Reliability of the assessment of disc degeneration on the lateral DXA scans

Abstract Introduction: Given the prevalence and costs induced by osteoarthritis (OA), it is necessary to find a cheap and safe technique to evaluate it reliably. Objective: To assess the value of the lateral dual energy X-ray absorptiometry (DXA) spine scans for the diagnosis of disc degeneration. Method: Seventy-seven individuals aged 18 and over, with or without disc degeneration, had both lateral thoracolumbar spine radiographs and DXA spine scans (≤ 6 months between both exams). Disc degeneration was assessed using the Lane score. The images of 20 randomly selected individuals were assessed by two readers. Results: Almost 13% of the thoracic levels were not assessable on the DXA scans. For the identification of the intervertebral levels on the DXA scans as interpretable or not, the intra-reader agreement was good (xf06b;=0.81) and the inter-reader agreement was fair (xf06b;=0.27–0.36). For the diagnostic criteria (osteophytes, disc space narrowing, osteosclerosis, overall grade), the intra-reader agreement was excellent for the radiographs (xf06b;=0.89-0.92), good for the DXA scans (xf06b;=0.64–0.83) and fair to moderate for the between-method comparison (xf06b;=0.25-0.44). The inter-reader agreement was moderate to good for the radiographs (xf06b;=0.49–0.66) and fair to good for the DXA scans (xf06b;=0.32–0.74). In the per patient analysis (the most severe grade), the intra-reader agreement was excellent for the radiographs (xf06b;=0.85–0.94), moderate to excellent for the DXA scans (xf06b;=0.53–0.85) and poor to good for the between-methods comparison (xf06b;=0.17–0.63). Conclusion: Our results do not support the use of DXA scans for the assessment of thoracolumbar disc degeneration.