News |  Sitemap |  Contact
PDF Export  | 

Long Bone Geometry Determined with an ImageJ Macro


The long bones of the skeleton are known to scale allometrically: bone diameter increases proportionally more rapidly than bone length as body mass increases, which is an important factor in maintaining limb structural integrity. The relationships between body mass, bone length and bone diameter have been well studied across a broad range of species using traditional 'balance and rule' instrumentation (Carter, Wong & Orr 1991). However, previous studies were limited to minimal, simplistic measurements while assuming cylindrical bone geometry and so failed to investigate true bone structure. We are extending knowledge of bone allometric scaling by analysing cross-sectional geometry of long bones to determine mechanically important but previously unmeasured bone features. Computed tomography (CT; Picker PQ5000) scans were made of appendicular long bones from quadrupeds with body masses between 5 and 5000 kg. Scans were saved as Hounsfield unit (HU) calibrated 512×512 pixel 16-bit DICOM images. An ImageJ macro was written that automatically determined cortical diameter, perimeter, thickness and cross-sectional area for each CT slice. An implementation of the rotating caliper method was developed for minimum and maximum cortical diameter. The macro incorporated MomentMacroJ-1.3 (Warfel, Serafin, DeLeon 2006) for determination of centroids, principal axes, second moments of area, polar moment of inertia and section moduli. MomentMacroJ was extended to include HU weighting as an approximation of volumetric bone density, so to avoid threshold-related error caused by edge voxels and to incorporate all bone (trabecular and cortical) in moment calculations. The macro could process either single slices or all slices in a stack, requiring < 10s to process each slice, or approximately 10 minutes to process a 200-slice stack on a dual core 2.2 GHz processor. Bone length was measured and midshaft slice identified by clicking on two points within the stack corresponding to most proximal and most distal landmarks while running a sub-macro. Data were written to the Results window and saved for later analysis. Centroids and principal axes were plotted on a copy of the CT stack, allowing contextual visualisation of their variation along the bone using the ImageJ 3D Viewer (Schmid 2007). This macro has allowed rapid acquisition of cortical bone geometric data which will contribute to advanced understanding of long bone allometry.


Carter DR, Wong M, Orr TE. Musculoskeletal ontogeny, phylogeny and functional adaptation. J Biomech 24 Suppl 1:3-16 (1991)

Schmid B. ImageJ 3D Viewer. First public version (2007)

Warfel M, Serafin S, DeLeon VB. MomentMacroJ. Version 1.3 (2005)

Bone, computed tomography, macro, allometry

Dr Michael Doube

Imperial College London


Short Biography   
I studied veterinary science at Massey University in New Zealand, graduating with a Bachelor of Philosophy in 2002 and Bachelor of Veterinary Science in 2004. After a brief stint as a full-time clinician I began a PhD on racehorse fracture at Queen Mary, University of London under the supervision of Professor Alan Boyde. I am now employed as a Postdoctoral Research Associate at Imperial College London's Department of Bioengineering. My research involves imaging approaches to orthopaedic tissues including confocal and electron microscopy, radiography, DXA, microCT and clinical CT. I also work in small animal clincs in London.

© Luxembourg Institute of Science and Technology | Legal Notice