RETURN TO ABSTRACT LISTING


Objective:
While dual energy X-ray absorptiometry (DEXA) is the gold standard for osteoporosis diagnoses, it is underutilized and often inaccessible to patients. CT opportunistic bone density screening has been established as a potential method to mitigate the screening gap. Prior studies have established useful Hounsfield unit (HU) screening thresholds for noncontrast CTs and have shown that intravenous contrast-enhanced (CE) CTs need higher thresholds. However, investigators have yet to establish a similar HU threshold for CECTs. The objective of this study is to determine a CECT L1 vertebral body bone mineral density (BMD) threshold, both as BMD alone and normalized to adjacent muscle density (MD), that can be used to screen for osteopenia and osteoporosis.

Materials and Methods:
Our study was IRB-approved and HIPAA-compliant. We found 18 consecutive adults (12 women, 6 men; mean age, 66.6 years) who received CECT and DEXA scans = 2 months apart (mean interval, 31.8 days). Patients with histories of bone metastasis, chemotherapy, osteoporosis treatment, hormone therapy, or dialysis were excluded. L1 vertebral bodies that were fractured, had sustained > 20% height loss, or contained hardware were also excluded. Mean BMD was measured in the L1 vertebral body on an axial slice using a region of interest (ROI) attenuation in the trabecular bone, just above the mid-vertebral body, in an area where there was no cortical bone or obvious abnormal sclerosis or lucency. MD measurements of the right and left psoas and paraspinal muscles measured on the same slice were used to calculate the mean MD. The BMD-MD ratio was defined as BMD divided by mean MD. DEXA T-scores and diagnoses were recorded according to the WHO classification system (> -1.0 = Normal, - 2.5 <-1.0 = Osteopenia, < -2.5 = Osteoporosis). Area under the receiver operating characteristic (AUC) analyses were performed to determine the correlation between the BMD-MD ratios and DEXA diagnoses, which were dichotomized into classes of normal and abnormal (osteopenia or osteoporosis). The diagnostic utilities of the BMD alone and the BMD-MD ratio for abnormal bone density were assessed.

Results:
BMD alone was not statistically significantly associated with abnormal DEXA (<em>p</em> = 0.859), so an optimized diagnosis threshold could not be determined. The BMD-MD ratio was lower for patients with DEXA-defined osteoporosis, and this difference approached statistical significance (<em>p</em> = 0.09). The AUC for diagnosing abnormal bone density using BMD-MD ratio was 0.738 (95% CI: 0.489 - 0.986). A BMD-MD ratio of 3.0 demonstrated a 62.5% (28.9 - 96.1%) sensitivity, 80.0% (55.2 - 100%) specificity, 71.4% (38.0 - 100%) positive predictive value (PPV), and 72.7% (46.4 - 99.1%) negative predictive value (NPV).

Conclusion:
The high specificity and high PPV and NPV of BMD-MD ratios in the detection of abnormal bone density suggests their potential utility in the diagnosis of osteoporosis. Further study with a larger cohort is warranted.