Comparison of bone mineral density quantification in dogs using spectral detector computed tomography versus...

The generation of virtual monoenergetic images using spectral detector computed tomography (SDCT) may facilitate the measurement of bone mineral density (BMD) without the requirement of a phantom. This study has two primary objectives: (1) To compare the BMD values obtained from SDCT maps with those derived from phantom-calibrated values using quantitative computed tomography (QCT) in specific phantom densities and predetermined locations in canine subjects; and (2) to assess the reproducibility, measurement precision, and the potential bias associated with phantom-based measurements.

SDCT examinations of 49 dogs included a phantom containing four hydroxyapatite inserts. BMD values were manually measured in 18 anatomical locations. A linear model was used to convert Hounsfield units to BMD values (mg/cm³). A paired Wilcoxon signed-rank test with Bonferroni-correction and Pearson correlation were used for statistical analysis. A p-value of ≤ 0.05 was considered significant.

The statistical analysis demonstrated consistently lower BMD values derived from SDCT data within the phantom volume of interest. However, when compared to QCT, higher BMD values were noted across all anatomical sites. QCT data provided BMD values closer to the density of the phantom, while SDCT data appeared to be less sensitive to phantom positioning and body weight. The absolute differences in phantom values were influenced by the number of voxels without completely correcting the generally observed differences in the measured values.

BMD values from both methods demonstrated significant systematic differences, highlighting the need for further research to optimize SDCT for clinical use.