Prediction of vertebral body mechanical parameters using opportunistic CT data
Abstract
Introduction:
The purpose of the study was to test a method describing the mechanical properties of bone using clinically available CT data.
Material and Methods:
The samples, 50 L3 vertebrae taken from males 22 to 81 years old, were examined with dual-energy X-ray absorptiometry and quantitative CT. An analysis of CT images and their image histograms was performed. The greyscale means – XC1, XC2, their standard deviations – SD1, SD2, and the areas under the curves – X1, X2 characterizing the organic matrix and bone material, respectively, were calculated by fitting two Gaussian functions. The compression tests were performed to determine the elastic modulus (E), ultimate stress (σmax), ultimate strain, and the ratio of work to fracture and the volume of the vertebra.
Results:
It was found that E and σmax were most precisely described by the parameter related to the trabecular bone density (XC2) obtained from the histogram analysis. Using the linear model, the coefficient of determination (R2) equals to 0.706 and 0.846 for E and σmax, respectively. For volumetric (vBMD) and areal bone mineral density (aBMD), R2 is 0.641 and 0.208 for E, while for σmax equals 0.784 and 0.356. After correction of vBMD using the histogram parameters R2 for E and σmax rise to 0.692 and 0.835, respectively.
Conclusions:
The superiority of the new method of E and σmax estimation based on clinically available CT data was confirmed. The proposed method does not require calibration and predicts the mechanical parameters of the vertebrae more precisely than vBMD or aBMD separately.
© 2024 Eugeniusz Rokita, Grzegorz Tatoń, published by Polish Society of Medical Physics
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.