Dual-Energy Computed Tomography-Based Quantitative Bone Marrow Imaging in Non-Hematooncological Subjects: Associations with Age, Gender and Other Variables

Background: Our aim is to assess the utility and associations of quantitative bone marrow attenuation (BMA) values measured on clinical dual-energy computed tomography (DECT) exams in non-hematooncologic subjects with skeletal regions, patient age, gender, and other clinical variables. Methods: Our local ethics committee approved this retrospective image data analysis. Between July 2019 and July 2021, 332 eligible patients (mean age, 64 +/- 18 years; female, 135) were identified. Inclusion criteria were the availability of a standardized abdominopelvic DECT data set acquired on the same scanner with identical protocol. Eleven regions-of-interest were placed in the T11-L5 vertebral bodies, dorsal iliac crests, and femur necks. Patient age, gender, weight, clinical, habitual variables, inflammation markers, and anemia were documented in all cases. Results: Multi-regression analyses (all, p < 0.05) identified age as the strongest predictor of lumbar BMA (standardized coefficient: beta = -0.74), followed by CRP (beta = 0.11), LDH (beta = 0.11), and gender (beta = -0.10). In the lower thoracic spine, age was the strongest predictor (beta = -0.58) of BMA, followed by gender (beta = -0.09) and LDH (beta = 0.12). In femoral bones, age was negatively predictive of BMA (beta = -0.12), whereas LDH and anemia were positively predictive (beta = 0.16 both). Heart insufficiency significantly decreased (beta = 0.12, p = 0.034) a BMA value gradient from higher to lower HU values along the vertebrae T11 and L5, whereas age significantly increased this gradient (beta = -0.2, p <= 0.001). Conclusions: DECT-based BMA measurements can be obtained from clinical CT exams. BMA values are negatively associated with patient age and influenced by gender, anemia, and inflammatory markers.