Scaling VOI size in 3D μCT studies of trabecular bone: A test of the over-sampling hypothesis

Authors

  • Richard A. Lazenby,

    Corresponding author
    1. Anthropology Program, University of Northern British Columbia, 3333 University Way, Prince George, BC, Canada V2N4Z9
    • Anthropology Program, University of Northern British Columbia, 3333 University Way, Prince George, BC, Canada V2N4Z9
    Search for more papers by this author
  • Matthew M. Skinner,

    1. Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology, Deutscher Platz 6, 04103 Leipzig, Germany
    Search for more papers by this author
  • Tracy L. Kivell,

    1. Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology, Deutscher Platz 6, 04103 Leipzig, Germany
    Search for more papers by this author
  • Jean-Jacques Hublin

    1. Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology, Deutscher Platz 6, 04103 Leipzig, Germany
    Search for more papers by this author

Abstract

For comparative 3D μCT studies of trabecular bone, the use of a volume of interest (VOI) scaled to body size may avoid over-sampling the trabecular mass in smaller versus larger-bodied taxa and comparison of regions that are not functionally homologous (Fajardo and Müller: Am J Phys Anthropol 115 (2001) 327–336), though the influence on quantitative analyses using scaled versus nonscaled VOIs remains poorly characterized. We compare trabecular architectural properties reflecting mass, organization, and orientation from three volumes of interest (large, scaled, and small) obtained from the distal first metacarpal in a sample of Homo (n = 10) and Pan (n = 12). We test the null hypotheses that neither absolute VOI size, nor scaling of the VOI to metacarpal size as a proxy for body size, biases intraspecific analyses nor impacts the detection of interspecific differences. These hypotheses were only partially supported. While certain properties (e.g., bone volume fraction or trabecular thickness) were not affected by varying VOI size within taxa, others were significantly impacted (e.g., intersection surface, connectivity, and structure). In comparing large versus scaled VOIs, we found that the large VOI inflated the number and/or magnitude of significant differences between Homo and Pan. In summary, our results support the use of scaled VOIs in studies of trabecular architecture. Am J Phys Anthropol, 2011. © 2010 Wiley-Liss, Inc.

Ancillary