Relationship between structural parameters, bone mineral density and fracture load in lumbar vertebrae, based on high-resolution computed tomography,quantitative computed tomography and compression tests

Different noninvasive techniques for the assessment of the individual fract ure risk in osteoporosis are introduced, and the relation between structura l properties of high-resolution computed tomography (HR-CT) images of verte bral bodies, their bone mineral density (BMD) and the fracture load is anal yzed. In 24 unfractured lumbar vertebrae with different degrees of deminera lization from six specimens, the trabecular and cortical BMD was determined using quantitative CT. A lateral X-ray image revealed the number of fractu res in the entire spine. A structural analysis of spongy and cortical bone was performed based on the HR-CT images. In the spongiosa, the fractal dime nsion was calculated as a function of the threshold value. In the cortical shell, the maximum number of clusters of low BMD was determined at varying threshold values. After the CT measurements the vertebrae were excised and compressed until fractured. On the basis of the spongiosa BMD and the numbe r of fractures, 3 cases were found to be severely osteoporotic; the other 3 cases showed osteopenia. The average fracture loads were determined as 353 3 N for the non-osteoporotic cases (range 2602-5802 N) and 1725 N for the o steoporotic cases (range 1311-2490 N). The parameters were determined as fo llows: average spongiosa BMD 115.2 mg/ml (101.8-135.3 mg/ml) for the nonost eoporotic cases, 46.2 mg/ml (34.8-57.6 mg/ml) for the osteoporotic cases; a verage cortical BMD 285.1 mg/ml (216.4-361.9 mg/ml) for the non-osteoporoti c cases, 136.1 mg/ml (142.5-215.2 mg/ml) for the osteoporotic cases; spongi osa structure: average 0.5 (range 0.32-0.75) for the non-osteoporotic cases , average 1.05 (range 0.87-1.24) for the osteoporotic cases; cortical struc ture: average 81 (range 55-104) for the non-osteoporotic cases), average 13 6 (range 102-159) for the osteoporotic cases. Single parameters (BMD and st ructure) and weighted sums of these parameters were correlated with the fra cture load, resulting in correlation coefficients of r(sBMD) = 0.82 (spongi osa BMD), r(cBMD) = 0.82 (cortical BMD), r(sStr) = -0.75 (spongiosa structu re) and r(cStr) = -0.86 (cortical structure). The weighted sum of cortical and spongiosa BMD resulted in r(BMD) = 0.86, of cortical and spongiosa stru cture in r(Str) = -0.86. A weighted combination of all four parameters corr elates with the fracture load at r(4) = 0.89, all correlations being statis tically significant (p<0.0001). The four individual parameters show only a slight overlap between nonosteoporotic and osteoporotic subjects. The high correlation of the cortical BMD and the structural parameter in cortical bo ne indicates the important contribution of the cortical shell to vertebral stability. A weighted sum of multiple parameters results in a higher correl ation with the fracture load and does not show an overlap between the two g roups. It is best suited to estimate the individual fracture risk. The pres ented methods are generally applicable in vivo; and allow an improvement of the diagnosis of osteoporosis compared with the measurement of the BMD alo ne.