Regional Hounsfield unit measurement of screw trajectory for predicting pedicle screw fixation using cortical bone trajectory: a retrospective cohort study

Background The sufficiency of screw anchoring is a critical factor for achieving successful spinal fusion; however, no reliable method for predicting pedicle screw fixation has been established. Recently, Hounsfield units (HU) obtained from computed tomography (CT) was developed as a new reliable tool to determine the bone quality. The purpose of the present study was to demonstrate the utility of regional HU measurement of the screw trajectory to predict the primary and long-term fixation strength of pedicle screws. Method The insertional torque of pedicle screws using the cortical bone trajectory technique was measured intraoperatively in 92 consecutive patients who underwent single-level posterior lumbar interbody fusion. The cylindrical area of each screw was plotted on the preoperative CT image by precisely confirming the screw position, and the screw trajectory was measured in HU. First, three parameters: the bone mineral density (BMD) of the femoral neck and lumbar vertebrae, and regional HU values of the screw trajectory, were correlated with the insertional torque and compared among three groups. Next, pedicle screw loosening was evaluated by postoperative CT obtained 12 months after surgery, and clinical and imaging data were analyzed to assess whether regional HU values could be used as a predictor of screw loosening. Results Regional HU values of the screw trajectory (r = 0.75, p < 0.001) had stronger correlation with the insertional torque than the femoral BMD (r = 0.59, p < 0.001) and lumbar BMD (r = 0.55, p < 0.001). The incidence of screw loosening was 4.6% (16/351). Multivariate logistic regression analysis revealed that regional HU value (odds ratio = 0.70; 95% confidence interval = 0.56-0.84; p = 0.018) was an independent risk factor significantly affected screw loosening. Conclusions Regional HU values of the screw trajectory could be a strong predictor of both primary and long-term screw fixation in vivo.