During skeletal tissue differentiation processes such as distraction osteogenesis, fracture healing, adaptation to skeletal implants, bone development and regeneration, the structure and composition of the mineralized bone tissue, referred to as woven bone or immature bone, continually change due to growth and responses to the tissue environment, including physical, chemical, and mechanical factors (Wehrle et al., 2014; Tourolle Né Betts et al., 2020; Mora-Macías et al., 2021; Blázquez-Carmona et al., 2021a; Paul et al., 2022). The organic and inorganic matrices of woven bone are highly dynamic and undergo alterations in their structure and composition (Isaksson et al., 2010; Turunen et al., 2011; Martínez-Reina et al., 2018; Reznikov et al., 2018). The inorganic phase is primarily composed of nonstoichiometric hydroxyapatite crystals. This mineral content is deposited in two phases: an initial phase in which 70% of the final mineral content is incorporated within a few weeks, followed by a second phase in which the mineral content increases over months and years (Roschger et al., 2020). The organic phase is predominantly composed of type I collagen, along with various biomacromolecules, including proteoglycans and non-collagenous proteins (Boskey and Robey, 2013). During the initial stages of healing, naïve collagen fibers are randomly oriented due to their rapid formation. As the maturation progresses, these fibers increase in density, reorient, and structurally mature (Blázquez-Carmona et al., 2022). This maturation is understood as the process in which the fibers are crosslinked and packaged to accommodate mineralization. These microstructure arrangements result in a time-dependent mechanical behavior of the tissue, with initial stiffness much lower than that of the matured cortical and trabecular bones.

The microstructure and composition of bone play crucial roles in its macroscopic structural and mechanical behavior. Their relationship has been extensively studied in cortical and trabecular bones (Schaffler and Burr, 1988; Rho et al., 1995; Sevostianov and Kachanov, 2000; Budyn et al., 2012). It has been shown that the elastic modulus of the mature bones strongly depends on mineral content, while toughness correlates with the quality of the collagen matrix (Zioupos and Currey, 1998). Correlations between the elastic modulus and the bone mineral density have been established to estimate mechanical properties from the image-based measurements of the bone mineral density at both the organ and tissue scales (Nobakhti and Shefelbine, 2018). At the organ level, the elastic modulus (E) is related to the apparent mineral density ρ , mineralized bone mass/bulk volume, through a power law relationship ( E ∝ ρ b ) . For cortical bones, b ranges between 4 and 7.4 (Schaffler and Burr, 1988; Currey, 1988) and for trabecular bones, it is normally lower, b = 1.27-2.57 (Rice et al., 1988; Keller, 1994; Bouxsein and Radloff, 1997), indicating lower apparent stiffness. One limitation of these models is that they do not distinguish the influence of bone volume fraction from ash fraction. Therefore, at the macro- level, these correlations are improved when porosity (Schaffler and Burr, 1988; Currey, 1988) or fabric orientation are included (Hodgskinson and Currey, 1993; Martin and Boardman, 1993; Nicholson et al., 1997). However, this relationship is weak at the tissue level, likely due to the effect of microstructural features at small length scales (Nobakhti and Shefelbine, 2018). In cortical bone, hierarchical micromechanical models have been also developed to explain how mechanical interactions between components at different observation scales regulate the tissue’s effective elastoplasticity (Fritsch and Hellmich, 2007; Hamed et al., 2010; Kumbolder et al., 2024). However, they traditionally assume invariant structural and mechanical properties of these elementary components, including the collagen and hydroxyapatite (Hellmich et al., 2022). This assumption is not valid for an evolving tissue like woven bone. In the woven bone, 6th-degree polynomial equations were proven to best fits the elastic modulus versus the ash fraction, predicted by a multiscale computational homogeneization model (García-Rodríguez and Martínez-Reina, 2017). In the in silico model, García-Rondríguez and Martínez-Reina (García-Rodríguez and Martínez-Reina, 2017) assumed isotropic orientation of collagen fibrils and used the ash fraction reported in the literature (Martínez-Reina et al., 2018). A simple exponential regression equation based on the ash fraction α , E = 12.88 ⋅ α 2.75 , was also provided with a significant correlation coefficient. To date, no previous work has proposed power laws based on experimental data to predict the apparent mechanical properties of woven bone at micro- or macro-scales considering the evolutionary structure of this tissue.

Various techniques have been developed to assess the bone inorganic matrix at macroscopic scale to measure bone mineral density or mineral content: computed tomography (Snyder and Schneider, 1991; Rho et al., 1995), micro-computed tomography, dual-energy X-ray absorptiometry (Hodgskinson and Currey, 1993; Bouxsein and Radloff, 1997), ashing the sample (Currey, 1988; Martínez-Reina et al., 2018) or weighing it (Schaffler and Burr, 1988; Martin and Boardman, 1993; Keller, 1994; Nicholson et al., 1997). To account for local variations in mineral content, scanning small angle X-ray scattering and wide angle X-ray scattering have been used to measure mineral crystal length and thickness (Roschger et al., 2020). Quantitative back-scattered electron microscopy is a validated method for determining spatially resolved calcium content (Roschger et al., 2020), and Fourier Transform infrared microspectroscopy and Raman microspectroscopy have been widely used to study the inorganic chemical compositional changes of the bone (Tarnowski et al., 2002; Akkus et al., 2004; McCreadie et al., 2006; Yerramshetty et al., 2006; Morris and Mandair, 2011), including mineralization, crystallinity, or carbonate substitution amongst others. For instance, investigating the inorganic chemical compositional changes during bone ageing using infrared and Raman microspectroscopy, it was concluded that Raman microspectroscopy is more sensitive for the inorganic matrix (Turunen et al., 2011). Bone healing processes have also been monitored with Raman spectroscopy (Uthgenant et al., 2007; Gamulin et al., 2013; Ahmed et al., 2018). Ahmed et al. (2018) studied early healing in calvarial defects, which heal spontaneously, with Raman Spectroscopy. They showed an increase in mineral/matrix and crystallinity ratios and a reduction in the carbonate/phosphate ratio after 14 days of healing. However, the long-term temporal evolution of the mechanical, chemical and morphological properties of woven bone during healing in critical-sized bone defects remains unknown. Beyond ash analysis, using Raman spectra or back-scattered electron signals as reliable input data for building power laws to predict the mechanical properties of bone tissue has not been explored to date.

Methods for evaluating the mechanical properties of bone tissue are more standardized. Macroscopic mechanical properties of cortical and trabecular bones have been traditionally evaluated through different mechanical testing including tension, compression, torsion, three-point bending, or buckling (Townsend et al., 1975; Ryan and Williams, 1989; Woo et al., 1991; Bayraktar et al., 2004; Beaupied et al., 2007). Local mechanics of mature tissue has been also assessed with ultrasound microscopy (Eriksen et al., 1994; Hengsberger et al., 2002; Bala et al., 2013), nanoindentation (Rho et al., 1995; 1999; Rho and Pharr, 1999; Zysset et al., 1999; Thurner, 2009) or atomic force microscopy (Asgari et al., 2019). However, for woven bone and its evolutionary properties, studies have used both in vivo (Dwyer et al., 1996; Aarnes et al., 2005; Mora-Macías et al., 2015a; b; Blázquez-Carmona et al., 2021a) and ex vivo approaches (Ohyama et al., 1994; Floerkemeier et al., 2010; Blázquez-Carmona et al., 2022) to provide values of local and apparent stiffness at different time-points during healing. Typically, in vivo mechanical properties are indirectly measured through instrumentation of surgically implanted fixations (Dwyer et al., 1996; Aarnes et al., 2005; Mora-Macías et al., 2015a; b; Blázquez-Carmona et al., 2021a). For ex vivo techniques, nanoindentation tests have proven valid to account for nanoscale heterogeneity and examine bone quality at a lower scale (Leong and Morgan, 2008; Manjubala et al., 2009; Mora-Macías et al., 2017; Mora-Macías et al., 2021). For instance, Manjubala et al. (2009) and Mora-Macías et al. (2017) measured the nanoindentation modulus for woven bone during fracture healing and bone transport processes, respectively.

In silico models are valuable for the understanding the course of healing from a mechanobiological perspective. Many researchers have sought to establish relationships between the mechanical properties of undifferentiated tissue and ultimate tissue phenotype formed in a wide variety of bone regeneration processes, such as fracture healing (Borgiani et al., 2017; Ghiasi et al., 2017), distraction osteogenesis (Isaksson et al., 2007; Reina-Romo et al., 2011; 2012) or tissue engineering (Byrne et al., 2007; Stops et al., 2010), amongst others. As a limitation, all these works develop phenomenological rules based on empirical observations and input parameters that are not sufficiently validated with experimental data, leaving room for improvement in accuracy. In particular, most of these models assumed constant values for the elastic modulus and porosity of the woven bone, regardless of its mineralization state or degree of matrix organization, which is not explicitly modeled. A constant porosity value of 80% and an elastic modulus of 1,000 MPa are typically assigned to the woven bone in most of these mechanobiological evolutionary models (Lacroix and Prendergast, 2002; Isaksson et al., 2007; Burke and Kelly, 2012).

The main aim of this study is to identify time-related changes at the macro- and micro-scales at the chemical, mechanical, and morphological properties of the woven bone during healing of a critical-size bone defect. There is a need for a better understanding of the woven bone micromechanics and the development of a reliable mechanical model of this tissue to improve the prediction of its stiffness. For this purpose, at the microscopic scale, micromechanical properties will be measured with nanoindentation; mineral and organic characteristics will be quantified through Raman spectroscopy; and structural features through micro-CT. At the macroscopic scale, ash and element analyses will be assessed together with a numerical model to reproduce the woven bone bulk mechanical behavior based on physical measurements. All these experimental data will be used to build power laws relating the elastic modulus with the structural and chemical properties of the woven bone for the first time.

Quantifying the structural, chemical, and mechanical properties of the woven bone is crucial for understanding the mechanobiology behind bone regeneration and assessing the effectiveness of ortophaedic approaches for healing critical-sized bone defects. Equally important is the parallel development of numerical models that optimize key engineering parameters to control this biological process and prevent non-unions. There is a notable lack of micromechanical models of woven bone in the literature based on ex vivo experimental data that consider changes in tissue properties at different regeneration phases. The multiscale research presented here analyzed the evolution of the trabeculae organization in this immature tissue through micro-CT, its elastic modulus using nanoindentation, its mineral quality via Raman spectroscopy, and its chemical composition with ash analysis. From these experiments, power laws based on physical measurements were defined for the first time, relating the elastic modulus of the woven bone with its porosity and mineralization state. In addition, the data and power laws provided in this study are highly extrapolated to human bone tissues due to the ovine nature of the samples. High similarities in bone composition were reported between mature sheep and humans in the literature (Newman et al., 1995). Similarly, comparable metabolic and bone remodeling rates were found using dual energy X-ray absorptiometry (den Boer et al., 1999). The results obtained in this study also try to address some open questions in the field: Why is the stiffness of woven bone lower than that of cortical bone? Is it solely due to the mineral content, or is it influenced by the structural factors as well? Do structural and mineral changes equally contribute to the bulk mechanical properties of woven bone? Are Raman spectra valid inputs for defining power laws?

The mechanical properties of the woven bone tissue exhibited a slight increase over the analyzed regeneration period, as indicated by the elastic modulus measured by nanoindentation, which ranged from 3.07 to 9.25 GPa (Figures 3A, B; Supplementary Figure S1). However, the stiffness levels characteristic of cortical tissue, approximately 16.44 ± 2.79 GPa, were not reached even after 161 days post-surgery. Previous studies also used nanoindentation to measure the elastic modulus of bone calluses formed during other regeneration processes (Leong and Morgan, 2008; Manjubala et al., 2009; Mora-Macías et al., 2017). Manjubala et al. (2009) reported values in the range of 2–13 GPa over the first 9 weeks of a fracture healing experiment in an ovine tibia. Thus, the woven tissue associated to this regeneration process reached higher stiffness at a greater speed, likely due to the smaller size of the bone gap in fracture healing, approximately 3 mm in this study. Similarly, Leong and Morgan (Leong and Morgan, 2008) conducted indentation tests on rodent fracture healing calluses, reporting an elastic modulus ranging from 0.03 to 1.01 GPa at day 35 after fracture. These results are considerably lower than those observed in the current study, possibly due to the different animal model used and the earlier healing stage analyzed. In another study, Mora-Macías et al. (2017) also measured elastic modulus values ranging from 7 to 14 GPa in bone transport calluses of the ovine metatarsus with the same critical size of 15 mm. Specifically, they reported an elastic modulus of 11 GPa after 161 days in bone transport (Mora-Macías et al., 2017), closely matching the 9.25 ± 2.93 GPa measured in this study during bone lengthening at the same time-point. The possible lower degree of mechanical maturation in the bone lengthening callus could be explained by a lower mechanical stimulation associated to the less mobility and bearing capacity of the specimens after the indirect elongation of the surrounding soft tissues (Blázquez-Carmona et al., 2021b). These findings reflect a slower recovery of the mechanical properties of the woven bone at the tissue scale throughout its remodeling phase. It is important to highlight that nanoindentation tests specifically excluded the influence of inter-trabecular macroporosities on the average data provided. Consequently, the observed lower stiffness can only be attributed to the presence of microporosities or other microstructural features at smaller length scales, or to differences in the chemical composition and organization of the deposited mineral phase, as discussed later.

The mineral content in woven bone has been previously examinated in the literature. Manjubala et al. (2009) measured calcium content directly from back-scattered electron microscopic images (2D mineral content), showing inhomogeneity in the callus mineralization with a maximum Ca content of 20% of Ca that saturates at week 9 after surgery. In cortical bone, Turunen et al. (2011) investigated the temporal evolution of mineral content during aging with Infrared and Raman spectroscopy. They found that while mineralization increases over time, hydroxyapatite crystals mature more slowly. Ahmed et al. (2018) reported an increase in the mineral-to-matrix and crystallinity ratios during the first 14 days of healing in calvarial defects through Raman Spectroscopy. In our study, bone mineralization during healing was quantified at the microscopic scale using several Raman parameters, including mineral-to-matrix and carbonate-to-matrix ratios. Figure 3 demonstrates the positive correlation between the Tyrosine and Amide III mineral-to-matrix and healing time, reaching values comparable to those of cortical bone at day 161. These measurements provide a quantitative assessment of the extent of bone mineralization, directly related to ash weight (Gourion-Arsiquaud et al., 2009). Crystallinity, which is a measure of the crystal size development (Turunen et al., 2011), remained nearly constant during the consolidation phase, with values similar to those of the cortical bone (Figure 3C). Given that osteoid is deposited by osteoblasts at different locations and timing within the callus, this finding suggests that the maximum crystal size is obtained in an early stage of regeneration. Despite the increase in the carbonate-to-matrix ratio (Figure 3H), the carbonate-to-phosphate ratio (Figure 3D) showed no trend during the consolidation phase, maintaining values akin to those of cortical bone. Similar results were reported in calvarian healing (Ahmed et al., 2018). This ratio represents type-B carbonate substitution into hydroxyapatite, a process not fully understood but associated with increased solubility and decreased mechanical performance of the tissue (Pan and Darvell, 2010). Literature suggests this substitution is more related to aging, as carbonate peak areas and intensities increase with age (Boskey and Robey, 2013; Turunen et al., 2013). Regarding the hydroxyproline-to-proline ratio, the literature proved that it negatively correlates with the maturation of the tissue (Shah et al., 2019). However, it keeps constant in woven bone and within the standard deviation range of cortical bone (Figure 3G). A similar trend was observed for the Amide I-to-Amide III ratio, which remained approximately constant above 1 (Figure 3J). It follows a very similar trend to crystallinity, as both processes are closely correlated with the bone formation. Initially, naïve collagen fibers are created followed by an enzymatically crosslinking (Knott and Bailey, 1998). These fibers then mineralized and served as scaffolds for the nucleation and growth of the additional mineral crystals (Gourion-Arsiquaud et al., 2009). A high Amide I-to-Amide III ratio indicates a predominance of functional and mature collagen crosslinks, whereas a predominance of Amide III suggests high activity of nonfunctional precursor collagen with reducible crosslinks (Coelho et al., 2014; Mandair and Morris, 2015; Fraulob et al., 2020). Thus, woven bone exhibits an increase in procollagen molecules and immature crosslinks during the analyzed regeneration phase.

Figure 4 also reported the chemical analysis performed by ash analysis and elemental study. The ash fraction, phosphorous and calcium content increased nonlinearly with healing time with a plateau nearly the cortical reference values. Phosphorous and calcium contents are also in line with the infrared and X-ray analysis on natural hydroxyapatite (Janicki et al., 2012). Numerous substitutors were reported in hydroxyapatite crystals in the literature, including carbon, potassium, magnesium, sodium, or strontium (Bergstrom and Wallace, 1954; Buddhachat et al., 2016; Katzenberg, 2020). For potassium and sodium, the decrease with time was quantified (Figures 4D, F). However, none of the above substitutes seems to play a key role during the bone regeneration process.

Regarding the structural features of the woven bone at the trabecular scale, as measured by micro-CT, the BV/TV increased up to 0.69. This increase is primarily due to an increased in the trabecuale thickness (Tb.Th) as the different mineralization fronts advance (Figure 2). Previous studies have already examined bone volume fraction using different techniques and length scales. For instance, Blázquez-Carmona et al. (2021a) and Mora-Macías et al. (2021) measured the bone volume fraction from CT macroscopic images in bone lengthening and bone transport, respectively. In these studies, the volume fraction ranged from 88.18% to 93.97% in bone lengthening and 87.55%–100% in bone transport at days 121–161 after surgery. The lower values in our study are more consistent with estimates reported in bone transport calluses (Mora-Macías et al., 2021), which used direct segmentation of 2D micrographs based on x-ray grey-scale images. Thus, the porosity at the trabecular scale appears to significantly impact the estimation of the bulk volume fraction. This structural feature likely contributes to the differences found in the same period of time between the micro-scale elastic modulus measured in this study (3.07–9.25 GPa), and the lower apparent elastic modulus measured in vivo by Blázquez-Carmona et al. (2021a) (0.01–8.53 GPa), input in a later discussed power law. Compared to the trabecular tissue, woven bone only exhibits BV/TV levels similar to those observed in this study or those reported in femoral bone at the early stages of regeneration (Mittra et al., 2005; Teo et al., 2007). The other three-dimensional microarchitectural parameters of woven bone (Figure 2) could not be directly compared to other regenerating tissues due to the lack of reported data in the literature, to our knowledge. Therefore, comparisons are limited to cancellous bone tissue data from this study or literature sources (Mittra et al., 2005; Teo et al., 2007; Turunen et al., 2013). The aforementioned increase in Tb.Th above cancellous values occurs alongside a progressive reduction of the number (Tb.Nm) and interconnectivity (Conn.D) of woven trabeculae, more similar to that of healthy metatarsal trabecular tissue, 3.26 ± 0.87 c m − 1 and 12.25 ± 3.96, respectively. These reference trabecular values are slightly higher than those of healthy ovine or human femoral trabecular tissue reported in the literature (Mittra et al., 2005; Teo et al., 2007; Turunen et al., 2013), probably due to physiological differences between bones. The stabilization of the woven trabecular separation around 0.13–0.17 mm (Figure 2C) contrasts with the constant increase in Tb.Th, which can only be explained by the remodeling of already mineralized tissue in the medullary cavity or the outer interzones of the callus farthest from the original cortical bone. The degree of anisotropy (DA) remained similar to that of trabecular tissue in the same bone (Figure 2F) or to values reported in human femoral bone by Turunen et al. (2013), Finally, SMI evolved negatively from around 0, typical for trabecular bone (Mittra et al., 2005; Turunen et al., 2013), indicating a transition from a rod-and-plate structure typical of trabecular bone to a denser structure characterized by void spaces (Hildebrand and Rüegsegger, 1997). These results highlight the dynamic evolution of woven bone microstructure, from numerous disorganized thin trabeculae at the beginning of the consolidation phase to a structure with fewer but thicker and better-interconnected trabeculae.

The previous experimental data facilitated the definition of the power laws presented in Table 2, which serve as valuable tools for predicting the elastic modulus of woven bone in future in silico models at both macro- and micro-scales. The varied laws allow for the assignment of woven bone tissue mechanical properties based on different experimental results, not necessarily mechanical testing. Except for the current study, woven bone correlations have never been established using mechanical, structural and chemical parameters measured directly from ex vivo samples. The exponents of these expressions ( b and c fitting constants in Table 2) provide insights into the degree of influence of chemical composition (measured at the macro- scale by ash and elemental analyses and at the micro-scale by Raman spectroscopy) and microstructural morphology on the mechanical properties of the woven bone at both macro- and micro-scale. For instance, the ash fraction appears to have a lower relative influence on the elastic modulus measured by nanoindentation ( b = 1.26 ) over a broad range of ash fraction values compared to the calcium or phosphorous contents ( b = 2.62 − 2.86 ) . This exponent is also slightly lower than those in equivalent power laws reported in the literature (Schaffler and Burr, 1988; Currey, 1988; Hernandez et al., 2001). The differences can be attributed to the use of compact and cancellous bone tissues from different mammalian species, the range of ash fractions analyzed, and the different loading conditions in the mechanical characterization. The degree of significance of all correlation (R-square in range of 0.55–0.90 and most of p-values below 0.05) is consistent with previous works (Schaffler and Burr, 1988; Currey, 1988; Hernandez et al., 2001). However, Raman parameters, Amide III mineral-to-matrix and Tyrosine mineral-to-matrix, were found to correlate more significantly with elastic modulus than ash, calcium or phosphorus fractions (Figures 3A–C, E, J; Table 2). We hypothesize it could also be due to significant chemical heterogeneity in the tissue. Note that the Raman spectra were taken in the same areas of woven bone as those measured by nanoindentation. We also define a two-parameter function to explain more of the variance in the bulk mechanical properties of the woven bone than single-parameter functions based solely on chemical factors. Thus, the combination of ash fraction with the evolution of the BV/TV provides a significant correlation with the apparent elastic modulus (Figure 3D; Table 2), representing mechanical behavior at a macro-scale. In this case, the substantial difference between the ash fraction ( b = 3.88 ) and the BV/TV exponent values ( c = 9.98 ) indicates a greater influence of the bone volume evolution over the total callus volume, encompassing both mineralized and non-mineralized phases, on the bone callus’ apparent mechanical properties during healing, compared to the influence of the ash fraction. This finding contrasts with previous reports on cortical and trabecular tissue in the literature, where ash fraction seems to have a significant greater influence on bone strength and modulus (Hernandez et al., 2001), possibly due to the less structural remodeling and time depending nature of healthy bone tissues.

There are several limitations associated with the methodology employed in this study. Firstly, it is expected that the mechanical, structural and chemical parameters analyzed ex vivo are not significantly influenced by differences in bone maturity of sheep during the regeneration period in vivo since the longest experiment (161 days) represents only around 4% of the animal life cycle. Secondly, the methods used for the indentation test (Oliver and Pharr, 1992) assume linear elastic behavior in the tissue, despite the possibility that the woven bone could exhibit viscoelastic behavior in some callus interzones, particularly during early stage of mineralization. However, the mechanical characterization performed in this study is focused on the mineralized phase, where the influence of the viscoelastic behavior should not be significant compared to a soft tissue (Blázquez-Carmona et al., 2021b). Using the Oliver-Pharr method in the evaluation of the nanoindentations also presents several limitations. The elastic modulus of surrounding tissue sub-domains, along with the presence of microcracks or defects, may influence the homogenized and averaged elastic modulus measured by the indenter (Kariem et al., 2015; Pastrama et al., 2018; Königsberger et al., 2022). Additionally, the use of different specimens for each time-point introduces an inevitable source of variation in the measurements. Despite this, the results appear to follow logical trends, and the power laws derived are significant. Finally, a higher number of animals would have enhanced the statistical significance of the conclusions of this study. As previously mentioned, the authors prioritized the use of an ovine animal model, whose results are crucial for construction of more realistic micromechanical models and whose conclusions can be readily extrapolated to human clinical cases.

In conclusion, this work contributes to the development of novel micromechanical models applicable to numerical simulations of bone regeneration processes. It also enhances the understanding of the relationship between chemical composition, structural morphology, and the mechanical properties of woven bone at different length scales by combining multiple ex vivo techniques: micro-CT, nanoindentation, Raman spectroscopy, and chemical and elemental analysis. Addressing the questions posed at the beginning of the discussion, woven bone exhibited a significant increase in bone volume, primarily controlled by the thickening and unification of trabeculae, alongside a simultaneous increase in apparent ash fraction, calcium, and phosphate content. The power law established for the bulk elastic modulus highlights a more significant influence of the evolving trabecular microarchitecture on the apparent mechanical behavior of the bone callus. At the tissue scale, the Raman mineral-to-matrix ratio increased over the analyzed regeneration period, reaching values and crystallinity levels comparable to those of cortical bone. However, the local elastic modulus did not increase sufficiently to reach cortical values, potentially due to the role of microporosities within the trabeculae. Power laws predicting the evolution of the woven bone elastic modulus at this scale demonstrated that Raman parameters correlate better with the elastic modulus than apparent mineral content measured through ash analysis and elemental studies. Therefore, we advocate for the combined use of Raman spectroscopy and nanoindentation in constructing future power laws for regenerating tissue, potentially strengthened by including a second parameter to account for structural features at the microscale.