The influence of 3D printing on inter- and intrarater reliability on the classification of tibial plateau fractures.

ABSTRACT

PURPOSE: Tibial plateau fractures continue to be a challenging task in clinical practice and current outcomes seem to provide the potential for further improvement. Especially presurgical understanding of the orientation of fracture lines and fracture severity is an essential key to sufficient surgical treatment. The object of this study was to evaluate the reliability of modern axial CT-based classification systems for tibial plateau fractures. In addition, the diagnostic-added value of 3D printing on the classification systems was investigated. METHODS: 22 raters were asked to classify 22 tibial plateau fractures (11 AO B- and 11 AO C-fractures) with the AO, the 10-Segment and the Revisited Schatzker classification in a three-step evaluation first only using CT scans, second with 3D volumetric reconstructions and last with 3D-printed fracture models. Inter- and intraobserver agreement and the subjective certainty were analyzed. Statistics were done using kappa values, percentage match and a univariant one-way analysis of variance. RESULTS: The AO classifications interobserver percentage match and kappa values improved for all raters and recorded an overall value of 0.34, respectively, 43% for the 3D print. The 10-Segment classification interobserver agreement also improved with the 3D-printed models and scored an overall kappa value of 0.18 and a percentage match of 79%. Equally the Revisited Schatzker classification increased its values to 0.31 and 35%. The intraobserver agreement showed a moderate agreement for the AO (0.44) and Revisited Schatzker classification (0.42) whereas the 10-Segment classification showed a fair agreement (0.27). Additionally, the raters changed their classification in 36% of the cases after evaluating the fracture with the 3D-printed models and the subjective certainty regarding the decisions improved as categories of self-reliant diagnostic choices were selected 18% (p < 0.05) more often after using the 3D-printed models. CONCLUSION: Based on the measured outcomes it was concluded that the new classification systems show an overall slight to fair reliability and the use of 3D printing proved to be beneficial for the preoperative diagnostics of tibial plateau fractures. The 10-Segment classification system showed the highest percentage match evaluation of all classification systems demonstrating its high clinical value across all levels of user experience.