The choice of region of interest after spinal procedures alters bone mineral density measurements

PURPOSE: Vertebrae affected by artifacts, such as metallic implants or bone cement, should be excluded when measuring the spine bone mineral density (BMD) by dual-energy X-ray absorptiometry (DXA). Exclusion may be performed using two methods: first, the affected vertebrae are included in the region of interest (ROI) and subsequently excluded from the analysis; second, the affected vertebrae are completely excluded from the ROI. This study aimed to investigate the influence of metallic implants and bone cement on BMD with and without the inclusion of artifact-affected vertebrae in the ROI. METHODS: DXA images of 285 patients, including 144 with spinal metallic implants and 141 who had undergone spinal vertebroplasty from 2018 to 2021, were retrospectively reviewed. Spine BMD measurements were performed when the images were evaluated using two different ROIs for each patient during the same examination. In the first measurement, the affected vertebrae were included in the ROI; however, the affected vertebrae were excluded from the BMD analysis. In the second measurement, the affected vertebrae were excluded from the ROI. Differences between the two measurements were evaluated using a paired t-test. RESULTS: Among 285 patients (average age, 73 years; 218 women), spinal metallic implants led to an overestimation of bone mass in 40 of 144 patients, whereas bone cement resulted in an underestimation of bone mass in 30 of 141 patients when the first measurement was compared with the second measurement. The opposite effect occurred in 5 and 7 patients, respectively. Differences in results between the inclusion and exclusion of the affected vertebrae in the ROI were statistically significant (p<0.001). Spinal implants or cemented vertebrae included in the ROI might significantly alter BMD measurements. Additionally, different materials were associated with varying modifications in BMD. CONCLUSION: The inclusion of affected vertebrae in the ROI may notably alter BMD measurements, even when they are excluded from the analysis. This study suggests that the vertebrae affected by spinal metallic implants or bone cement should be excluded from the ROI.