TIBIAL PLATEAU FRACTURE CHARACTERISTICS: RELIABILITY AND DIAGNOSTIC ACCURACY OF 3D-PRINTED MODELS

Abstract

Classification systems for tibial plateau fractures suffer from poor interobserver agreement, and their value in preoperative assessment to guide surgical fixation strategies is limited. For tibial plateau fractures four major characteristics are identified: lateral split fragment, posteromedial fragment, anterior tubercle fragment, and central zone of comminution. These fracture characteristics support preoperative assessment of fractures and guide surgical decision-making as each specific component requires a respective fixation strategy. We aimed to evaluate the additional value of 3D-printed models for the identification of tibial plateau fracture characteristics in terms of the interobserver agreement on different fracture characteristics.Preoperative images of 40 patients were randomly selected. Nine trauma surgeons, eight senior and eight junior registrars indicated the presence or absence of four fracture characteristics with and without 3D-printed models. The Fleiss kappa was used to determine interobserver agreement for fracture classification and for interpretation, the Landis and Koch criteria were used.3D-printed models lead to a categorical improvement in interobserver agreement for three of four fracture characteristics: lateral split (Kconv = 0.445 versus K3Dprint = 0.620; P < 0.001), anterior tubercle fragment (Kconv = 0.288 versus K3Dprint = 0.449; P < 0.001) and zone of comminution (Kconv = 0.535 versus K3Dprint = 0.652; P < 0.001).The overall interobserver agreement improved for three of four fracture characteristics after the addition of 3D printed models. For two fracture characteristics, lateral split and zone of comminution, a substantial interobserver agreement was achieved.Fracture characteristics seem to be a more reliable way to assess tibial plateau fractures and one should consider including these in the preoperative assessment of tibial plateau fractures compared to the commonly used classification systems.