Pierre Pottecher, Presenter: Nothing to Disclose

Klaus Engelke PhD, Abstract Co-Author: Nothing to Disclose

David Mitton, Abstract Co-Author: Nothing to Disclose

Thomas Moser MD, Abstract Co-Author: Research Consultant, Horizon Sciences & Technologies Inc

Oleg Museyko, Abstract Co-Author: Nothing to Disclose

Eric Vicaut, Abstract Co-Author: Nothing to Disclose

Judith Elizabeth Adams MD, Abstract Co-Author: Nothing to Disclose

Wafa Skalli, Abstract Co-Author: Nothing to Disclose

Jean-Denis Laredo MD, Abstract Co-Author: Research Consultant, Cardinal Health, Inc Research Consultant, Laurane Medical Research Consultant, F. Hoffman-La Roche Ltd Research Grant, SERVIER

Valerie Bousson, Abstract Co-Author: Nothing to Disclose

To predict the strength of the proximal femur with three imaging modalities: plain radiographs (XR), dual X-ray absorptiometry (DXA), quantitative computed tomography (QCT) with a dedicated three-dimensional image analysis tool (MIAF-Femur)and finite element model (FEM).

The proximal ends of forty pairs of excised femurs (82 +/-12 years) were studied. Each femur was analyzed using (1) radiographs to measure geometric parameters: lengths, angle, cortical thicknesses; (2) DXA (gold standard) to measure bone mineral densities (aBMD); (3) QCT with a three-dimensional (3D) analysis tool (medical image analysis framework (MIAF-Femur)) to determine bone mineral densities (vBMD) and geometric variables (hip axis length, cortical thicknesses, volumes, moments of inertia) of cortical and trabecular bone; (4) CT-based FEM to calculate a numerical value of failure load. The eighty femurs were also studied using mechanical tests to failure either in stance or lateral configuration (random assignment of the two femurs from each pair to one mechanical configuration). The variables were described with mean, standard deviation, and range. Univariate correlations and multivariate models were computed for each imaging modalities, and FEM, to predict failure load in both configurations.

In multivariate analysis, models explained 66% (XR), 73 % (DXA), 76 % (QCT) and 87% (FEM) of the variance of the fracture load and 63 % (XR), 79 % (DXA), 83 % (QCT) and 84 % (FEM) in stance and lateral configurations respectively

Simple measurements of geometric variables using radiographs, or simple measurements of densitometric variables using DXA, explains a large part of femoral failure load. However femoral failure load is best explained by a combination of geometric and densitometric variables as provided by QCT-MIAF or CT-based FEM.

The overall aim of our study is to improve osteoporotic hip fracture prediction to guide appropriate prevention.

Pottecher, P, Engelke, K, Mitton, D, Moser, T, Museyko, O, Vicaut, E, Adams, J, Skalli, W, Laredo, J, Bousson, V, Prediction of Osteoporotic Hip Fractures: An in vitro Study of 80 Femurs Assessed by Three Imaging Modalities and Finite Element Models. The European Femur Fracture Study (EFFECT).  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14009599.html