

Objective
This study aimed to assess whether three-dimensional (3D) CT imaging improves the inter- and intra-observer reliability of peri-knee fracture classifications, compared to two-dimensional (2D) CT imaging.
Methods
A retrospective analysis was conducted on 23 patients with peri-knee fractures, using both 2D and 3D-CT scans. Three radiologists classified distal femur, patella, and tibial plateau fractures according to Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association (AO/OTA) and Schatzker systems. Reliability was measured using Cohen’s kappa, with evaluations conducted at two separate intervals to assess intra- and inter-observer consistency.
Results
The intra-observer reliability for 2D-CT was substantial for distal femur (κ = 0.737, IQR 0.615–0.788) and tibial plateau (κ = 0.732, IQR 0.615–0.819) fractures, improving slightly with 3D-CT (κ = 0.775, IQR 0.658–0.869; κ = 0.768, IQR 0.628–0.882 respectively). Patella fracture classification showed almost perfect reliability (κ = 0.823, IQR 0.707–0.882) with 2D-CT, further improving with 3D-CT (κ = 0.865, IQR 0.764–0.951). However, inter-observer reliability showed no significant improvement with the addition of 3D-CT across all fracture types.
Conclusion
While 3D-CT marginally enhances intra-observer reliability for peri-knee fractures, the difference in inter-observer reliability compared to 2D-CT was not statistically significant.