Correlation Between Cone‐Beam Computed Tomography and High‐Resolution Peripheral Computed Tomography for Assessment of Wrist Bone Microstructure. (26th March 2019)

• Record Type: Journal Article
• Authors: Mys, Karen; Varga, Peter; Gueorguiev, Boyko; Hemmatian, Haniyeh; Stockmans, Filip; van Lenthe, G Harry
• Is Part Of: Journal of bone and mineral research. Volume 34:Number 5(2019)
• Page Start: 867
• Page End: 874
• Publication Date: 2019-03-26
• Subjects: HR‐PQCT -- MICRO‐COMPUTED TOMOGRAPHY -- CONE‐BEAM COMPUTED TOMOGRAPHY -- BONE MORPHOMETRY -- ACCURACY; Bones -- Metabolism -- Periodicals; Mineral metabolism -- Periodicals; 612.392
• DOI: 10.1002/jbmr.3673
• Languages: English
• ISSNs: 0884-0431

Abstract

ABSTRACT: High‐resolution peripheral quantitative computed tomography (HR‐pQCT) is considered as the best technique to measure bone microarchitecture in vivo. However, a breakthrough for medical applications is inhibited because of the restricted field of view (∼9 mm) and a relatively long acquisition time (∼3 minutes). The goal of this study was to compare the accuracy of cone‐beam computed tomography (CBCT) and HR‐pQCT and to determine the agreement between CBCT and HR‐pQCT in quantifying bone structural parameters. Nineteen trapezia of arthritic patients were scanned four times ex vivo: 1) CBCT (NewTom 5G, Cefla, at 75 μm); 2) HR‐pQCT (XTremeCT‐I, Scanco, at 82 μm); 3) HR‐pQCT (XTremeCT‐II, Scanco, at 60.7 μm); and 4) microCT (SkyScan1172, Bruker, at 19.84 μm). XTremeCT‐I and XtremeCT‐II were reconstructed, segmented, and analyzed following the manufacturer's guidelines. CBCT was reconstructed with in‐house developed software and analyzed twice: once with an adaptive segmentation technique combined with a direct analysis method (AT‐DM) and once with a Laplace‐Hamming filtering technique combined with an indirect analysis method (LH‐IM). Parameters of interest included bone volume fraction (BV/TV) and trabecular thickness (Tb.Th), separation (Tb.Sp), and number (Tb.N). The analyses of the CBCT data showed that the AT‐DM analysis correlated better with microCT for BV/TV, Tb.Sp, and Tb.N, whereas the LH‐IM technique correlated better for Tb.Th. Evaluated over all parameters, the coefficient of determination for XtremeCT‐I, XtremeCT‐II, and CBCT were higher as R 2 = 0.68, 0.72, and 0.67, respectively. For CBCT, the correlations improved when three samples with very thin trabeculae close to each other were excluded and became similar to those for XtremeCT‐I and XtremeCT‐II. Interesting for clinical practice is that those bones could be identified automatically with the CBCT scanner. We conclude that CBCT produced similar accuracy as HR‐pQCT in bone morphometric analyses of trapezia. The broader range of application, larger field of view, and shorter acquisition time make CBCT a valuable alternative to HR‐pQCT. © 2019 American Society for Bone and Mineral Research.