Background/Purpose: 18F sodium fluoride (NaF) positron emission tomography (PET) allows quantitative assessment of osteoblastic bone synthesis in specific skeletal sites. Previous studies showed that increased 18F NaF uptake at vertebral corners was linked with syndesmophytes formation in patients with ankylosing spondylitis (AS), but the association between bone metabolism measured by 18F NaF PET and bone mineral density (BMD) has not been studied. Thus, we investigated whether baseline 18F NaF uptake of vertebral bodies or vertebral corners on PET is associated with changes in BMD at the corresponding lumbar vertebrae in AS.   

Methods: In 12 male AS patients, 18F NaF PET/computed tomography (CT) was performed at baseline and dual energy X ray absorptiometry (DEXA) was performed at baseline and the 1-year follow-up. The maximum standardized uptake value (SUVmax) of the vertebral body and both upper and lower vertebral corners from the L1 to L4 were measured in the mid-sagittal plane of the 18F NaF PET/CT image. By using the squared region of interest (ROI), the SUVmax of the vertebral body was acquired (Figure 1a). For the SUVmax of the corners, we used the arithmetic mean SUVmax of the upper and lower vertebral corners which were evaluated with 1cm-sized circular ROIs (Figure 1b). The BMD of the L1 to L4 were calculated from conventional DEXA. The association of the SUVmax of the vertebral bodies or corners with the changes in BMD during 1-year follow-up was analyzed at the lumbar vertebral level using generalized estimating equations (GEE) to adjust within-patient correlation for a total number of lumbar vertebrae.

Results: We analyzed 48 lumbar vertebrae in 12 AS patients. At the lumbar vertebral level, the mean (SD) baseline SUVmax of the vertebral body and the median (IQR) baseline SUVmax of the vertebral corners were 6.7(1.8) and 6.5(5.3-7.1), respectively. The mean (SD) BMD of each vertebra at baseline and 1-year follow-up were 1.18 (0.21) g/cm² and 1.19 (0.19) g/cm², respectively. In correlation analyses, changes in BMD at lumbar vertebrae was positively correlated with the baseline SUVmax of the vertebral bodies (r=0.39, p=0.036) but not with that of the vertebral corners (r=-0.07, p=0.657). In multivariable GEE analysis, the baseline SUVmax of the vertebral bodies was significantly associated with changes in BMD at the lumbar vertebrae [B(SE)=0.009(0.004), p=0.017], but that of vertebral corners did not showed this association [B(SE)=-0.003(0.002), p=0.174]. The baseline BMD of the vertebra also showed significant associations with changes in BMD.

Conclusion: Baseline 18F NaF uptake of the vertebral bodies but not that of the vertebral corners on PET was associated with 1-year changes in BMD at the lumbar vertebrae in AS patients. Our data suggests different mechanisms of bone metabolism between the vertebral bodies and corners that have separate effects on BMD in AS.