Correlation of cancellous bone microarchitectural parameters from microCT to CT number and bone mechanical properties

Cancellous bone density alone is not a sufficient predictor of bone mechanical stiffness and strength. Microarchitectural parameters of bone may improve the accuracy of prediction. It is well established that bone density can be inferred from the intensities of clinical CT image datasets, however it is unknown if microarchitectural parameters manifest any relationship to the image intensities. Thirty-eight porcine cancellous bone cubes were subjected to clinical CT and microCT imaging to obtain the CT number and microarchitectural parameters respectively. They were subsequently compressed to failure to obtain the mechanical properties. Linear regression analyses were carried out for the three testing modalities to determine any possible correlations. Significant correlations (R² ≥ 0.5) were obtained between Hounsfield units and (1) volume fraction (BV/TV), (2) bone trabecular pattern (Tb.Pf), (3) trabeculae spacing (Tb.Sp), (4) structure model index (SMI) and (5) trabeculae number (Tb.N). Significant correlations were also discerned for microarchitectural parameters and mechanical properties. In conclusion, we established linear regression models between image-measured density in terms of Hounsfield units (HU) and microarchitectural parameters as well as models between cancellous bone microarchitectural parameters and mechanical properties. It should be noted that both density and microarchitectural parameters important to bone research could be inferred from clinical CT images.