The available scores to clinically evaluate fracture consolidation encounter difficulties to interpret progression towards consolidation in long-bone non-union, particularly when incorporating biomaterials in the surgical treatment. The aims of this study were to validate the REBORNE bone healing scale in tibia, humerus and femur non-unions treated by a combination of mesenchymal stromal cells (MSCs) and biomaterials, through the interclass correlation (ICC) among raters, and to define reliability and concordance in anteroposterior and lateral radiographs, compared to computed tomography (CT).
Methods: Twenty-six cases from the EudraCT 2011-005441-13 clinical trial underwent bone healing evaluation, if at least 3 out of 4 cortical views clearly identified. Three senior orthopaedic surgeons evaluated radiographs and CTs at 3 and 6 months FU. All cases included preoperative imaging and radiographs at 12 months. The 4-stage scale score was obtained from each cortical view in orthogonal radiographs or CTs. A score of 0.6875 (11/16) was set as a threshold for bone healing. Statistically, ICC evaluated agreement among raters. Cronbach's alpha coefficient tested reliability. Lin's concordance correlation coefficients (CCC) were estimated between mean CT scores and mean radiographic scores. Bland and Altman graphs provided the limits of agreement between both imaging techniques. Sensitivity and specificity were assessed in radiographs (against CT), and the Area Under the Receiver Operating Characteristics (ROC) Curve was estimated. The probability to predict bone consolidation with REBORNE scores obtained from radiographs was modelled.
Results: An ICC of 0.88 and 0.91 (CT and radiographs) confirmed agreement in the REBORNE score for non-union bone healing, with an inter-rater reliability of 0.92 and 0.95. Scores through the radiographic evaluation were found equivalent to the CTs at 6 months FU. A CCC of 0.79 was detected against CT. The radiographic scores in the REBORNE bone healing scale correctly classified bone consolidation in 77%, with an accuracy of 83% based on ROC curves.
Conclusions: The REBORNE score measured with CT or radiographic images was reliable among raters at a follow-up time above 6 months for long bone non-union fractures. The REBORNE scale measured with radiographs proved valid to assess consolidation against CT measurements.
Keywords: Bioceramics; Internal consistency; MSC; Non-union; Radiographic; Reliability; Score; Treatment; Validation.
Copyright © 2020. Published by Elsevier Ltd.