The predictive value of Hounsfield units for titanium mesh cage subsidence after anterior cervical corpectomy and fusion.

Abstract

To investigate whether bone mineral density (BMD) measured in Hounsfield units (HUs) correlates with titanium mesh cage (TMC) subsidence after anterior cervical corpectomy and fusion (ACCF). A total of 64 patients who underwent one or two levels of ACCF with TMC with a mean follow-up of 19.34 ± 7.86 months were analysed. HU values were measured three times in 3 different planes in the upper and lower vertebrae according to published methods. Subsidence was defined as segmental height loss of more than 3 mm. Pearson correlation analysis was performed. Receiver operating characteristic (ROC) curve analysis was used to obtain optimal thresholds. A multivariate logistic regression analysis was also conducted. Twenty-two patients (34.38%) had evidence of TMC subsidence on follow-up x-ray. The mean HU values in the subsidence group (317.34 ± 32.32, n = 22) were significantly lower than those in the nonsubsidence group (363.07 ± 25.23 n = 42, p < 0.001, t test). At last follow-up, mean disc height loss was 4.80 ± 1.16 mm in the subsidence group and 1.85 ± 1.14 mm in the nonsubsidence group (p < 0.001). There was a negative correlation between HU values and disc height loss (Pearson's coefficient -0.494, p < 0.001). HU values decreased gradually from the C3 vertebra to the C7 vertebra, and the HU values of the C5, C6, and C7 vertebrae in the nonsubsidence group were significantly higher than those in the subsidence group (p < 0.05). Furthermore, there were significant differences between the groups in the segmental angle at the last follow-up and the mean changes in segmental angle (p < 0.05). The area under the ROC curve was 0.859, and the most appropriate threshold of the HU value was 330.5 (sensitivity 100%, specificity 72.7%). The multivariate logistic regression analysis showed that older age (p = 0.033, OR = 0.879), lower LIV HU value (p < 0.001, OR = 1.053) and a greater segmental angle change (p = 0.002, OR 6.442) were significantly associated with a higher incidence of TMC subsidence after ACCF. There are strong correlations between a lower HU value and TMC subsidence after ACCF. More accurate assessment of bone quality may be obtained if HU measurement can be used as a routine preoperative screening method together with DXA. For patients with HU values <330.5, a more comprehensive and cautious preoperative plan should be implemented to reduce TMC subsidence.