The optimal treatment for different types of displaced partial articular radial head fractures is still debatable. Fractures involving the articulating portion of the radial head can be defined also as bi-articular and are often underestimated. Moreover, the complete loss of periosteal contact between fracture fragments is a marker of instability. Available classifications do not emphasize these aspects. The purpose of the present study is to describe two assessment methods to identify fractures involving the articulating portion of the radial head and complete loss of periosteal contact with the help of 2D-3D CT scan. The second purpose is to propose a classification of the displaced partial articular radial head fractures basing on these two assessment methods.

We hypothesize that the proposed classification is reliable.

By observing the position of the bicipital tuberosity with respect to the ulna in reference to the coronal plane in the 3D-CT scan and by observing the location of the fracture fragment in the 2D axial scan, it is possible to understand if a displaced partial articular fracture involves the articulating portion of the radial head. Also, it is possible to understand the presence or absence of a complete loss of periosteal contact between the two fracture fragments by observing the coronal and sagittal 2D-CT scans. We identified 20 displaced partial articular radial head fractures on a series of 149 patients with radial head fractures. Following the above-described assessment methods, these 20 fractures were classified in three subgroups. Also, seven evaluators were asked to use these assessment methods to classify these 20 fractures in the three subgroups, in order to evaluate inter- and intra-observer agreement.

Eight fractures involved the non-articulating portion of the radial head with complete loss of periosteal contact, eight involved the non-articulating portion of the radial head without complete loss of periosteal contact, and four involved the articulating portion of the radial head with or without complete loss of periosteal contact. The kappa for intra-observer reliability ranged from 0.46 to 0.84. The average kappa for inter-observer reliability was 0.570 (range: 0.526 to 0.676). The Kendall's coefficient for inter-observer concordance was 0.673.

Prognostically, displaced partial fractures of the articulating portion of the radial head could differ from the other types, regardless of whether or not there is a complete loss of periosteal contact. Underestimating this fracture pattern can lead to poor results due to risk of forearm rotation blockage. Moreover, underestimating complete loss of periosteal contact in displaced partial fractures of the non-articulating portion of the radial head could lead to poor results. The described evaluation methods have moderate reliability, but can represent, along with other described methods, a good starting point to better understand and treat these insidious fractures.

III; retrospective study.