Accuracy and Robustness of Bone Volume Fraction Assessment by Photon-Counting, Dual-Energy, and Quantitative CT using Micro-CT as Standard of Reference.

Abstract

We aimed to quantitatively compare different computed tomography (CT)-based bone volume fraction (BV/TV) measurements. We hypothesize that phantom-less measurement using virtual monochromatic images (VMI) reconstructed form dual-energy CT (DECT) or photon counting CT (PCT) is less affected by tissue variations in trabecular bone than quantitative CT (QCT) and that PCT allows measurements with a lower radiation dose. The BV/TV of bovine trabecular bone samples were measured using four CT scanning methods. Eight different measurements were compared using three soft tissue substitutes (air, saline solution, and fat) to investigate its effect on BV/TV estimation. For this purpose, the samples' bone marrow was removed and replaced with one of the three substitutes in succession for CT scanning. For QCT, a standard bone phantom was used to derive BV/TV form CT values. While DECT- and PCT-based measurements were based on a system of energy-dependent equations established by reconstructing VMIs for specific photon energies. T-test, ANOVA tests and pairwise comparison were performed using the micro-CT (µCT) measurements as reference standard. QCT showed a significant difference between the fat, saline solution, and air. DECT and especially PCT showed no differences between the substitutes. PCT showed no significant differences between radiation doses. Our results highlight the complexity of BV/TV measurements and emphasize the impact of the trabecular bone components on measurement accuracy. Despite these challenges, VMIs from "low" dose PCT provide a reliable alternative to standard QCT. They have the potential to improve the estimation of bone conditions, as well offering valuable insights for clinical, and forensic applications.