Analysis of bone architecture using fractal-based TX-Analyzer™ in adult patients with osteogenesis imperfecta.

ABSTRACT

BACKGROUND: Osteogenesis imperfecta (OI) is a rare genetic disorder characterized by impaired bone quality and quantity. Established imaging techniques have limited reliability in OI. The TX-Analyzer™ is a new, fractal-based software allowing a non-invasive assessment of bone structure based on conventional radiographs. We explored whether the TX-Analyzer™ can discriminate OI patients and healthy controls. Furthermore, we investigated the correlation between TX-Analyzer™ parameters and (i) bone mineral density (BMD) by Dual Energy X-ray Absorptiometry (DXA), (ii) trabecular bone score (TBS), and (iii) bone microstructure by high-resolution peripheral quantitative computed tomography (HR-pQCT). MATERIAL AND METHODS: Data of 29 adult OI patients were retrospectively analyzed. Standard radiographs of the thoracic and lumbar spine were evaluated using the TX-Analyzer™. Bone Structure Value (BSV), Bone Variance Value (BVV), and Bone Entropy Value (BEV) were measured at the vertebral bodies T7 to L5. Data were compared to a healthy, age- and gender-matched control group (n = 58). BMD by DXA, TBS, and trabecular bone microstructure by means of HR-pQCT were correlated to TX-Analyzer™ parameters in OI patients. The accuracy of the TX-Analyzer™ parameters in detecting OI was assessed with area under curve (AUC) analysis of receiver operating characteristic (ROC). RESULTS: BEV of the thoracic and the lumbar spine were significantly lower in OI patients compared to controls (both p < 0.001). BEV of the thoracic spine was significantly correlated to TBS (ρ = 0.427, p = 0.042) as well as trabecular number (Tb.N) at the radius (ρ = 0.603, p = 0.029) and inhomogeneity of the trabecular network (Tb.1/N.SD) at the radius (ρ = -0.610, p = 0.027), when assessed by HR-pQCT. No correlations were found between BEV and BMD by DXA. BEV of the thoracic and the lumbar spine had an AUC of 0.81 (95% confidence interval [CI] 0.67-0.94, p < 0.001) and 0.73 (95% CI 0.56-0.89, p = 0.008), respectively. BSV and BVV did not differ between OI patients and controls. CONCLUSION: The software TX-Analyzer™ is able to discriminate patients with OI from healthy controls. ROC curves of BEV values suggest a suitable clinical applicability. Low to no correlations with conventional methods suggest, that the TX-Analyzer™ may indicate a new and independent examination tool in OI.