Endocrine Abstracts

Background: Vertebral Fractures (VF) are recognized as an important aspect of bone health in children and adolescents. The clinical utility of vertebral fracture assessment (VFA) using dual-energy X-ray absorptiometry (DXA) has not been evaluated in the paediatric population.

Method: VFA was performed independently by two non-radiologist observers, in 165 patients (77M/88F) as part of their investigation for low bone mineral density. Lateral thoracolumbar X-ray images (LXR) were obtained in 20/165 patients. The median age of the patients was 13.4 years (3.6, 18). Lateral DXA images of the spine from T6 to L4 were obtained using Lunar Prodigy DXA device. The diagnosis of VF was performed according to Genant’s semi-quantitative classification.

Results: Interobserver agreement in vertebral readability using VFA was 94% (κ, 0.73 (95% CI, 0.68, 0.73)). The vertebrae not readable by both observers were 287/1815 (16%) and 266/287 (93%) were located between T6 and T9. Conversely, 1134/1155 (98%) of vertebrae from T10 through L4 were adequately visualised (P<0.0001). Among the 1528 vertebrae visualised by both observers, 72 (4.7%) in 45 (27%) patients and 84 (5.5%) in 48 (29%) patients were classified as VF by observer 1 and by observer 2, respectively. Interobserver per-vertebra agreement for the presence of VF was 99% (κ, 0.85 (95% CI, 0.79, 0.91)). Interobserver per-patient agreement was 91% (κ, 0.78 (95% CI, 0.66, 0.87)). The two observers had in common 67 (4.5%) VF in 39 (24%) patients and 18 (27%) of them were classified as moderate or severe. The anatomical distribution of VF was biphasic, with peaks located on T9 (odds ratio, 2.1 (1.1, 4.2)) and L4 (odds ratio, 1.7 (1.0, 3.4)). Among those who underwent both LXR and VFA, 24 (11%) VF in 6 (30%) patients and 20 (9%) VF in 5 (25%) patients were identified by LXR and VFA, respectively. Per-vertebra agreement was 95% (κ, 0.79 (95% CI, 0.62, 0.92)) and per-patient agreement was 95% (κ, 0.88 (95% CI, 0.58, 1.0)). Specificity of VFA was 98.4% per-vertebra and 100% per-patient.

Conclusion: VFA reaches an excellent level of agreement between observers and a high level of specificity in identifying VF in paediatric population. The readability of vertebrae from T6 to T9 is suboptimal and interpretation at this level should be exercised with caution.