Reliability of Radiographic Union Scoring in Humeral Shaft Fractures.

Abstract

To establish the reliability of 2 radiographic union scoring systems for nonoperative humeral shaft fractures. Retrospective medical record review. Patients identified had humeral shaft fractures and radiographs at various follow-up time points, which were graded according to the both the standard (RUST) and modified radiographic union scoring systems (mRUST). A single North American Level-1 Trauma center in Connecticut, including emergency department and clinic follow-up visits. Forty-five adult patients (162 image sets) met the following inclusion criteria: diaphyseal humerus fracture, initial nonoperative management, and greater than 2 weeks of follow-up with imaging. All 162 image sets of anterior-posterior and lateral radiographs were scored and divided into 4 tiers based on increasing score. Anterior-posterior and lateral image sets were randomly selected from each tier for a total of 50 that were then scored by 7 different reviewers using both the RUST and mRUST systems. The intraclass correlation coefficients for the cortical and system scores for the RUST and mRUST systems. Interobserver reliability was 0.795 for the RUST system and 0.801 for mRUST. Intraobserver reliability was 0.909 for RUST and 0.949 for mRUST. For mRUST, 92% of values were within ± 1 point from each other. The RUST and mRUST systems can be applied to humeral shaft fractures with excellent reliability. They have the potential to assist in the diagnosis of humeral shaft union by providing an objective and standardized method to assess healing of bone over time.