[Comparative evaluation of trabecular bone density in Hounsfield units in the lumbar native CT cross-section for osteoporosis diagnosis and fracture risk determination by different examiners].

Background: An increasing loss of bone mineral density (BMD) in the axial skeleton leads to osteoporosis and fractures, with an increase found in the thoracic and thoracolumbar regions.

Research Question: The extent to which an examiner-independent assessment of the extent of osteoporosis and fracture risk determination is possible by determining the trabecular density in Hounsfield units (HU) in the spine should be examined. The next question was whether quantitative BMD values can be calculated from the HU values.

Patients And Methods: 225 patients (pt.) with an average age of 64.9 ± 13.1 years and a body-mass-index (BMI) of 26.8 ± 6.8 kg/m, of which 37 were men and 188 were women, were examined to determine whether they had osteoporosis. The BMD was determined in mg/cm using quantitative computed tomography (QCT) in the lumbar region. After anonymization by three experienced radiologists, an additional measurement of the trabecular bone density in HU, was carried out in the same vertebral bodies (a total of 675 vertebral bodies), each using a region of interest (ROI) positioned in the midvertebral cancellous space in the sagittal reformed CT image. In additional lateral X‑rays of the thoracic and lumbar spine, vertebral fractures were detected and graded. Sacral insufficiency fractures that occurred at the same time were also recorded.

Results: The median BMD was 73.2 (57.05-104.17) mg/cm and the median HU was 89.93 (67.90-126.95). With a correlation of 0.988 (p < 0.001), quantitative values in mg/cm can be calculated using the following formula: Xq = 12.1 + 0.68 × HU. With HU values less than 69.84 and a BMD of the lumbar spine below 59.54 mg/cm, there was a significantly increased number of OVF. At least one OVF was found in 137/225 pt. In 17/137 pt., sacral fractures were also found; these patients showed the significantly lowest values with a median BMD of 41.81 (16.2-53.7) mg/cm. Comparable HU values were determined independently of the examiners (p > 0.05).

Discussion: The trabecular density measurements in HU values can be converted into quantitative BMD values in mg/cm, which enables a good assessment of osteoporosis and fracture risk. Taking the results obtained into account, an opportunistic evaluation using HU values in native CT alone seems quite possible. Experienced examiners have arrived at comparable results.