Adjusting for the partial volume effect in cortical bone analyses of pQCT images

Rittweger, Jörn, Michaelis, Ingo, Giehl, M., Wüsecke, P. and Felsenberg, Dieter (2004) Adjusting for the partial volume effect in cortical bone analyses of pQCT images. The Journal of Musculoskeletal and Neuronal Interactions, 4 (4). pp. 436-441.

File not available for download.

Abstract

Quantitative analyses of computed tomography images are prone to errors due to the partial volume effect which affects objects (e.g., bones) that have a different size or are assessed with different resolution. We have developed a set of equations suitable for both modeling the partial volume effect in cortical bone and for performing the corresponding adjustment. Seven hollow cylinders and 2 cuboid phantoms were made out of Al with 1% Si. The specimens were scanned with a pQCT machine (XCT2002, Stratec Medizintechnik, Pforzheim, Germany) and analyzed with the integrated software, version 5.50. Measurements were performed at different resolutions (voxel size = 0.20 to 0.75 mm), both in air and in Ringer solution, and analyses were performed at different detection thresholds. Applying the correcting equations set we could reduce the errors in cortical density by about 80%. The cortical area was assessed with a negligible error at a threshold (Theta0) that is equivalent to the mean of the cortical bone density and of the background density. On chosing Theta0 as the detection threshold the error in density was lowered to less than 2%. We propose to assess cortical area and cortical density in several steps, first assessing the area and density thereafter. Applying this method should be beneficial whenever ‘true world’ values are required, or objects of different size are compared.