Holger Frank Boehm MD, Presenter: Nothing to Disclose

Christoph W. Raeth PhD, Abstract Co-Author: Nothing to Disclose

Roberto Adrian Monetti PhD, Abstract Co-Author: Nothing to Disclose

Dirk Mueller MD, Abstract Co-Author: Nothing to Disclose

Ernst Josef Rummeny MD, Abstract Co-Author: Nothing to Disclose

Thomas Marc Link MD, Abstract Co-Author: Nothing to Disclose

To use global topological texture measures based on the Minkowski Functionals in 2D (MF) for the analysis of bone mineral distribution from DXA-scans of the proximal femur in combination with bone mineral density (BMD) in standard regions of interest (ROI) for enhanced prediction of the bio-mechanical strength in-vitro.

Bone mineral density was assessed for 48 proximal femur specimens in the standard ROIs (shaft, trochanter, Ward’s area, femoral neck) using a clinical narrow fan beam DXA scanner with multi-view image reconstruction. For topological analysis of the pattern of bone mineral distribution, the MF corresponding to area, perimeter, and Euler-Poincaré characteristic were obtained as a function of gray-level from the original DXA scan images. Following imaging procedures, the specimens were tested destructively in a mechanical loading device to obtain the maximum compressive strength (MCS). An optimized texture parameter (MF_SW) was extracted from the spectra of MF with maximum correlation to MCS and minimum correlation to BMD and therefore providing information complementary to that of densitometry with respect to predicting bone strength. A multi-regressional model combining topological information and densitometry was employed for the assessment of MCS.

R² for the correlation between MCS and BMD ranged from .67 (trochanteric ROI) to .70 (neck ROI), p<.001. The optimized topological parameter MF_SW was correlated with MCS with R² = .62, whereas correlation with BMD ranged from .42 to.48. By combining the information of bone mineral distribution given by MF_SW and BMD using a multi-regressional model, the prediction of MCS was significantly improved with an R² of .89 (p<.001).

Topological analysis of bone mineral distribution in the proximal femur based on the MF and standard densitometric evaluation both provide widely complementary information with respect to fracture prediction of the hip. In our study, we demonstrated that the assessment of the bio-mechanical integrity is significantly enhanced by combining BMD and topological properties of the bone mineral distribution in the proximal femur as depicted in standard DXA scans.

Boehm, H, Raeth, C, Monetti, R, Mueller, D, Rummeny, E, Link, T, Enhanced Prediction of Hip Fracture Risk: Application of the Minkowski Functionals in 2D for Topological Analysis of the Bone Mineral Distribution in Standard DXA Scans of the Proximal Femur in-Vitro.  Radiological Society of North America 2004 Scientific Assembly and Annual Meeting, November 28 - December 3, 2004 ,Chicago IL. http://archive.rsna.org/2004/4403147.html