Methods
Technique for HU Assessment of the Spine
Schreiber et al. described a simple technique for assessing regional BMD of the spine that was correlated with results of DXA scans as well as compressive strength in osseous models. Using standard picture archiving and communication system software, an ROI is drawn at 3 points on axial images that are obtained as parallel to the endplates as possible: just inferior to the superior endplate, midvertebral body, and just superior to the inferior endplate (Figure 1). The ROI is drawn encapsulating only cancellous bone and avoiding cortical edges, osseous abnormalities, and voids such as vascular channels. The software calculates the average HU in the ROI for each image. An average of the 3 measurements determines the HU for individual vertebral levels. Others have found that a single midbody ROI is as reliable as the average of 3 measurements and have used an automated software program to generate values.
(Enlarge Image)
Figure 1.
Computed tomography scans illustrating the technique for calculating vertebral BMD with HUs. A: Sagittal slice of the L-3 vertebral body demonstrating axial planes of interest. B–D: Axial images showing HU values generated by the imaging software program. The mean value of B–D is used for calculating vertebral body bone density.
Study Cohorts
All studies had institutional review board exemptions. Several patient cohorts were used to answer specific clinical questions. Twenty-five consecutive patients who underwent CT scanning to evaluate possible spinal injury and who had recently undergone DXA were used to evaluate reliability and correlations. Eighty patients who were evaluated for trauma were used to measure normative values. Twenty patients with adjacent-segment fractures after spinal fusion along with 20 matched nonfracture controls were used to assess the association between HU and adjacent-segment fracture. The CT scans of 28 patients with a combined 52 levels of lumbar interbody fusion were used to assess the association between HU and fusion success.
Reliability of HU Measurements
The intra- and interobserver reliability of HU measurements was performed using an interclass correlation coefficient. Two observers independently measured HU at 3 locations of 4 vertebrae in 25 subjects.
Normative Data
Normative data stratified by sex and decade of life were determined by Schreiber et al. based on CT scans from 80 consecutively presenting trauma patients grouped by decade from 10–19 years to 80–89 years, with 10 patients in each decade. These data were compared with other published normative data.
Correlation of HU and DXA
Twenty-five patients undergoing lumbar CT and DXA within 6 months of each other were identified, and the BMD, T-score, and HU value were measured from L-1 to L-4. These patients had a mean age of 71.5 years with a range of 33–87 years. There were 18 female and 7 male patients. Correlation of T-score and BMD from DXA to HU measurement was performed using the Pearson correlation coefficient. Each patient was grouped into normal (T-score ≥ −1.0), osteopenia (T-score −1.0 to −2.5), and osteoporotic (T-score ≤ −2.5). The HU threshold values for each bone health state were reported as the mean and 95% confidence intervals.
HU Value and Fracture Risk
Meredith et al. matched 20 patients with adjacent-segment fractures after spinal fusion with nonfracture controls of similar age, sex, body mass index, and surgical procedure. Hounsfield unit measurements were taken throughout the spine from preoperative CT scans, and a Student t-test was used to compare means between the groups.
HU Value and Fusion Success
Schreiber et al. compared the HU of vertebral bodies adjacent to 38 successful fusions to the HU values adjacent to 14 nonunions. There were no differences in age, sex, body mass index, or smoking status between the 2 cohorts. Hounsfield unit measurements both within the fusion construct and throughout the adjacent spine were recorded using a previously described technique, and mean values between the groups were assessed using a Student t-test.