Abdominal CT was done using multidetector CT scanners (LightSpeed Series, GE Healthcare) calibrated daily to ensure accurate vertebral CT-attenuation numbers, which reflect underlying BMD (Figure 1). We retrospectively accessed the CT images and evaluated vertebral BMD on a standard radiology picture archiving and communication system workstation, with images viewed in soft tissue and bone windows (windows define gray-scale assignment of the image display to emphasize particular tissues and do not influence attenuation or BMD values [Figure 1]) (14 (link)). We assessed vertebral BMD by placing a single oval click-and-drag region of interest (ROI) over an area of vertebral body trabecular bone and then measuring CT attenuation in Hounsfield units (HU), with lower HU (lower attenuation) representing less-dense bone, at each of the T12 through L5 levels (Figures 1 and 2); this process is identical to that used for measuring CT attenuation for other clinical conditions (for example, adrenal adenomas, renal lesion enhancement, and fatty liver assessment). We avoided placing the ROI near areas that would distort the BMD measurement (posterior venous plexus; focal heterogeneity or lesion, including compression fracture; and imaging-related artifacts).
We assessed the presence of vertebral compression fractures by using sagittal CT views of the lumbar spine (Figure 2, B) by employing the Genant visual semiquantitative method (15 (link)), a widely accepted way of assessing vertebral fractures on conventional radiography that can be easily applied to sagittal CT images. We counted only obvious moderate (grade 2, 25% to 40% loss of height) or severe (grade 3, >40% loss of height) compression deformities to avoid ambiguity related to more subjective borderline or mild compression deformities. All potential moderate-to-severe compression fractures identified on the initial review were verified in a separate reading session for final confirmation, further excluding any questionable mild fractures.