MRI-based biomarkers of bone marrow adaptation in anemia: A quantitative evaluation

Abstract Objective To evaluate anaemia-related alterations in lumbar bone marrow using apparent diffusion coefficient (ADC), fat fraction (FF), and signal intensity ratio (SIO), and to assess relationships with haematologic markers. Methods Ninety-three patients with abdominal MRI and available haemoglobin (Hb), red cell distribution width–coefficient of variation (RDW-CV), red cell distribution width–standard deviation (RDW-SD), and quantitative MRI parameters were included. Patients were classified into four Hb-based groups. Group differences were assessed using Kruskal–Wallis and Dunn–Bonferroni tests. Spearman correlation, age- and BMI-adjusted partial correlations and intraclass correlation coefficients (ICCs) were computed. Results Hb showed a positive correlation with FF (rho = 0.386, p < 0.001) and a negative correlation with ADC (rho = − 0.467, p < 0.001). RDW-CV correlated negatively with FF (rho = − 0.341, p = 0.001) and positively with ADC (rho = 0.300, p = 0.004). RDW-SD showed a weak positive correlation with ADC (rho = 0.208, p = 0.045). After age adjustment, RDW-SD correlated with FF (partial rho = − 0.350, p = 0.001). The initially borderline RDW-SD–ADC association became significant after age (partial rho = 0.279, p = 0.007) and BMI adjustment (partial rho = 0.219, p = 0.035). FF and ADC differed across groups ( p < 0.05). FF was lowest in Group 4 (all p < 0.05). ADC was higher in Group 4 than in Groups 1 and 2 ( p ≤ 0.001) and higher in Group 3 than Group 1 ( p = 0.035). All ICC values exceeded 0.90. Conclusion Quantitative MRI parameters reflect anaemia-related lumbar bone marrow microstructural alterations.