Metabolic and densitometric correlation between atherosclerotic plaque and trabecular bone: an 18F-Natrium-Fluoride PET/CT study

Abstract

Increasing evidence links atherosclerosis to a decreased bone thickness. This correlation could reflect a bone/plaque interaction. Hereby we analyzed Hounsfield density (HU) and mineral turnover in bone and in the arterial calcifications (AC), using a computational method applied to PET/CT data. 79 ¹⁸F-NaF PET/CT from patients with AC were retrospectively analyzed. Mean AC density and background-corrected uptake (TBR) were estimated after semi-automatic isocontour segmentation. The same values were assessed in the trabecular bone, using an automatic adaptive thresholding method. Patients were then stratified into terciles, according to their mean HU plaque density ("light", "medium" or "heavy" calcifications"). 35 ¹⁸F-NaF PET/CT from patients without AC served as controls. Vertebral density and TBR were lower in patients than in controls (137±25 vs. 160±14 HU, P<0.001); (6.2±3.9 vs. 8.4±3.4, P<0.05). Mean trabecular TBR values were 8.3±4, 4.5±2.1 and 3.5±1.8 in light, medium and heavy AC groups, respectively (P<0.05 for light vs. medium and P<0.01 for light vs. heavy). Similarly, mean trabecular HU was 143±19, 127±26 and 119±18 in the three groups, respectively (P<0.01 for light vs. heavy). Mean AC density was inversely associated with the trabecular HU (R=-0.56, P<0.01). Conversely, plaques' TBR directly correlated with the one in trabecular bone (R=0.63, P<0.001). At multivariate analysis, the sole predictor of vertebral density was plaque HU (P<0.05). Our data highlight a correlation between plaque and bone morpho-functional parameters and suggest that observing skeletal bone characteristics could represent a novel window on atherosclerosis pathophysiology.

Keywords: Positron emission tomography; atherosclerosis; bone density; computational analysis.

Figure 1

Density trabecular bone and in the aortic plaque. In the three top panels are displayed low-dose CT images depicting examples of typical “light”, “medium” and “heavy” plaques. Directly below (middle panels) are displayed CT slices of the trabecular bone from the same patients. After terciles stratification according to mean plaque density, subjects with a higher mean plaque thickness presented a decreased vertebral density and metabolism (lower histogram).

Figure 2

Mineral metabolic activity in patients and controls. The histograms depict the target-to-background ratio of subjects with a ACs and of controls in trabecular bone (upper) and the corresponding value within the plaques. Subjects with ACs had a lower TBR than controls; however, this metabolic decrease was circumscribed to patients with medium and heavy calcifications. Likewise, plaque TBR was higher in less heavily calcified plaques.

Figure 3

Correlation between vertebral density and plaque features. A higher plaque density was associated to a lower trabecular density; conversely, TBR in the AC was directly associated with the one in the trabecular skeleton.

Figure 4

Correlation between plaque and trabecular mineral metabolism. Patients with thicker arterial calcifications showed a reduced mineral metabolism within the plaque itself and in the trabecular bone.