Bone Material Properties Distributions from quantitative Raman analysis: a distribution-based approach to human bone matrix characterization

Background: Bone mechanical competence depends not only on bone mass but also on the spatial organization of mineral and organic matrix properties. Whereas quantitative backscattered electron imaging (qBEI) characterizes mineral heterogeneity using Bone Mineral Density Distributions (BMDDs), no equivalent framework currently exists for describing the biochemical heterogeneity of the bone matrix by Raman microspectroscopy. Methods: We developed a quantitative Raman analysis (qRA) workflow to generate Bone Material Properties Distributions (BMPDs). Raman spectra were acquired along full-thickness line scans crossing successive bone structural units in human trabecular and cortical bone. Pixel-wise Raman-derived biochemical parameters describing collagen organization, advanced glycation end-products, glycosaminoglycans, mineral-to-matrix ratios, carbonate substitution, and mineral crystallinity were converted into normalized frequency distributions and modeled using gaussian functions. BMPDs were generated from undecalcified pMMA-embedded iliac biopsies (n=38) and routinely processed decalcified paraffin-embedded osteomedullary biopsies (n=21). The analytical framework was validated against conventional two-dimensional Raman mapping, and reproducibility, interindividual variability, and demographic influences were evaluated. Results: BMPDs were consistently well described by gaussian functions (typically R2 > 0.95), allowing extraction of three descriptors for each biochemical parameters: Mean, Peak, and Width. One-dimensional line scans showed excellent agreement with conventional two-dimensional Raman mapping while reducing acquisition time approximately 100-fold. Technical variability remained substantially lower than biological variability for all parameters, and a single trabecular line scan provided reliable estimates of BMPD descriptors for nearly all Raman-derived indices. Cortical bone exhibited narrower BMPDs than trabecular bone, indicating lower biochemical heterogeneity. Comparable BMPDs were obtained from undecalcified and routinely decalcified specimens, demonstrating robustness to tissue processing. No significant associations with age or sex were observed after correction for multiple testing. Reference BMPDs further enabled visualization and standardized quantification of deviations in representative pathological bone biopsies. Conclusions: Quantitative Raman analysis introduces BMPDs as a novel framework for assessing the spatial heterogeneity of bone extracellular matrix composition. By extending Raman microspectroscopy beyond conventional mean measurements, BMPDs provide robust, reproducible descriptors of bone material organization and establish a methodological foundation for investigating alterations of bone quality in metabolic bone diseases.