Validation of the modified radiographic union score for tibia fractures (mRUST) in murine femoral fractures

Abstract

Bony union is a primary predictor of outcome after surgical fixation of long bone fractures. Murine models offer many advantages in assessing bony healing due to their low costs and small size. However, current fracture recovery investigations in mice frequently rely on animal sacrifice and costly analyses. The modified Radiographic Union Score for Tibia fractures (mRUST) scoring system is a validated metric for evaluating bony healing in humans utilizing plain radiographs, which are relatively inexpensive and do not require animal sacrifice. However, its use has not been well established in murine models. The aim of this study was to characterize the longitudinal course of mRUST and compare mRUST to other conventional murine fracture analyses. 158 mice underwent surgically created midshaft femur fractures. Mice were evaluated after fracture creation and at 7, 10, 14, 17, 21, 24, 28, 35, and 42 days post-injury. mRUST scoring of plain radiographs was performed by three orthopaedic surgeons in a randomized, blinded fashion. Interrater correlations were calculated. Micro-computed tomography (μCT) was analyzed for tissue mineral density (TMD), total callus volume (TV), bone volume (BV), trabecular thickness, trabecular number, and trabecular separation. Histomorphometry measures of total callus area, cartilage area, fibrous tissue area, and bone area were performed in a blinded fashion. Ultimate torque, stiffness, toughness, and twist to failure were calculated from torque-twist curves. A sigmoidal log-logistic curve fit was generated for mRUST scores over time which shows mRUST scores of 4 to 6 at 7 days post-injury that improve to plateaus of 14 to 16 by 24 days post-injury. mRUST interrater correlations at each timepoint ranged from 0.51 to 0.86, indicating substantial agreement. mRUST scores correlated well with biomechanical, histomorphometry, and μCT parameters, such as ultimate torque (r=0.46, p<0.0001), manual stiffness (r=0.51, p<0.0001), bone percentage based on histomorphometry (r=0.86, p<0.0001), cartilage percentage (r=-0.87, p<0.0001), tissue mineral density (r=0.83, p<0.0001), BV/TV based on μCT (r=0.65, p<0.0001), and trabecular thickness (r=0.78, p<0.0001), among others. These data demonstrate that mRUST is reliable, trends temporally, and correlates to standard measures of murine fracture healing. Compared to other measures, mRUST is more cost-effective and non-terminal. The mRUST log-logistic curve could be used to characterize differences in fracture healing trajectory between experimental groups, enabling high-throughput analysis.