Volumetric Assessment of Bone Mineral Density and Bone-Architecture Using Spectral Photon-Counting CT

Abstract

The goal of this research is to explore the feasibility of the MARS Spectral Photon Counting CT (SPCCT) in assessing site-specific volumetric bone mineral density (BMD) and micro-architecture, which are vital for bone health assessment. Diagnostic imaging is essential for confirming or ruling out bone conditions. Current imaging techniques available in clinics face limitations in the quantitative and qualitative assessments of bone conditions such as osteoporosis and osteopenia. This study investigates the emerging potential of MARS SPCCT for clinical applications, specifically in providing accurate bone health assessments. In this study we used four bovine bone samples which were obtained from a commercial butcher. Samples were sectioned to f it the small bore of the MARS Microlab 5X120 scanner, equipped with Cadmium Zinc-Telluride (CZT) Medipix3RX detectors. Calibration data were collected using a customized QRM spectral CT calibration phantom that included varying concentrations of calcium hydroxyapatite (49.2, 102.4, 201.5, 406.9, and 809.8 mg/cm³), as well as lipid and water. Both the biological samples and the calibration phantom were reconstructed using the following predefined energy bins: 7-40, 40-50, 50-60, 60-79, 79-118 keV, with an 80 micron voxel size. Images for each energy bin were processed using an iterative reconstruction algorithm. Using the built-in MARS material decomposition algorithm, the reconstructed multi-energy CT images of bone samples were transformed into material-specific images showing concentrations of hydroxyapatite (a bone-like material), lipid, and water. The hydroxyapatite density images were segmented to study the cortical and trabecular bone compartments separately. Firstly, the segmentation was conducted using slice-by-slice manual contouring in ImageJ. Density and architectural parameters such as cortical BMD and thickness, trabecular BMD, trabecular thickness, and spacing were quantified using the BoneJ plugin. Secondly, an automated segmentation algorithm was developed using MATLAB® and validated against manual segmentation methods. The comparison between the manual and automated segmentation algorithm across all quantified parameters demonstrated good agreement, with percentage differences < 5% for density parameters and <10% for architectural parameters. A comparative analysis of bone parameters derived from MARS SPCCT was performed with Bruker micro-CT. The comparability of the density and architectural bone parameters obtained using MARS SPCCT with those from micro-CT showed variation, with percentage differences ranging from 3% to 11% for density parameters, and from 3% to 23% for architectural parameters. Our findings indicate that MARS SPCCT enables high-resolution, three-dimensional imaging, quantification, and differentiation among various tissue types such as fat, soft tissue, and bone. Notably, it distinguishes calcified areas of cortical and trabecular bone from bone marrow. This capability is instrumental in assessing the risks associated with osteoporosis and fractures, presenting a feasible alternative to the gold-standard DEXA assessments. Overall, this research establishes a foundation for the capabilities of the MARS SPCCT and sets the stage for future enhancements in bone quality assessment using this cutting-edge technology. Additionally, it outlines a methodology for future cross-modality studies of bone quality, including the handling, preservation, and sectioning of bone samples.

Date of Award	24 Jun 2024
Original language	American English
Supervisor	AAMIR Raja (Supervisor)