Axial fragility fractures are about 60% more common in chronic lymphocytic leukemia (CLL) than in age-matched peers, even when dual-energy X-ray absorptiometry (DXA) is normal. This finding implies that leukemic infiltration may silently remodel the bone-marrow niche long before trabecular mass is lost. Patients with CLL therefore face vertebral and pelvic fractures that current imaging fails to predict. This clinical gap highlights a need to investigate the impact of leukemic infiltration on the structural and cellular integrity of the bone-marrow microenvironment. To address this, we used high-resolution synchrotron micro-computed tomography (SRμCT) on 18 bone marrow core biopsies from 17 CLL patients, stratified into low (≤5%; Lo-CLL) and high (>5%; Hi-CLL) CLL involvement. Bone samples were collected during routine bone marrow biopsies, fixed in 70% ethanol, and stored at 4°C prior to scanning. Imaging was performed at a resolution of 1.6 μm, enabling precise visualization of both mineralized and soft tissue compartments. Consequently, we employed a combined image processing and deep learning segmentation pipeline, allowing for accurate quantification of bone, osteocyte lacunae, and adipose tissues. Segmentation outputs were manually quality-checked for each sample to ensure accuracy. From the validated masks, we quantitatively analyzed density and morphometric features of lacunae, adipose tissue, and bone. Group differences were tested with two-tailed t-tests, and associations with continuous CLL burden with Pearson correlation.

Included patients had a median age of 70 yr (63-85) and BMI 29.3 kg m⁻² (18.8-41.0); 61 % were male. The median interval from CLL diagnosis to biopsy was 3 yr (1-15). Since most of the Lo-CLL samples were collected during post-treatment remission, the median number of prior CLL therapies was higher in the Lo-CLL group [1 (range (1-6)] compared to the Hi-CLL group [0 (range 0-3)]. No patient received chemoimmunotherapy (CIT) immediately prior to sample collection.

Hi-CLL group showed a 40.3% reduction in osteocyte lacuna density (17 100 vs 25 664 #/mm³, p < 0.05) and a 101% increase in adipose surface-area-to-volume ratio (SA:V) (118.0 vs 58.7, p < 0.05), indicating extensive adipose tissue fragmentation. These changes occurred despite similar BMI values between groups, and additional analyses confirmed no significant differences in age or sex between groups. Beyond lacuna density, lacunae in Hi-CLL were 12 % larger, and lacunar volume correlated with both tissue mineral density (r = 0.50, p < 0.05) and adipose density (r = 0.70, p < 0.05), underscoring a coordinated, multi-tissue remodeling. Treating CLL infiltration as a continuous variable, lacuna density correlated negatively with CLL burden (r = –0.49, p < 0.05) whereas adipose SA:V correlated positively (r = 0.56, p < 0.05), and neither association was explained by demographic factors. One patient who underwent biopsy both before and after CLL-directed therapy with ibrutinib (BTKi) demonstrated restoration of osteocyte lacuna density and reduction in adipose fragmentation following treatment with a reduction in CLL involvement from 60% to <1%. This finding suggests that CLL-induced microstructural damage may be partially reversible.

Osteocytes play a central role in bone remodeling, coordinating responses to mechanical stress and regulating osteoblast and osteoclast activity. A decline in lacuna density implies either impaired osteoblast-to-osteocyte transition or increased osteocyte apoptosis, weakening bone microarchitecture. Inflammatory cytokines secreted by CLL may be driving this mechanism. Similarly, fragmentation of adipose tissue may reflect CLL-driven metabolic remodeling of the bone marrow niche. Based on preclinical studies of other blood cancers, oversecretion of GDF15 and TGF-β, could be contributing to this mechanism.

This is the first study to obtain sub-cellular and 3D metrics from both mineralized and soft-tissue compartments in human CLL bone marrow. By quantifying osteocyte loss, lacunar enlargement, and adipose tissue fragmentation that precedes any decline in DXA-measured density, we provide a mechanistic explanation for the unexplained fracture risk in CLL. The SRμCT-derived biomarkers, such as lacuna density, lacuna volume, and adipose SA:V, can identify at-risk patients, monitor treatment response, enrich microenvironment focused trials in CLL and other hematologic malignancies.

This content is only available as a PDF.
2025
Sign in via your Institution