Repository logo

Measurement of the bone endocortical region using clinical CT.

Accepted version



Change log


Pearson, RA 
Treece, GM 


The extent of the endocortical region and cortical bone mineral density (cBMD) throughout the proximal femur are of interest as both have been linked to fracture risk and osteoporosis treatment response. Non-invasive in-vivo clinical CT-based techniques capable of measuring the cortical bone attributes of thickness, density and mass over a bone surface have already been proposed. Several studies have robustly shown these methods to be capable of producing cortical thickness measurements to a sub-millimetre accuracy. Unfortunately, these methods are unable to provide high quality cBMD estimates, and are not designed to measure any attributes over the endocortical region of cortical bone. In this paper, we develop a cortical bone mapping based technique capable of providing an improved cBMD estimate and a measure of the endocortical width, while maintaining similar quality cortical thickness and trabecular bone mineral density (tBMD) estimates. The performance of the technique was assessed using a paired dataset of ex-vivo QCT and HR-pQCT scans across 72 proximal femurs. The HR-pQCT scans were analysed using a new method developed for this study: high resolution tissue classification (HRTC). In HRTC the cortical, endocortical and sub-surface trabecular bone features are extracted from the partially resolvable microarchitectural details in the HR-pQCT scan. We demonstrate that measurement of the endocortical extent from QCT is possible with an accuracy of -0.15±0.71mm, and that local cBMD can be measured down to densities of 300 mg/cm3.



Bone mineral density, Computed tomography, Cortical thickness, Endocortical region, Aged, Aged, 80 and over, Algorithms, Bone Density, Cadaver, Female, Femur, Humans, Male, Middle Aged, Radiographic Image Interpretation, Computer-Assisted, Reproducibility of Results, Tomography, X-Ray Computed

Journal Title

Med Image Anal

Conference Name

Journal ISSN


Volume Title



Elsevier BV