What question did this study set out to answer?

April 25, 2026Open Access

Correlation between bone marrow blood perfusion and bone mineral density in middle-aged and elderly population: a preliminary study based on IVIM-DWI

Key Points

To investigate the relationship between lumbar magnetic resonance perfusion parameters and volumetric bone mineral density in middle-aged and elderly individuals.
Analyzed 212 subjects categorized into normal, osteopenia, and osteoporosis groups using quantitative computer tomography.
Evaluated IVIM-DWI parameters including diffusion coefficient, pseudo-diffusion coefficient, and perfusion fraction.
Developed logistic regression models for predicting low bone mass based on clinical parameters and IVIM-DWI values.
In males, the perfusion fraction (f value) in the low bone mass group was significantly lower than in the normal group (P < 0.001).
In females, IVIM-DWI parameters D, D*, and f were significantly lower in the low bone mass group (P = 0.029, P = 0.020, and P = 0.002, respectively).
Combined prediction model showed AUCs of 0.684 for f value and 0.744 for the model in males, and 0.676 for f value and 0.831 for the model in females.

Abstract

To investigate the relationship between lumbar magnetic resonance perfusion parameters and volumetric bone mineral density (vBMD) in middle-aged and elderly people. A total of 212 subjects (104 males and 108 females) were categorized into normal, osteopenia, and osteoporosis groups based on quantitative computer tomography (QCT) for the lumbar spine. Differences in intravoxel incoherent motion diffusion weighted imaging (IVIM-DWI) parameters diffusion coefficient (D), pseudo-diffusion coefficient (D*), and perfusion fraction (f) and levels of sex hormone and bone turnover markers between normal and low bone mass (LBM, including osteopenia and osteoporosis) groups were analyzed. The areas under the receiver-operating characteristic curves (AUCs) for the clinical parameters, serum markers, and IVIM-DWI parameters selected by logistic regression analysis for prediction of LBM were calculated. In male group, f value in LBM group was significantly lower than that in normal group (P < 0.001). After adjusting for age, f value was weakly positively correlated with vBMD (r = 0.230, P = 0.020). Using logistic regression, a model (6.888 + 0.075age − 0.138BMI − 55.271f) was set up as a predictor for LBM. The AUCs of f value and the model for predicting LBM were 0.684 and 0.744, respectively. In female group, D, D*, and f values in LBM group were significantly lower than those in normal group (P = 0.029 for D value; P = 0.020 for D* value; and P = 0.002 for f value). After adjusting for age, f value was weakly positively correlated with vBMD (r = 0.241, P = 0.012). Using logistic regression, a model (-3.771 + 0.143age − 32.462f − 0.003estradiol + 1.639NTX) was set up as a predictor for LBM. The AUCs of f value and the model for predicting LBM were 0.676 and 0.831, respectively. IVIM-DWI parameters (especially f value) can reflect the changes of bone marrow blood perfusion for different genders with varying bone mass. The combined prediction model including f value and clinical information can effectively predict LBM. ChiCTR1900024511 (Registered 13-07-2019).

اسأل الذكاء الاصطناعي

Bookmark

View Full Paper