Microvascular anatomy and mapping of intraosseous arterioles in the human L1 vertebral body: a histological study, literature comparison, and relevance for fracture healing

Introduction

The vertebral body, together with the intervertebral discs, constitutes the primary anatomical structure tasked with bearing, distributing, and modulating anteriorly directed axial loads. This complex plays a fundamental role in preserving the biomechanical equilibrium of the spinal column. In upright bipedal posture, the vertebral bodies are primarily exposed to significant compressive stresses, while the posterior spinal elements experience tensile forces and are critical for controlling and stabilizing rotational dynamics. The interplay between these opposing forces ensures the functional integrity of the spinal motion segment. Disruption or alteration of any of these components can lead to biomechanical imbalance and compromise spinal stability. Anatomically, the vertebral body—structurally supported by the adjacent intervertebral discs—transmits the majority of compressive axial load, accounting for approximately 80% of the force distribution along the spinal axis. Traumatic events or structural deformities, particularly those resulting in vertebral body compromise such as kyphotic collapse, significantly affect the mechanical behavior and overall function of the spinal unit. Flattening of the vertebral body leads to an anterior shift in the axis of load transmission, resulting in increased compressive stress on the already compromised vertebral structure, while concurrently intensifying tensile forces on the posterior spinal elements. This alteration disrupts the biomechanical equilibrium of the spinal segment, progressively subjecting adjacent vertebrae and posterior structures to abnormal mechanical loads. Over time, the sustained imbalance contributes to functional deterioration and frequently manifests clinically through pain or other symptomatic presentations associated with structural overload (1-12). In light of these considerations, there is growing support for the anatomical restoration of vertebral body fractures, aiming to closely recreate the original vertebral architecture. This restorative approach encompasses the correction of angular deformities—particularly kyphotic misalignments—the reestablishment of vertebral body height, and the reconstruction of endplate morphology. These objectives align with the broader principles of anatomical restoration commonly applied in the treatment of other synovial joints. The overarching goal is to recover the spine’s physiological load-bearing capacity and reinstate its biomechanical integrity and functional stability (13-22).

The incidence of post-traumatic necrosis of the vertebral body appears to be increasing, a trend likely influenced by the aging population. This condition is most frequently identified at the thoracolumbar junction, particularly involving T12 and L1, and predominantly affects elderly patients, especially those with underlying osteoporosis. Despite its clinical significance, post-traumatic necrosis of the vertebral body remains substantially underdiagnosed. Current literature estimates its prevalence in approximately 7% to 37% of vertebral compression fractures. It is more commonly associated with highly comminuted fractures, severe vertebral body collapse, and injuries in anatomically predisposed regions with limited vascular supply. These factors contribute to compromised bone healing and represent key risk indicators for delayed union, non-union, and the eventual development of vertebral pseudarthrosis. Posttraumatic necrosis signifies a pathological interruption in the normal reparative process of vertebral bone and continues to be one of the most challenging and unpredictable complications encountered in the context of spinal trauma management (23-30). Although definitive scientific evidence remains lacking, the prevailing hypothesis suggests that the progression of vertebral fractures toward non-union is primarily attributed to trauma-induced disruption of the intraosseous and/or extraosseous arterial network supplying the vertebral body. This vascular compromise leads to insufficient perfusion of the osseous tissue, resulting in ischemia that delays—or in some cases completely inhibits—the bone healing process. Consequently, such an environment becomes favorable to the development of osteonecrosis, non-union, and, ultimately, pseudarthrosis. An intact and functional vascular supply is critical for successful bone regeneration, as it ensures the targeted delivery of essential biological components to the injury site. These include oxygen, growth factors, immune and inflammatory cells, nutrients, minerals, cytokines, collagen, and various hormones—all of which play indispensable roles in coordinating the multifaceted cellular and molecular events required for effective bone repair (23-28,31-36). Nevertheless, to date, no diagnostic modality exists that can reliably identify or predict, from a biological and vascular standpoint, whether a given vertebral body fracture pattern has compromised essential vascular structures and is therefore at risk of progressing to non-union. As a result, vascular factors have not yet been systematically incorporated into the clinical decision-making algorithms for the management of thoracolumbar fractures (29,30). For this reason, the decision to either allow time for bone healing in vertebral fractures with a favorable prognosis for union or to proceed immediately with vertebral body replacement in fractures unlikely to heal—whether by interior augmentation or corpectomy and total replacement—remains unclear and lacks consensus (29,30). As a result, current treatment strategies for vertebral fractures are primarily guided by the mechanical characteristics of the injury. The decision to pursue vertebral body replacement ad initium in fractures deemed unlikely to heal spontaneously remains largely reliant on individual clinical judgment, as well as the experience and preferences of the treating surgeon and institution. However, this approach may be insufficient, given the critical role of vascular supply in the bone healing process. Integrating this essential biological perspective into the clinical decision-making framework has the potential to improve treatment quality—particularly by informing and standardizing early decisions regarding vertebral body replacement in selected cases. Such a strategy may offer a more effective means of preventing non-union and the development of post-traumatic vertebral necrosis (29,30,36).

Beyond its apparent important role in bone healing, the vertebral body’s vascular network may have broader significance in the pathophysiology of various spinal disorders, including vertebral fracture healing; dissemination of metastatic cells and septic emboli; the pathophysiology of vertebral osteomyelitis; the development of vascular tumors such as hemangiomas; the progression of vertebral osteoporosis; the pathogenesis of degenerative disc disease; and the emergence of congenital spinal deformities—such as hemivertebrae and kyphoscoliosis—through disruption of normal vertebral development and ossification, among others (36-43).

To effectively discuss, investigate, and deepen the understanding of these topics, a detailed knowledge of the arterial vascularization of the vertebral bodies is essential. Therefore, we present a summary of the scientific evidence to date regarding the arterial anatomy of the lumbar vertebral body. This vascular anatomy has historically been documented by a limited number of anatomical studies of the spine, which nonetheless provide a critical foundation for both experimental research and clinical investigation into the role of vascular injury in spinal pathologies. The majority of seminal research on spinal vascularization and development was conducted during the nineteenth century, utilizing methods such as anatomical dissection, arteriography, diaphanization, and vascular corrosion casting. Early descriptions of vertebral body vascular anatomy were closely associated with studies of the venous system, notably those by Gilbert Breschet in 1819 and later Oscar Batson, who acknowledged earlier foundational work by anatomists such as Guillaume Dupuytren, François Chaussier, and Vidus Vidius, particularly regarding the basivertebral veins (37,44-48). In this context, the venous vascular anatomy of the adult lumbar vertebral body, both extraosseous and intraosseous, has been comprehensively described and is generally accepted within the scientific community (44-46). In turn, the anatomical description of the vertebral body extraosseous arteries (Figure 1) is also well-established and widely accepted worldwide, showing only minor variations since its original characterization by Antoine Portal in 1803 (49-52). However, research on the intraosseous arterial vascularization of the lumbar vertebral body remains limited, with only a few studies addressing this topic—most of which show considerable variability in anatomical detail (37,38,49-66). Some studies primarily focus on describing the extraosseous arterial anatomy, offering only limited insight into intraosseous circulation. In contrast, others provide a more comprehensive characterization of the arterial network within the vertebral body. Notably, most early investigations into vertebral vascularization relied on angiographic techniques involving barium contrast injection. However, these methods were unable to visualize the fine arteriolar networks due to the high radiodensity of the contrast material (37,38,49-66). More recent studies have begun employing selective microarteriography, often in combination with diaphanization and corrosion casting techniques, which have enabled the visualization of smaller-caliber arterioles. Nevertheless, the assessment of intraosseous microvasculature remains challenging, particularly for vessels with a diameter smaller than 100 µm (38,59,63,67). In light of the limited scientific evidence available, the anatomy of the intraosseous arterial microvasculature of the adult lumbar vertebral body remains a topic of ongoing debate among anatomists. One of the central controversies concerns the primary source of arterial supply—specifically, whether the vertebral body receives most of its blood flow from posterior or anterior arterial branches (37,38,49-66,68,69). Currently, the most widely accepted and comprehensive anatomical descriptions of the intraosseous arteries of the adult lumbar vertebral body are those provided by Jorge Draper Mineiro [1965], Markhashov [1965], Henry Crock and Hiroshi Yoshizawa (1976 and 1977), and J. F. Ratcliffe [1980] (37,38,50,58). In addition to offering detailed anatomical descriptions, these authors defined angiosomes within the vertebral body by segmenting the intraosseous arterial vascularization according to the specific regions they supply (Table 1) (37,38,50,58). For the purpose of standardization, this article adopts the descriptive nomenclature established by J. F. Ratcliffe (38). A concise anatomical overview of the intraosseous arterial microvasculature of the lumbar vertebral body—based on this nomenclature—is presented below and originally illustrated by the authors of the present paper in Figure 2 (38). The lumbar segmental arteries (La), originating from the aorta (Ao), extend bilaterally and course posteriorly along the midportion of the anterolateral surfaces of the lumbar vertebral body. Along this course, the lumbar arteries typically give rise to approximately 10–20 ascending and descending branches, known as primary periosteal arteries (Ppa). These vessels further generate secondary periosteal branches that extend horizontally and penetrate the anterolateral surfaces of the vertebral body. According to their distribution within the vertebral body, three categories of intraosseous arteries can be identified: short peripheral arteries (Pa), which supply the outer third of the vertebral body; anterolateral equatorial arteries (Alea), responsible for vascularizing the central third; and metaphyseal arteries (Ma), which irrigate the intermediate zone between the regions supplied by the other two arterial types (38). At the level of the intervertebral foramina, the lumbar arteries divide into three principal branches: the medial spinal branch, the lateral branch, and the dorsal muscular branch. With regard to vertebral body vascularization, the medial spinal branch further gives rise to the anterior spinal canal branch, the posterior spinal canal branch, and the radicular branch. Upon entering the intervertebral foramen, the anterior spinal canal branch bifurcates into ascending and descending divisions, which form anastomoses within the vertebral canal with corresponding branches from adjacent levels, as well as with the contralateral homologous artery via intervertebral anastomoses, thereby creating retrocorporeal H-shaped arterial arcades. From the transverse anastomosis between the two anterior spinal canal branches, two posterior nutrient arteries—also referred to as basivertebral or nutrient arteries (Na)—originate. These vessels enter the vertebral body through the basivertebral foramen and supply its posterior central region (Figures 1,2) (38).

Figure 1 Extraosseous arterial vascularization of the lumbar vertebral body, showing the collateral circulation of the anterior third versus the posterior two-thirds of the vertebral body—anterolateral versus posterior surfaces. Sagittal view (left picture)—note that the blood supply entering through the anterolateral surfaces originates mainly from the lumbar artery, which runs centrally along each vertebral body (unisegmental arterial origin), and is supplied by small, short, terminal penetrating arteries. Posterior view (right picture)—note that the blood supply entering the posterior surface has a bisegmental origin, arising from the ascending and descending branches of the anterior spinal canal branch at the level of the intervertebral foramina. These branches anastomose with those from the adjacent levels, resulting in arterial vascularization of the posterior portion of the vertebral body being supplied by four arteries: two from the supradjacent intervertebral space and two from the infradjacent space. Ab, ascending branch; Ao, aorta; Ascb, anterior spinal canal branch; Db, descending branch; Ia-Ascb, intervertebral anastomose Ascb; La, lumbar artery; MAn, metaphyseal anastomosis; Na, nutritive artery; Ppa, primary periosteal artery; Va, vertical anastomoses at the disc level connecting the horizontal metaphyseal anastomoses of adjacent vertebrae.

Figure 2 Intraosseous arterial vascularization of the lumbar vertebral body. Currently accepted pattern of intraosseous arterial vascularization of lumbar vertebral body, shown in coronal, mid-sagittal, and axial (intermediate axial and metaphyseal axial) views. Alea, anterolateral equatorial artery; Alea-Asc, ascending branch of the anterolateral equatorial artery; Alea-Dsc, descending branch of the anterolateral equatorial artery; Ao, aorta; La, lumbar artery; Ma, metaphyseal artery; MAn, metaphyseal anastomosis; Na, nutritive artery; Na-Asc, ascending branch of the nutritive artery; Na-Dsc, descending branch of the nutritive artery; Pa, peripheral artery.

The present study aims to conduct a detailed anatomical investigation of the intraosseous arterial microvasculature of the L1 vertebral body—one of the vertebrae most commonly affected by burst fractures and post-traumatic vertebral necrosis—through a comprehensive, sequential histological analysis of adult vertebrae, with the objective of complementing, refining, and comparing our findings to the existing anatomical literature (23,24,30). We present this article in accordance with the STROBE reporting checklist (available at https://jss.amegroups.com/article/view/10.21037/jss-2025-aw-179/rc).

Methods

Three L1 vertebrae were harvested using an anterior retroperitoneal approach from fresh adult Caucasian human cadavers donated to Nova Medical School, Lisbon, Portugal. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by institutional ethics committees of Nova Medical School, Lisbon, Portugal (No. 135/2019/CEFCM) and Faculty of Medicine, University of Coimbra, Coimbra, Portugal (No. CE-141/2023/FMUC). The specimens included one male aged 61 years and two females aged 73 and 75 years, respectively. None of the donors had any known spinal pathologies, bone metabolic disorders, chronic inflammatory diseases, systemic vascular conditions, hematological disorders, bone or disc infections, bone neoplasms or metastases, or advanced systemic diseases that could potentially affect bone metabolism. The L1 vertebral bodies were fixed in 10% paraformaldehyde and decalcified using an 8% hydrochloric acid/formic acid solution. In order to accommodate the size limitations of histological slides, each axial section of the vertebral body was divided into four quadrants (Q): anterior right (ARQ), anterior left (ALQ), posterior right (PRQ), and posterior left (PLQ) (Figure 3). Accordingly, the vertebral bodies underwent sequential axial sectioning throughout their entire height, from the upper to the lower endplate, at 300-micrometer intervals. This process yielded upper, intermediate, and lower slices of each quadrant, enabling a comprehensive anatomical study of the entire vertebral body structure along its full height (Figure 3). Thus, each histological slide corresponded to a single quadrant of an axial section of the vertebral body. Subsequently, the sections were prepared for histological examination using hematoxylin-eosin and Masson’s trichrome stains. Immunohistochemical analysis was additionally conducted using CD31 staining to visualize endothelial structures (70-72).

Figure 3 Anterior view of human L1 vertebral body—upper, intermediate and lower slices and their division in quadrants and areas: ARQ, ALQ, PRQ, and PLQ, CA, IA and PA. VB—anterosuperior and left view of the L1 vertebral body corresponding to the anatomical specimen after its resection from the human cadaver; VB-Q—superior view of an axial slice of the vertebral body divided into quadrants. ALQ, anterior left quadrant; ARQ, anterior right quadrant; CA, central area; IA, intermediate area; PA, peripheral area; PLQ, posterior left quadrant; PRQ, posterior right quadrant.

High-resolution digital images of the histological sections were generated and then examined with QuPath (version 0.5.1), an open-source tool developed for viewing, annotating, and performing automated quantitative analysis of histological images. To enhance the localization capacity of structures within the vertebral body, each quadrant was further subdivided into three distinct topographic areas: the peripheral area (PA), corresponding to the outer third adjacent to the cortical layer of the vertebral body; the intermediate area (IA), corresponding to the middle third; and the central area (CA), corresponding to the central third of the vertebral body (Figure 3). Additionally, the sequential histological sections, ordered from the superior to the inferior endplate, were divided into three homogeneous regions, resulting in groups of upper, intermediate, and lower slices. This division allowed for the delineation of the superior, intermediate, and inferior thirds of the vertebral body, thereby enabling both individual and comparative analyses (Figure 3). To enable automated tissue classification, a Random Trees (RTrees) algorithm was implemented using QuPath’s Train Object Classifier module. This machine learning approach, based on an ensemble of decision trees, enhances classification performance while reducing the risk of overfitting. The model was trained using manually annotated regions, establishing four categories for histological classification: vascular tissue, bone, intertrabecular stroma, and bone marrow elements. For quantitative assessment, QuPath’s Cell Detection function was applied to identify and count nuclei within each tissue category. In vascular regions, the detected nuclei were attributed to endothelial cells; in bone tissue, they corresponded to osteoblasts, osteocytes, and osteoclasts. In the intertrabecular stroma, nuclei predominantly represented fibroblasts and inflammatory cells, whereas in bone marrow areas they were associated with diverse hematopoietic cell populations. The Cell Detection function enabled the identification of individual cellular elements, which were subsequently classified according to tissue compartment.

To assess the distribution of cellular elements, two principal metrics were determined: nuclear density (expressed as the number of nuclei per unit area, n/mm²) and nuclear proportion (defined as the percentage of specific nuclei relative to the total nuclei identified). These parameters were evaluated across the vascular, bone, intertrabecular stroma, and bone marrow compartments. The endothelial component of this analytical algorithm provides an overview of the global vascular distribution within the vertebral body; however, it does not distinguish between arterial, venous, and capillary vessels. Therefore, to achieve a more detailed and focused analysis of the arterial system, a manual evaluation of the intraosseous arterial vessels within the vertebral bodies was conducted. All identifiable arteries in the high-resolution digitized histological sections were manually annotated. For each artery, the following morphometric parameters were quantified: luminal diameter, cross-sectional area, perimeter, and orientation. All vessel measurements were conducted using calibrated image analysis software within QuPath. Vessel orientation was classified as either vertical (appearing with a circular lumen in axial sections) or oblique (appearing with an elongated lumen), based on their angle relative to the axial plane of the histological section. Vessels were classified based on luminal diameter as follows: small-caliber arteries (200–2,000 µm), arterioles (10–200 µm), and capillaries (<10 µm) (73,74). Also according to their luminal diameter, arterioles were further categorized by hierarchical branching order, following the adapted criteria below in ascending sequence: precapillary or terminal arterioles (10–15 µm], third-order arterioles (15–30 µm] (3rd order), second-order arterioles (30–100 µm] (2nd order), and first-order arterioles (100–200 µm] (1st order) (73,74). To increase detail within the larger luminal diameter ranges, we further subdivided first- and second-order arterioles into subtypes. First-order arterioles were categorized as subtype A (150–200 µm] and subtype B (100–150 µm], while second-order arterioles were categorized as subtype A (50–100 µm] and subtype B (30–50 µm]. Intervals of arteriole luminal diameter used to define arteriole order and subtype are expressed as (a – b], where a < x ≤ b, meaning the lower bound is exclusive and the upper bound is inclusive. Luminal diameters smaller than 10 µm were considered capillary vessels and were not included in this artery-focused analysis (73,74). The distinction between intraosseous arterial and venous vessels was performed using established morphological criteria (73,74). Arterial vessels were identified by the presence of a well-defined tunica media composed predominantly of smooth muscle cells, as well as relatively thicker vessel walls in proportion to the luminal diameter. In contrast, venous vessels exhibited thinner walls with little to no muscular layer. Vessel shape was also taken into account, with arteries typically displaying a more regular, circular profile, while veins often appeared irregular or collapsed (73,74). Given that oblique or longitudinal orientations of vessels in histological sections can distort their morphological features, multiple vessels within each microscopic field were comparatively assessed to enhance classification accuracy. Special attention was paid to the arrangement and thickness of the vascular wall layers, as well as the overall vessel architecture. Additionally, in cases where there was uncertainty in distinguishing between arterioles and venules, the vessels were reviewed by two expert pathologists. If consensus was not reached, the vessel was excluded from the final sample, thereby ensuring that only those unanimously identified as arterial were included in the analysis.

Image analysis and tissue component quantification were performed independently for each of these predefined topographic areas to assess potential regional heterogeneity, meaning for each quadrant and area, across the entire analyzed vertebral body, including upper, intermediate and lower slices. The results were then compared across different quadrants and its areas, both individually and in grouped analyses, to identify differences between the anterior and posterior halves of the vertebra, as well as between the right and left portions and central, intermediate and peripherical areas. Additionally, differences were investigated between the upper, intermediate, and lower slices of the vertebral body.

Statistical analysis

Statistical analyses were performed using IBM SPSS Statistics for Windows, Version 28.0 (IBM Corp., Armonk, NY, USA). Nominal variables are presented as absolute counts (n) and relative frequencies (%), whereas quantitative variables are expressed as means or medians, depending on their distribution, as determined by the Shapiro-Wilk normality test. Comparisons between groups were conducted using nonparametric tests. P values <0.05 were considered statistically significant.

Results

From an initial total of 902 histological slides obtained from axial sections, 162 slides were excluded after a preliminary histological review due to the presence of discal or cartilaginous components of the endplate. Only sections exhibiting continuous bone tissue across the entire quadrant were included, resulting in 740 slides. Of these, only those allowing reliable analysis with assured spatial orientation were considered for the final evaluation. Consequently, 140 slides were excluded due to absent or distorted cortical layer orientation, which prevented clear identification of the PA and the delineation of the three topographic areas: PA, IA, and CA (Figures 4,5). Additionally, 194 slides displaying excessive bone fragmentation, which could negatively impact or limit the analysis, were also excluded. This selection process yielded a final set of 406 slides, each represented by hematoxylin-eosin (Figures 5,6), Masson’s trichrome (Figure 7), and CD31 immunohistochemistry staining (Figure 8). These correspond to 406 quadrants, representing a total of 1,218 analyzed areas derived from three L1 vertebral bodies: vertebra 1 with 152 slides, vertebra 2 with 127 slides, and vertebra 3 with 127 slides. Furthermore, superior, intermediate, and inferior sections were defined to facilitate comparative analysis, resulting in 136 superior sections, 136 intermediate sections, and 134 inferior sections (Figure 3).

Figure 4 Distribution of absolute numbers and percentages of all detected arterioles throughout the entirety of the height of the analyzed vertebral bodies, organized by predefined quadrants and regions. Quadrants and areas with the highest absolute numbers and percentages of arterioles in each group are highlighted in red. Distribution across the entirety of the vertebral body height (left panel), superior third (upper right panel), intermediate third (middle right panel), and inferior third (lower right panel). ALQ, anterior left quadrant; ARQ, anterior right quadrant; CA, central area; IA, intermediate area; PA, peripheral area; PLQ, posterior left quadrant; PRQ, posterior right quadrant.

Figure 5 Hematoxylin-eosin stained slide depicting a quadrant from an axial section of the L1 vertebral body, with the topographic areas PA, IA, and CA outlined in red lines. CA, central area; Co, cortical wall; IA, intermediate area; PA, peripheral area.

Figure 6 Hematoxylin-eosin stained slide depicting a quadrant (ALQ) from an axial section of the L1 vertebral body. The enlarged image on the right highlights an obliquely oriented arteriole in the CA with a luminal diameter of 48.75 µm, corresponding to a second-order arteriole, subtype A. CA, central area; Co, cortical wall.

Figure 7 Masson’s trichrome stained slide depicting a quadrant (ARQ) from an axial section of the L1 vertebral body. The enlarged image on the right highlights an obliquely oriented arteriole in the PA with a luminal diameter of 184.75 µm, corresponding to a first-order arteriole, subtype A. Note the arteriole wall showing clearly defined intima, media, and adventitia layers. Smooth muscle fibers in the media appear red, while collagen fibers in the adventitia are highlighted in blue/green. ARQ, anterior right quadrant; Co, cortical wall; PA, peripheral area.

Figure 8 Immunohistochemical CD31 stained slide depicting a quadrant (PLQ) from an axial section of the L1 vertebral body. The enlarged image on the right highlights an obliquely oriented arteriole in the CA with a luminal diameter of 24.12 µm, corresponding to a 3rd order arteriole. Note the positive labeling of endothelial cells, visible as brown staining due to the DAB chromogen, lining the vascular lumen. The thick arteriole wall is counterstained in light blue. CA, central area; Co, cortical wall.

The vascular component was evaluated at both the endothelial cellular level and the specific arterial level. The analysis begins with the results concerning the density and proportion of identified vascular endothelial nuclei. Cell nuclear density and nuclear proportion for each analyzed parameter were assessed in the upper, intermediate, and lower slices of the L1 vertebral body, with detailed results presented in Tables 2-4. The median vascular density and cellular proportion across all analyzed slices were 215.84 endothelial nuclei per mm² and 8.06%, respectively. When analyzing the vertebral bodies as a whole, without distinguishing any regions other than the quadrants, the median endothelial vascular component density and proportion for each quadrant were as follows: the ARQ showed 202.89 nuclei/mm² and 6.79%, the ALQ had 205.11 nuclei/mm² and 7.39%, the PRQ exhibited 239.55 nuclei/mm² and 9.86%, and the PLQ displayed 220.45 nuclei/mm² and 8.96%. Analyzing only the areas, but without separating them by quadrant, median values for endothelial vascular component density and proportion were identified as follows: CA—212.18 nuclei/mm² and 6.60%; IA—205.00 nuclei/mm² and 5.87%; and PA—226.27 nuclei/mm² and 10.65%. All areas showed statistically significant differences between each other (P<0.001), observed both in the overall analysis and in the separate analyses of the upper, intermediate, and lower sections, confirming a vascular predominance in the PA, followed by the CA, and lastly the IA. When analyzing each area within the quadrants, the median values of vascular endothelial component density and proportion are as follows. In the CAs: ARQ-CA presents 195.00/mm² and 5.88%; ALQ-CA, 203.06/mm² and 5.61%; PRQ-CA, 225.67/mm² and 8.66%; and PLQ-CA, 221.58/mm² and 8.07%. In the IAs: ARQ-IA shows 183.28/mm² and 4.86%; ALQ-IA, 203.75/mm² and 4.81%; PRQ-IA, 229.77/mm² and 7.17%; and PLQ-IA, 204.23/mm² and 6.37%. In the PAs: ARQ-PA records 205.09/mm² and 8.08%; ALQ-PA, 213.01/mm² and 10.93%; PRQ-PA, 265.03/mm² and 12.69%; and PLQ-PA, 226.59/mm² and 12.67%. These results show that the PA of the posterior quadrants exhibit the highest vascular nuclear density and proportion, with the highest values found in the PA of the PRQ. Among the anterior quadrants, the PA of the ALQ shows the highest values for these parameters. In contrast, the lowest vascular component density is observed in the IA of the ARQ, while the lowest vascular proportion is recorded in the IA of the ALQ. By grouping the anterior quadrants with the posterior quadrants, two regions were defined: the anterior region (AR) and the posterior region (PR). The median endothelial vascular component density and proportion in these regions were 203.63/mm² and 7.11% for the AR, vs. 230.35/mm² and 9.55% for the PR, with statistically significant differences (P=0.001 for density and P<0.001 for proportion). When analyzing the different areas within the AR and PR separately: in the CA, the values were 197.64/mm² and 5.64% for AR, compared to 221.60/mm² and 8.27% for PR (statistically significant differences with P=0.006 for density and P<0.001 for proportion); in the IA, 198.56/mm² and 4.83% for AR, vs. 220.20/mm² and 6.87% for PR (P=0.016 and P<0.001, respectively); and in the PA, 210.26/mm² and 9.39% for AR, vs. 245.71/mm² and 12.67% for PR (both parameters with P<0.001). Statistically significant differences were also found in total vascular density and proportion between the ARQ, which had the lowest endothelial component, and both posterior quadrants: PRQ (P=0.005 for density and P=0.002 for proportion) and PLQ (P=0.024 and P=0.006, respectively). These results show higher endothelial cell density and proportion across all areas of the posterior quadrants when compared to the anterior quadrants. When analyzing the vertebral levels of the sections separately, this difference between anterior and posterior quadrants was more pronounced and statistically significant in the upper sections, while it became less evident in the middle and lower sections, where the endothelial distribution was more homogeneous. By grouping the right and left quadrants, the right region (RR) and the left region (LR) were defined. The median density and proportion of the endothelial vascular component were 218.34/mm² and 8.08% in the RR, compared to 212.81/mm² and 8.02% in the LR. Although a slight predominance of endothelial cells was observed in the RR, this difference was not statistically significant. When analyzing the individual areas, the values were similar, and no statistically significant differences were observed: in the CA, 212.68/mm² and 6.66% for the RR vs. 211.68/mm² and 6.50% for the LR; in the IA, 205.20/mm² and 5.87% for the RR vs. 203.75/mm² and 5.85% for the LR; and in the PA, 229.23/mm² and 10.13% for the RR compared to 223.97/mm² and 11.13% for the LR. In the analysis of the median endothelial vascular component density and proportion—without subdividing by area or quadrant—the values for the upper, intermediate, and lower slices were 182.86/mm² and 5.43%, 221.94/mm² and 6.42%, and 245.19/mm² and 10.14%, respectively. These values indicate a significantly lower vascular component density and proportion in the upper slices compared to the intermediate slices (P=0.004 and P<0.001, respectively) and the lower slices (both P<0.001). Regarding vascular proportion alone, it was also significantly higher in the lower slices compared to the intermediate ones (P=0.005). These findings demonstrate a progressive increase in both vascular density and proportion from the upper to the lower vertebral levels, indicating a greater endothelial representation in the lower third of the L1 vertebral body compared to the upper third. Regarding the different areas, and consistently in the order of upper-intermediate-lower slices, the median values were as follows. In the CA: 180.39/mm² and 4.43% vs. 223.22/mm² and 8.02% vs. 241.44/mm² and 8.82%. In the IA: 164.66/mm² and 3.95% vs. 217.07/mm² and 6.26% vs. 228.24/mm² and 7.69%. In the PA: 190.32/mm² and 7.25% vs. 229.86/mm² and 10.65% vs. 248.84/mm² and 13.95%. These results revealed statistically significant differences in endothelial vascular component density and proportion in the CA, with the upper slices showing lower values compared to both the intermediate slices (P=0.006 and P<0.001, respectively) and the lower slices (both P<0.001). Statistically significant differences were also observed in the IA, where the upper slices presented lower density and proportion compared to both the intermediate slices (P=0.003 and P<0.001, respectively) and the lower slices (both P<0.001). For vascular proportion alone, a significantly higher value was also observed in the lower slices compared to the intermediate slices (P=0.009). In the PA, statistically significant differences in density and proportion were likewise found, with the upper slices showing lower values compared to both the intermediate slices (P=0.03 and P<0.001, respectively) and the lower slices (P=0.001 and P<0.001, respectively). Once again, for vascular proportion alone, the lower slices showed significantly higher values than the intermediate slices (P=0.001). This analysis demonstrates a clear predominance of the endothelial vascular component in the lower third of the vertebral body, with the highest values observed in the PA, followed by the CA, and then the IA.

Regarding the specific detection of arterial vessels, only intraosseous arterial vessels with diameters between 10 and 200 µm were identified within the vertebral bodies. These fall within the classification of arterioles. Diameters below 10 µm correspond to capillaries and were not included in this analysis, while vessels with diameters greater than 200 µm were only observed outside the vertebral body, and not in an intraosseous location. In this sample, comprising three L1 vertebral bodies, a total of 13,712 arterioles were identified: 4,453 in vertebra 1, 4,068 in vertebra 2, and 5,191 in vertebra 3. In the initial analysis considering the entire height of the three vertebral bodies—i.e., without specifying whether the slices are upper, intermediate, or lower—the absolute and relative frequencies of arterioles detected in each quadrant revealed a higher prevalence in the ALQ (30.15%; n=4,135), followed by the ARQ (26.93%; n=3,693), then the PRQ (22.23%; n=3,048), and finally the PLQ (20.68%; n=2,836) (Figures 4). All quadrants showed statistically significant differences between each other in terms of arteriole percentage (P=0.006 between PRQ and PLQ and P<0.001 for all other comparisons). This distribution demonstrates a predominance of arterioles in the anterior quadrants. Accordingly, when the anterior quadrants were grouped into the AR and the posterior quadrants into the PR, a significant predominance of arterioles was confirmed in the AR (57.09%; n=7,828) compared to the PR (42.91%; n=5,884) (P<0.001). This difference was particularly attributed to the higher frequency of 2nd- and 3rd-order arterioles in the AR. In turn, when grouping the right quadrants into the RR and the left quadrants into the LR, a predominance of arterioles was observed in the LR (50.84%; n=6,971) compared to the RR (49.16%; n=6,741), although this difference was not statistically significant (P=0.05). Regarding the prevalence of arterioles across the different areas, the highest frequency was found in the PA (35.00%; n=4,793), followed by the CA (33.91%; n=4,650), and lastly the IA (31.13%; n=4,269) (Figure 4). Statistically significant differences were found between the PA and IA (P<0.001), as well as between the CA and IA (P<0.001), while the difference between PA and CA was not statistically significant (P=0.14). Regarding the types of arterioles found throughout the analyzed vertebral bodies, classified based on the vascular lumen diameter according to the intervals described in the Methods section, it was observed that the most frequent arterioles were the 3rd order (42.31%; n=5,802), followed by 2nd order subtype B (24.86%; n=3,409), precapillary arterioles (17.44%; n=2,391), 2nd order subtype A (13.64%; n=1,871), 1st order subtype B (1.47%; n=201), and finally 1st order subtype A (0.28%; n=38) (Figure 9). These frequencies showed statistically significant differences among all arteriole types (P<0.001). The distribution of arteriole types across the quadrants is illustrated in Figure 10 and corresponds to the same order of frequency observed for the complete axial section without quadrant division, that is, a predominance of 3rd order and 2nd order subtype B arterioles in all quadrants. Histological examples of the identified arterioles, oriented both obliquely and vertically, are presented for each category: 1st order (Figure 11), 2nd order subtype A (Figures 12,13), 2nd order subtype B (Figure 14, panels a and b), 3rd order (Figures 15,16), precapillary arterioles (Figure 17). Additionally, a representative histological section of the extraosseous lumbar arteries is shown in Figure 18.

Figure 9 Absolute frequencies (n) and relative frequencies (%) of arteriole types found across the entirety of vertebral body height. 1st order sub A: first-order subtype A arterioles. Arterioles were classified according to lumen diameter, with the mean lumen diameter indicated for each type. 1st order sub B: first-order subtype B arterioles. 2nd order sub A: second-order subtype A arterioles. 2nd order sub B: second-order subtype B arterioles. 3rd order: third-order arterioles.

Figure 10 Absolute frequencies (n) and relative frequencies (%) of arteriole types in each quadrant, across the entirety of vertebral body height. 1st order sub A: first-order subtype A arterioles. 1st order sub B: first-order subtype B arterioles. 2nd order sub A: second-order subtype A arterioles. 2nd order sub B: second-order subtype B arterioles. 3rd order: third-order arterioles. ALQ, anterior left quadrant; ARQ, anterior right quadrant; PLQ, posterior left quadrant; PRQ, posterior right quadrant.

Figure 11 Representative histological sections stained with hematoxylin-eosin, showing some of the identified 1st-order arterioles. (A) Vertically oriented subtype B arteriole from the CA of the PLQ (diameter =122.41 µm; area =18,440.61 µm²; perimeter =830.92 µm). (B) Vertically oriented subtype B arteriole from the CA of PRQ (diameter =100.10 µm; area =7,278.62 µm²; perimeter =609.41 µm). (C) Obliquely oriented subtype B arteriole from the PA of the PLQ (diameter =108.28 µm; area =42,061.67 µm²; perimeter =1,511.44 µm). (D) Obliquely oriented subtype A from the IA of ARQ (diameter =155.72 µm; area =42,905.41 µm²; perimeter =1,295.79 µm). (E) Obliquely oriented subtype B arteriole penetrating the Co through a foramen at the PLQ. The outer (Out) and inner (In) regions of the vertebral body are indicated. Left side: image showing the anatomical quadrant and the Co. Right side: magnified image highlighting an obliquely oriented subtype B arteriole (marked with an asterisk *) penetrating the Co. Arteriole measurements: diameter =105.93 µm; area =55,815.92 µm²; perimeter =2,808.38 µm. (F) Obliquely oriented subtype B arteriole (marked with an asterisk *) penetrating the Co through a foramen at the ALQ. The outer (Out) and inner (In) regions of the vertebral body are indicated (diameter =133.28 µm; area =19,021.12 µm²; perimeter =640.45 µm). ALQ, anterior left quadrant; ARQ, anterior right quadrant; CA, central area; Co, cortical wall; IA, intermediate area; PA, peripheral area; PLQ, posterior left quadrant; PRQ, posterior right quadrant.

Figure 12 Representative histological sections stained with hematoxylin-eosin, showing some of the identified 2nd-order arterioles-subtype A. (A) Vertically oriented arteriole from the IA of PRQ (diameter =72.60 µm; area =4,443.52 µm²; perimeter =552.18 µm). (B) Vertically oriented arteriole from the IA of ARQ (diameter =90.02 µm; area =6,512.20 µm²; perimeter =550.54 µm). (C) Obliquely oriented arteriole from the IA of ARQ (diameter =85.33 µm; area =63,898.00 µm²; perimeter =2,938.97 µm). (D) Obliquely oriented arteriole from the PA of PRQ (diameter =63.80 µm; area =14,829.92 µm²; perimeter =1,555.67 µm). (E) Obliquely oriented arteriole from the PA of PRQ (diameter =67.04 µm; area =32,598.34 µm²; perimeter =4,473.11 µm). ARQ, anterior right quadrant; IA, intermediate area; PA, peripheral area; PRQ, posterior right quadrant.

Figure 13 Representative histological sections stained with hematoxylin-eosin, showing some of the identified ramified or clustered 2nd-order arterioles-subtype A. (A) Axial and oblique arterioles in close proximity, from the IA of the PRQ (axial oriented arteriole—diameter =92.00µm; area =6,311.77 µm²; perimeter =965.94 µm) (oblique oriented arteriole—diameter =98.01 µm; area =14,057.79 µm²; perimeter =880.32 µm). (B) Cluster of obliquely oriented arterioles from the CA of the ALQ (anterior left quadrant) (middle arteriole—diameter =77.92 µm; area =21,540.67 µm²; perimeter =2,195.93 µm) (right arteriole—diameter =82.28 µm; area =9,184.88 µm²; perimeter =1,349.83 µm) (left arteriole—diameter =79.67 µm; area =5,856.67 µm²; perimeter =620.60 µm). (C) A vertically oriented 2nd order arteriole-subtype A (marked with a cardinal #) giving off a horizontal collateral branch, identified as a 3rd order arteriole (marked with an asterisk *), from the CA of the ALQ (vertical arteriole—diameter =59.50 µm; area =2,034.70 µm²; perimeter =311.93 µm) (oblique arteriole—diameter =15.19 µm; area =973.34 µm²; perimeter =607.02 µm). (D) A vertically oriented 2nd order arteriole-subtype A (marked with a cardinal #) giving off a horizontal collateral branch, identified as a 2nd order subtype B arteriole (marked with an asterisk *), from the CA of the ALQ (vertical arteriole—diameter =83.85 µm; area =5,446.62 µm²; perimeter =578.19 µm) (oblique arteriole—diameter =31.66 µm; area =1,523.92 µm²; perimeter =371.86 µm). (E) Oblique arterioles in close proximity, from the IA of the PRQ (arteriole marked with an asterisk *—diameter =33.07 µm; area =4,099.90 µm²; perimeter =838.04 µm) (arteriole marked with a cardinal #—diameter =48.78 µm; area =2,805.97 µm²; perimeter =395.04 µm). ALQ, anterior left quadrant; CA, central area; IA, intermediate area; PRQ, posterior right quadrant.

Figure 14 Representative histological sections stained with hematoxylin-eosin, showing some of the identified 2nd-order arterioles-subtype B. (A) Vertically oriented arteriole from the CA of ALQ (diameter =48.38 µm; area =2,459.39 µm²; perimeter =244.44 µm). (B) Vertically oriented arteriole from the CA of ALQ (diameter =35.33 µm; area =1,017.04 µm²; perimeter =379.733 µm). (C) Obliquely oriented arteriole from the PA of ALQ (diameter =31.63 µm; area =2,617.95 µm²; perimeter =577.21 µm). (D) Obliquely oriented arteriole from the IA of PRQ, showing markings for luminal diameter, area and perimeter measurements (diameter =40.28 µm; area =4,543.06 µm²; perimeter =1,001.82 µm). (E) Obliquely oriented bifurcated arteriole from the PA of ALQ (measurement of the 2nd-order arteriole subtype B marked with an asterisk *: diameter =36.96 µm; area =3,570.84 µm²; perimeter =937.97 µm) (measurement of the 3rd-order arteriole marked with a cardinal #: diameter =22.54 µm; area =1,668.17 µm²; perimeter =716.87 µm). (F) Obliquely oriented arteriole from the IA of ALQ (diameter =46.08 µm; area =6,623.43 µm²; perimeter =512.79 µm). (G) Obliquely oriented arteriole from the IA of PRQ (diameter =40.28 µm; area =4,034.43 µm²; perimeter =1,001.82 µm). (H) Two obliquely oriented arterioles in close proximity, from the IA of the ALQ (right arteriole—diameter =33.83 µm; area =2,651.03 µm²; perimeter =428.34 µm), (left arteriole—diameter =36.06 µm; area =3,398.36 µm²; perimeter =563.63 µm). (I) Obliquely oriented arteriole from the IA of PRQ (diameter =32.61 µm; area =5,302.07 µm²; perimeter =1,154.50 µm). (J) Obliquely oriented arteriole from the IA of PRQ (diameter =38.87 µm; area =6,824.28 µm²; perimeter =789.17 µm). (K) A vertically oriented 2nd order arteriole-subtype B (marked with a cardinal #) giving off two horizontal collateral branches, identified as a precapillary arterioles (marked with asterisks * and **), from the IA of the ARQ (vertical arteriole—diameter =45.64 µm; area =932.87 µm²; perimeter =126.43 µm) (oblique precapillary arteriole marked in one asterisk *—diameter =12.56 µm; area =346.57 µm²; perimeter =106.32 µm) (oblique arteriole precapillary marked in two asterisks **—diameter =11.87 µm; area =1,015.64 µm²; perimeter =257.57 µm). ALQ, anterior left quadrant; ARQ, anterior right quadrant; CA, central area; IA, intermediate area; PA, peripheral area; PRQ, posterior right quadrant.

Figure 15 Representative histological sections stained with hematoxylin-eosin, showing some of the identified 3rd-order arterioles. (A) Vertically oriented arteriole from the IA of ALQ (diameter =20.31 µm; area =291.12 µm²; perimeter =118.96 µm). (B) Vertically oriented arteriole from the PA of ALQ (diameter =24.68 µm; area =391.14 µm²; perimeter =180.87 µm). (C) Vertically oriented arteriole from the PA of ALQ (diameter =20.80 µm; area =336.63 µm²; perimeter =166.82 µm). (D) Obliquely oriented arteriole from the PA of ALQ (diameter =28.72 µm; area =772.85 µm²; perimeter =275.82 µm). (E) Obliquely oriented arteriole from the IA of ALQ (diameter =29.54 µm; area =2,785.40 µm²; perimeter =840.88 µm). (F) Obliquely oriented arteriole from the PA of PRQ (diameter =23.70 µm; area =1,697.43 µm²; perimeter =438.60 µm). (G) Obliquely oriented, ramified arteriole from the PA of the PRQ (measurement of the primary trunk: diameter =18.25 µm; area =1,404.36 µm²; perimeter =393.62 µm). (H) Obliquely oriented bifurcated arteriole from the IA of ARQ (measurement of the arteriole marked with an asterisk *: diameter =29.29 µm; area =2,572.58 µm²; perimeter =388.87 µm). ALQ, anterior left quadrant; ARQ, anterior right quadrant; IA, intermediate area; PA, peripheral area; PRQ, posterior right quadrant.

Figure 16 Representative histological sections stained with hematoxylin-eosin, showing selected 3rd-order arterioles with notable features in orientation and branching. (A) Obliquely and nearly horizontally oriented arteriole from the CA of the PLQ (diameter =21.66 µm; area =1,755.64 µm²; perimeter =645.53 µm). (B) Obliquely and nearly horizontally oriented arteriole from the CA of the PRQ, giving rise to two collateral branches marked with asterisks * (diameter =20.06 µm; area =2,809.84 µm²; perimeter =872.35 µm). (C) A vertically oriented 3rd order arteriole (marked with a cardinal #) giving off a horizontal collateral branch, identified as a capillary (marked with an asterisk *), from the CA of the ARQ (vertical arteriole—diameter =29.87 µm; area =1,773.50 µm²; perimeter =166.20 µm) (capillary—diameter =8.37 µm; area =177.07 µm²; perimeter =209.43 µm). ARQ, anterior right quadrant; CA, central area; PLQ, posterior left quadrant; PRQ, posterior right quadrant.

Figure 17 Representative histological sections stained with hematoxylin-eosin, showing some of the identified terminal or precapillary arterioles. (A) Vertically oriented arteriole from the CA of ALQ (diameter =14.76 µm; area =332.88 µm²; perimeter =212.79 µm). (B) Vertically oriented arteriole from the PA of PRQ (diameter =13.99 µm; area =106.20 µm²; perimeter =102.47 µm). (C) Vertically oriented arteriole from the IA of ALQ (diameter =13.27 µm; area =82.30 µm²; perimeter =46.71 µm). (D) Obliquely oriented arteriole from the PA of PRQ (diameter =14.67 µm; area =1,534.70 µm²; perimeter =419.00 µm). (E) Obliquely oriented arteriole from the IA of PRQ (diameter =10.34 µm; area =346.34 µm²; perimeter =331.81 µm). (F) Obliquely oriented arteriole from the PA of PRQ (diameter =14.73 µm; area =5,704.62 µm²; perimeter =868.49 µm). (G) Obliquely oriented bifurcated arteriole from the IA of PLQ (measurement of the primary trunk: diameter =14.13 µm; area =501.08 µm²; perimeter =167.24 µm). ALQ, anterior left quadrant; CA, central area; IA, intermediate area; PA, peripheral area; PLQ, posterior left quadrant; PRQ, posterior right quadrant.

Figure 18 Hematoxylin-eosin stained histological sections showing the anterior quadrants of an axial section through the L1 vertebral body (A: ARQ; B: ALQ). Red stars indicate the L1 lumbar arteries (A: diameter =958.42 µm; B: diameter =1,023.77 µm) in their horizontal course from anterior to posterior, curving along the anterolateral surfaces of the vertebral body, closely adjacent to the Co. ALQ, anterior left quadrant; ARQ, anterior right quadrant; Co, cortical wall.

Regarding the orientation of the arterioles, the majority exhibited an oblique orientation within the vertebral body (73.51%; n=10,080), while the remainder were vertical (26.49%; n=3,632), representing a statistically significant difference (P<0.001). This relationship was consistent across all quadrants, areas, and regions. When comparing orientation with arteriole type, the distribution of oblique versus vertical vessels varied significantly according to arteriole type (P<0.001). Although oblique vessels remained more frequent overall, as the arteriole lumen diameter increased, the proportion of oblique vessels decreased while vertical vessels increased. For example, precapillary arterioles showed 80.97% oblique vessels and 19.03% vertical vessels, whereas 1st order arterioles exhibited 51.88% oblique vessels and 48.12% vertical vessels. Although obliquely oriented vessels were always more frequent, the prevalence of vertically oriented arterioles was significantly higher in the CA (35.01% vertical vs. 64.99% oblique) compared to both the IA (22.32% vertical vs. 77.68% oblique) and the PA (21.93% vertical vs. 78.07% oblique), with the opposite pattern observed for oblique arterioles (P<0.001). The most frequently observed vertical arterioles in the CA were the 2nd order subtype B (31.94%), followed by the 3rd order arterioles (29.73%) and the 2nd order subtype A (26.66%). The most frequently observed oblique arterioles in the PA and IA were 3rd order arterioles, accounting for 45.62% in the PA and 44.93% in the IA. Subsequently, the same analyses were performed with the vertebral body slices separated into upper, intermediate, and lower levels, allowing for comparison across the three vertical thirds of the vertebral body. This approach revealed certain differences in the distribution of arteriole frequencies according to the level within the vertebral body (Figures 3,4). A significantly lower frequency of arterioles was observed in the lower slices (30.26%; n=4,149) compared to the upper slices (34.88%; n=4,783) and the intermediate slices (34.86%; n=4,780) (P<0.001). Regarding the absolute and relative frequencies of arterioles detected in the quadrants, differences in distribution were observed in the middle third of the vertebral bodies when compared to the previous analysis of the full vertebral height (Figure 4). While in the entire vertebra—as well as in the upper and lower slices—the highest prevalence of arterioles was found in the ALQ and ARQ, in the intermediate slices, although the ALQ still showed the highest number of arterioles (29.67%; n=1,418), the PRQ ranked second (26.72%; n=1,277), followed by the ARQ (23.83%; n=1,139), and finally the PLQ (19.79%; n=946) (Figure 4). Furthermore, the predominance of arterioles in the AR was maintained across all thirds of the vertebral body. However, the difference in the number of vessels detected in the AR compared to the PR was significantly smaller in the middle third (AR =53.49%, n=2,557 vs. PR =46.51%, n=2,223) than in the upper (AR =59.44%, n=2,843 vs. PR =40.56%, n=1,940) and lower thirds (AR =58.52%, n=2,428 vs. PR =41.48%, n=1,721), with statistical significance (P<0.001). In turn, when comparing the frequency of arterioles between the RR and the LR, no statistically significant differences were observed in the intermediate and lower thirds. However, in the upper third, a significantly higher number of arterioles was detected in the LR compared to the RR (52.46%, n=2,509 vs. 47.54%, n=2,274) (P<0.001). Regarding the prevalence of arterioles in the different areas, the distribution varied according to the vertebral body level analyzed, as shown in Figure 4. While in the entire vertebra and in the lower slices the highest prevalence of arterioles was observed in the PA and CA, this pattern was not consistent in the upper and middle thirds. In the lower third, the highest prevalence was found in the PA (35.00%; n=1,452), followed by the CA (33.98%; n=1,410), and then the IA (31.02%; n=1,287). In the upper third, the PA also had the highest prevalence (35.38%; n=1,692), but the IA appeared next (33.01%; n=1,579), followed by the CA (31.61%; n=1,512). Finally, in the middle third, the highest prevalence of arterioles was observed in the CA (36.15%; n=1,728), followed by the PA (34.50%; n=1,649), and lastly the IA (29.35%; n=1,403). Regarding the types of arterioles found across the different thirds of the vertebral body height, the distribution remained consistent with that observed for the full vertebral height, as previously described—namely, with a predominance of 3rd order arterioles, followed by 2nd order subtype B and precapillary arterioles (Figure 9). The same pattern was observed in the distribution of arteriole types across quadrants throughout all vertebral thirds, with no relevant differences between the upper, intermediate, and lower thirds when compared to the overall vertebral analysis (Figure 10). The mapping of arteriole types found within the various quadrants and their respective areas, along with the mean luminal diameters, areas, and perimeters, is presented in detail in Table 5 (upper third), Table 6 (middle third), and Table 7 (lower third). Regarding arteriole orientation, the upper third of the vertebral bodies showed a lower proportion of vertical arterioles (20.74%) compared to the middle (28.87%) and lower thirds (30.37%). The opposite trend was observed for oblique vessels (79.26% vs. 71.13% vs. 69.63%, respectively). This relationship was statistically significant (P<0.001). The predominance of vertical vessels in the CA, as well as the decrease in the proportion of oblique vessels and the corresponding increase in vertical vessels with increasing arteriole lumen diameter, were also consistently observed within each third of the vertebral body height when analyzed separately. In summary, the specific analysis of arterioles revealed a significant higher prevalence of intraosseous arterioles in the RA, PA and CA, as well as in the upper and middle thirds of the vertebral body. Conversely, the regions with lower arteriole presence were the PR, IA, and the lower third of the L1 vertebral body.

In addition to analyzing the vascular component at both the endothelial cellular and specific arterial levels, a quantitative analysis of cellular nuclei was also performed for other parameters—namely bone cells, intertrabecular stroma, and bone marrow—within each quadrant (Tables 2-4). The following paragraphs present the main statistically significant differences identified. Regarding bone cells, a significantly higher nuclear proportion was identified in the PA of the lower slices compared to the upper slices (0.72% vs. 0.48%, P=0.02). Additionally, the nuclear proportions of bone cells in the PA and IA were significantly higher on the PR of the vertebral body compared to the AR (PA: 0.62% vs. 0.48%, P=0.01; IA: 0.49% vs. 0.47%, P=0.02). In the upper slices, both the density and proportion of bone cells in the CA were significantly greater on the PR compared to the AR (density: 21.21/mm² vs. 13.12/mm², P=0.01; proportion: 0.53% vs. 0.29%, P=0.009). Bone cell density was significantly lower in the PA (11.50/mm²) compared to the CA (16.84/mm²) and IA (15.62/mm²) (P<0.001). Meanwhile, nuclear proportion showed significant differences between the PA and IA (P<0.001), a pattern observed throughout the entire vertebral bodies height. Notably, the nuclear proportion behaved inversely to density, presenting higher values in the PA (0.56%), which was significantly greater (P<0.001) than in the IA (0.47%). Finally, significantly higher values of both density and proportion of the total bone component were found in all areas on the RR of the vertebral body compared to the LR, throughout the entire vertebral height (density: 21.03/mm² vs. 14.10/mm²; proportion: 0.69% vs. 0.46%; both P=0.001, respectively) (Tables 2-4). Regarding the density and nuclear proportion of bone marrow across all areas, both parameters were significantly higher in the upper slices compared to the intermediate and lower slices (all P<0.001). Additionally, both the total bone marrow density and proportion were significantly greater in the AR of the vertebral body compared to the PR, particularly in the following: total bone marrow density (3,114.82/mm² vs. 2,573.60/mm², P=0.005); total bone marrow proportion (92.18% vs. 89.70%, P<0.001); bone marrow density in the CA (P=0.007); bone marrow proportion in the CA (P<0.001); density in the IA (P=0.004); proportion in the IA (P<0.001); and proportion in the PA (P<0.001) (Tables 2-4). When comparing the distribution of bone marrow across the CA, IA, and PA, statistically significant differences were found for both density and proportion (P<0.001), with densities of 3,180.57/mm² in the CA, 3,460.31/mm² in the IA, and 2,181.42/mm² in the PA; and proportions of 90.52%, 91.19%, and 84.15%, respectively. These values also varied depending on vertebral height level. In the upper slices, both bone marrow density and nuclear proportion differed significantly between areas (P<0.001), with the highest values observed in the IA, followed by the CA and then the PA. However, in the intermediate and lower slices, while the statistically significant differences persisted (P<0.001), the highest values were found in the CA, followed by the IA and then the PA. Moreover, in the upper slices, a significantly higher total bone marrow density was observed in the RR compared to the LR (2,529.46/mm² vs. 2,523.01/mm², P=0.02), mainly due to significantly higher densities in the IA (P=0.045) and PA (P=0.002). Finally, regarding the intertrabecular stroma, its nuclear proportion was significantly higher in the IA (0.23%) compared to both the CA (0.21%) and the PA (0.16%) (P<0.001). The total density of intertrabecular stroma was also significantly greater in the PR of the vertebral body (2.98/mm²) compared to the AR (2.64/mm²) (P<0.001). This difference was consistent across all areas and throughout the entire height of the vertebral bodies (Tables 2-4).

Discussion

Intraosseous arterial microvascular anatomy of the L1 vertebral body: prior knowledge, novel findings, and their integration with current scientific literature

The anatomical descriptions of the arterial vascularization of the lumbar vertebral bodies in humans, like all areas of anatomical science, have evolved over time, parallel to advancements in the precision and detail of the methods used to study them. Early investigations relied solely on anatomical dissection, followed by the use of angiographic techniques involving barium contrast injection. However, these techniques were unable to visualize the finer arteriolar networks due to the high radiodensity of the contrast material. More recent studies have begun employing selective microarteriography, often in combination with diaphanization and corrosion casting techniques, which have enabled the visualization of smaller-caliber arterioles, thereby naturally allowing for greater detail and accuracy in anatomical descriptions (37,38,49,50,53-66). Despite these advancements, studies that focus in detail on the intraosseous arterial microvasculature of the lumbar vertebral body remain scarce in the existing literature (37,38,49,50,53-66). As previously mentioned, currently, the most widely recognized and comprehensive anatomical descriptions of the intraosseous arterial network within the adult lumbar vertebral body are those provided by Jorge Draper Mineiro [1965], Markhashov [1965], Henry Crock and Hiroshi Yoshizawa (1976 and 1977), and J. F. Ratcliffe [1980] (Table 1) (37,38,50,58). The currently accepted anatomical description of the intraosseous arterial microvasculature of the lumbar vertebral body, employing nomenclature based on J. F. Ratcliffe’s 1980 classification, is illustrated in Figure 2 as an original drawing by the authors of this study and is succinctly outlined in the “Introduction” section (38).

Despite advances in anatomical vessel investigation techniques, the assessment of intraosseous microvasculature remains challenging, particularly for vessels with a diameter smaller than 100 µm (38,59,63,67). This was the reason that led us to perform a sequential histological analysis of three L1 vertebral bodies, harvested from fresh adult human cadavers to avoid vascular and bone damage caused by long-term preservation methods. As far as we are aware, no other anatomical study has conducted a histological analysis of the entire extent of the human lumbar vertebral body with the aim of characterizing intraosseous microvasculature. The study of histological axial sections is limited by its two dimensional nature and, unlike other approaches to assessing vascularization such as microarteriography, corrosion-fluorescence, and diaphanization techniques, does not allow tracking of the vessel’s trajectory or its three-dimensional configuration. However, it enables easy and detailed identification of very small vessels and provides morphometric analyses not available with other methods, including measurements of luminal diameter, area, and perimeter, as well as precise localization within the vertebral body. Therefore, we propose that our analysis complements existing three-dimensional anatomical studies and offers a comparative perspective on the location and dimensions of the identified arterioles. By dividing the vertebral body into distinct regions—namely, quadrants (ARQ, ALQ, PRQ and PLQ), and within each quadrant area, CA, IA, and PA—and by analyzing sections from the upper, middle, and lower thirds of the vertebral body separately, we were able to perform a detailed mapping of microvascular location (Tables 5-7), as well as compare its prevalence across these different regions.

In this study, intraosseous vessels were histologically identified by both detecting the nuclei of endothelial cells and directly visualizing and measuring the arterial vessels. We consider the cellular nuclei classification algorithm a valuable tool for high-resolution quantitative tissue analysis, as it enables accurate cell-type identification and tissue quantification, supports robust morphometric and spatial analysis within complex tissue architecture, and reduces the likelihood of error—thereby allowing for reliable, high-resolution tissue mapping. Since this algorithm identifies cellular nuclei but cannot distinguish the morphological differences between arterial and venous vessels, arterioles were manually identified in a sequential manner. This also enabled the characterization of their location, orientation, and dimensions—including luminal diameter, area, and perimeter—thereby allowing for their classification. Thus, the study of microvasculature based on the analysis of vascular endothelial nuclei density and proportion is inherently general, as it encompasses all vessel types—namely, arterioles, venules, and capillaries. Consequently, it provides a precise yet non-specific overview of overall vessel distribution, without distinguishing between individual vessel types. However, when this approach is combined with the specific identification of arterioles, it enhances anatomical detail and enables the formulation of hypotheses regarding venous and capillary contributions to endothelial density and proportion in specific regions, based on the relationship between endothelial nuclei detection and arteriole identification. This dual approach proved particularly useful for the detailed detection and description of vessels with diameters smaller than 100 µm, such as 2nd-order, 3rd-order, and precapillary arterioles. By combining these two complementary methods, we quantitatively mapped the intraosseous arterial microvasculature of the L1 vertebral body across its various regions, as detailed in Tables 2-7.

While the extensive outcome data obtained in this study provide a wealth of information on the intraosseous arterial microvasculature of the L1 vertebral body, we will focus on the most significant findings that advance the current understanding of this subject, particularly by enhancing and complementing existing three-dimensional anatomical models. The predominance of 3rd-order arterioles, 2nd-order subtype B arterioles, and precapillaries reflects the high frequency of small-caliber vessels, highlighting the rich arterial microvascular network within the vertebral body. This network is essential for bone vascularization and for supporting bone marrow activity, as it ensures the continuous delivery of oxygen, nutrients, and regulatory factors necessary for bone remodeling and hematopoiesis. A higher endothelial density and proportion were observed in all areas of the PR compared to the AR, despite a greater number of arterioles being identified in the AR. This likely indicates a predominance of venules, and possibly capillaries, in the PR, thereby increasing the number of endothelial cells in that region. This interpretation aligns with previous anatomical descriptions of venular convergence into the basivertebral vein, which exits the vertebral body through the retrovertebral foramen, suggesting that the PR is predominantly composed of endothelial cells derived mainly from venous sinusoids, venous plexuses, and a denser network of capillaries rather than from arterial vessels (37,44-48). In turn, the predominance of arterioles observed in the AR—most pronounced in the ALQ, likely due to the left-sided position of the aorta, from which the lumbar arteries originate, and its closer proximity to this quadrant (Figures 4,18)—is consistent with the concept of anterior arterial dominance, which proposes that the majority of the vertebral body’s blood supply originates from anterolateral penetrating arteries (Figure 11E,11F). Furthermore, despite lower endothelial density and proportion in the AR, the PA of the ALQ exhibits the highest values among the anterior quadrants, consistent with its characterization as the quadrant containing the greatest number of arterioles. Although this concept has been described in several anatomical studies, the primary source of arterial supply to the vertebral body remains a topic of ongoing debate in the scientific literature—specifically, whether it predominantly arises from posterior or anterior vascular branches. While some authors support the theory of a posteriorly dominant supply, in line with Antoine Portal’s original proposal regarding posterior nutrient arteries, others have not confirmed the presence of such vessels. Instead, these latter studies emphasize a predominant contribution from arteries entering through the anterolateral aspects of the vertebral body (37,38,49-66,68,69). Thus, for this controversy, our study findings now provide microvascular histological evidence supporting anterior arterial dominance in the L1 vertebral body. The differences in arteriole distribution observed in the middle third compared to the superior and inferior thirds—specifically, the significantly lower predominance of arterioles in the AR and the fact that the PRQ appears as the second quadrant with the highest number of arterioles after the ALQ—are most likely explained by the entry of the posterior nutritive arteries (or artery) into the posterior half of the vertebral body at this level. These arteries penetrate through the basivertebral foramen, which is located at the midpoint of the posterior wall of the vertebral body, significantly increasing the number of local arterioles and influencing the entire PR—a phenomenon not observed in the superior and inferior thirds (38). The analysis of endothelial density and proportion across the areas PA, IA and CA is consistent with the distribution of arterioles throughout the entire height of the vertebral bodies. It was observed that endothelial cell density and proportion, as well as arteriole prevalence, were highest in the PA, followed by the CA, and then the IA. This pattern can be explained by the marked predominance of arterioles in the PA, a region where venules are typically less abundant; thus, arterioles are the main contributors to the higher endothelial cell density and proportion in this area. Next is the CA, where both arterioles and venules are usually present in relatively high numbers, and finally the IA, which contains fewer arterioles but is expected to have a moderate presence of venules, given it is a trabecular transition zone with venules draining the cancellous bone toward the basivertebral vein (37,44-48,75). By combining the two histological evaluation methods employed in this study, we can reasonably infer that the highest prevalence of intraosseous venules is likely to occur in the posterocentral portion of the lumbar vertebral body. This finding is consistent with previous descriptions and is attributed to the location of the basivertebral vein—the main venous drainage channel of the vertebral body—which is surrounded by multiple converging venules (37,44-48,75). Nonetheless, it is important to acknowledge the potential contribution of capillaries to the quantification of endothelial cell density and proportion. The higher prevalence of arterioles in the PA is likely related to the peripheral entry of all arterioles into the vertebral body through foramina located along the cortical ring, particularly the anterolateral equatorial arterioles, posterior nutritive arterioles, and metaphyseal arterioles, as well as the overall significant number of peripheral arteries (Figures 2,11E,11F,12D,12E,14C,14E,15D,15F,15G,17D,17F). In turn, the higher prevalence in the CA corresponds to the cluster of vessels that converge toward the center of the vertebral body before becoming vertically oriented and ascending toward the endplates (Figures 2,11A,11B,13C,13D,16C). According to J. F. Ratcliffe’s angiosome system, the center of the vertebral body is vascularized by arterial branches arising from the anterolateral equatorial arteries in its anterolateral portion, and from the posterior nutritive arteries in its posterior portion (Table 1, Figure 2) (38). In our study, we found that the middle third of the vertebral body height was the only third in which arterioles were more prevalent in the CA rather than in the PA, a pattern that contrasts with what was observed in the superior and inferior thirds, where arteriole prevalence was higher in the PA (Figure 4). This finding is likely explained by two main factors. First, the lower number of arterioles in the PA of the middle third is likely due to the absence, in this region, of the 10 to 20 metaphyseal arteries that penetrate the anterolateral portion of the cortical ring, as well as 4 to 5 arteries that enter through its posterior aspect. These arteries are present only in the cephalic and caudal metaphyses of the vertebral body, which correspond to the superior and inferior thirds. In contrast, the middle region is supplied solely by the two anterolateral equatorial arteries and the posterior nutritive artery, along with peripheral arteries (Figure 2) (38). Second, the higher number of arterioles in the CA of the middle third is likely explained by the marked convergence of branches from the anterolateral equatorial arteries and the posterior nutritive artery toward the center of the vertebral body before giving rise to vertically ascending and descending branches. Conversely, the superior and inferior thirds show greater arteriole prevalence in the PA, most likely due to the combined presence of both metaphyseal and peripheral arteries in these regions, and they exhibit a lower number of vertical arterioles in their CA compared to the middle third (Figures 2,4) (38). On the other hand, the pattern of endothelial nuclear density and proportion across the upper, intermediate, and lower thirds is similar according to the areas, with predominance in the PA, followed by the CA, and lastly the IA. This likely reflects the significant contribution of venules and capillaries in these regions to the calculation of these parameters. Endothelial density and proportion, as well as arteriole quantification, did not show statistically significant differences when comparing the RR and LR, except in the upper third of the vertebral body, where the number of arterioles was significantly higher on the LR. This specific finding regarding arterioles does not support the concept that there are more penetrating arteries on the right side to compensate for the leftward deviation of the aorta (62). Instead, it aligns more closely with the anatomical proximity of the aorta and the origin of the lumbar arteries (Figure 18) to the anterolateral left region of the vertebral body—particularly at the ALQ, which is the quadrant where arteriole prevalence is highest. This spatial relationship may account for the greater representation of arterioles on the left side of the vertebral body. Endothelial density and proportion, as well as arteriole quantification assessed across the different vertical regions of the vertebral body, revealed a progressive increase in endothelial vascular density from the superior to the inferior third. The inferior third exhibited the highest endothelial density and proportion, followed by the middle third, and finally the superior third. In contrast, the specific analysis of arterioles showed a significantly higher prevalence in the superior and middle thirds compared to the inferior third. These findings are, once again, most likely attributable to a greater prevalence of other vessel types—particularly venules and capillaries—in the middle and inferior thirds of the vertebral body. As previously discussed, the predominance of venules is expected to occur in the middle third and posterior portion of the vertebral body, in accordance with the presence of the basivertebral vein, which is the main drainage vessel for intraosseous venules. This vein collects blood from both the superior and inferior thirds and drains it through the retrovertebral foramen, located at the midpoint of the posterior surface of the vertebral body (37,44-48,75). This is consistent with the middle third of the vertebral body exhibiting higher endothelial density and proportion compared to the upper third; however, it does not fully account for the even higher values observed in the lower third. This finding may be attributed to the presence of both venules and capillaries. Such a distribution could be influenced by gravitational effects in the vertebral venous system, which allows retrograde flow and favors venous stasis in the inferior third of the vertebral body. With aging, this phenomenon may contribute to an increased number and caliber of intraosseous venules in this region, which may be reflected in our sample, considering that the vertebrae analyzed were from individuals aged over 60 and 70 years. Regarding the orientation of the arterioles, the higher proportion of vertical arterioles found in the CA compared to other areas suggests that these may correspond to the previously described ascending and descending branches of both the anterolateral equatorial arteries and the posterior nutrient arteries, or even the metaphyseal arteries. After entering the periphery, these vessels course centripetally and, upon reaching the CA, send vertical branches directed towards the cephalic and caudal endplates (Figures 2,11A,11B,13C,13D,16C) (38). This predominance of vertical arterioles in the CA occurs throughout the entire height of the vertebral body, confirming the course of the vertical branches that extend from the center of the body in the middle third up to near the endplates. The size of these vertical branches may vary; however, they correspond to the types of vertical arterioles most frequently observed in this study within this region, namely 2nd order subtype B (31.94%), 3rd order arterioles (29.73%), and 2nd order subtype A (26.66%). Conversely, the higher proportion of oblique arterioles in the PA and IA may partly correspond to penetrating arterioles within the vertebral body and their smaller collateral branches, specifically from the anterolateral equatorial arteries, the posterior nutritive arteries, the metaphyseal arteries, and the peripheral arteries (Figures 2,11E,11F,12D,12E,14C,14E,15D,15F,15G,17D,17F), with these two areas showing a predominance of 3rd order arterioles (respectively 45.62% and 44.93%) (38). J. F. Ratcliffe describes metaphyseal and anterolateral equatorial arterioles as having luminal diameters ranging from 100 to 200 µm (38). Therefore, we can assume that the oblique arteries identified within this caliber range in the PA and IA regions, at the respective vertebral body height thirds—likely corresponding to first-order arteriole subtypes A and B—represent these arterioles (Figure 11C,11D). Metaphyseal arterioles appear in the superior and inferior thirds, while anterolateral equatorial arterioles correspond to those found in the middle third within the anterior quadrants (AR), whereas posterior nutritive arterioles correspond to those present in the same middle third but in the posterior quadrants (PR). In the CA, vertical arterioles with slightly smaller dimensions likely correspond to the main ascending and descending branches that emerge from the central arteries as they reach the center of the vertebral body (Figures 11A,11B,13C,13D). Meanwhile, the second-order arterioles found in the PA, particularly the oblique ones (Figures 12D,12E,14C,14E), likely correspond to the peripheral arteries described by J. F. Ratcliffe as having diameters between 50 and 75 µm (38).

Regarding the other parameters analyzed by the cellular nuclei detection algorithm, the observed distribution of cell densities and percentages reflects the expected microarchitecture of a normal vertebral body containing active red bone marrow (Tables 3-5). The predominance of hematopoietic cells, the presence of vascular endothelial cells, the sparse distribution of osteocytes, and the low abundance of intertrabecular stromal cells are consistent with typical vertebral histology and support the physiological integrity of the sample (76-78). The high overall cellularity observed in the marrow compartment reflects active hematopoiesis, while bone cells are embedded within the mineralized matrix of trabecular bone and reside in small cavities called lacunae. Although the vertebral body contains a substantial amount of trabecular bone, bone cells represent only a small proportion of the total nuclei identified in histological sections. This is due to the highly porous nature of trabecular bone, which forms a lattice-like network of thin bony struts surrounded by extensive intertrabecular spaces filled with bone marrow—tissue that is considerably more cellular than the mineralized bone component (76-78). Our results demonstrate a higher proportion of bone cells in the inferior third and the posterior half (PR) of the vertebral body. This finding is consistent with these regions representing areas of greater structural support, which require increased mechanical resistance provided by bone tissue. It may also help explain why fractures more commonly occur in the superior third and anterior portion of the vertebral body, as seen in classical wedge vertebral body fractures. Additionally, the higher density of bone cells in the CA and lower density in the PA, alongside an opposite trend in nuclear proportion—with lower proportion in the CA and higher in the PA—is likely explained by the lower presence of bone marrow in the PA and significantly higher marrow content in the CA. The predominance of bone marrow cells in the CA and IA is consistent with their previously described localization in the central region of the vertebral body, where hematopoietic activity is higher and greater intertrabecular space is available to accommodate marrow cells. The increased presence of red bone marrow in the AR, in the superior third, and on the right side of the L1 vertebral body, as observed in this study, may reflect a functional rather than purely anatomical distribution of hematopoietic tissue. Classically, it is considered that red bone marrow is more abundant in the posterior half of the vertebral body. However, factors such as local perfusion, trabecular microarchitecture, and oxygen availability may favor the maintenance of red marrow in ARs. Moreover, modern histological techniques offer higher resolution than earlier imaging-based studies, potentially revealing marrow distribution patterns that were previously underestimated (79,80). The presence of both higher bone cell density and proportion, along with increased bone marrow content on the right side of the L1 vertebral body, may reflect an adaptive response to asymmetric mechanical loading. According to Wolff’s law, bone undergoes remodeling in response to the mechanical stresses it experiences. If the right side of the vertebral body is subjected to greater biomechanical stress—whether due to posture, movement patterns, or spinal asymmetry—this may lead to increased osteoblastic activity, heightened metabolic demand, and enhanced vascularization (81). This, in turn, may promote the retention of more active red bone marrow on that side, supporting both bone remodeling and hematopoietic activity. Intertrabecular stromal cells provide structural support and contribute to the extracellular matrix within the marrow. Although functionally important for maintaining the hematopoietic niche and regulating marrow architecture, their quantitative presence is minimal. Consequently, their low density and proportion align with normal vertebral histology. The predominance of intertrabecular stromal cell nuclei in the IA and PR regions likely reflects the role of these areas—particularly the PR—in coping with greater mechanical loading, both of which stimulate increased cellular activity and remodeling, as well as expansion of the intertrabecular stromal tissue.

Clinical relevance of vertebral body intraosseous microvascular anatomy: potential applications in medical practice

The critical role of vascularization in numerous physiological processes within the vertebral body underscores the importance of a detailed understanding of its intraosseous arterial microvasculature, which may hold significant clinical implications. This vascular network plays a central role in the pathogenesis and progression of various conditions, including fracture healing, hematogenous dissemination of metastatic cells and septic emboli, vertebral osteomyelitis, and vascular tumors such as hemangiomas. Moreover, it may contribute significantly to the pathophysiology of vertebral osteoporosis. Ischemic mechanisms have also been implicated in the etiology of congenital spinal deformities—such as hemivertebrae and kyphoscoliosis—by disrupting normal vertebral development and ossification. Additionally, impaired perfusion of the vertebral body has been linked to degenerative disc disease, as it compromises the diffusion of nutrients and oxygen to the intervertebral disc, thereby accelerating disc degeneration (36-43).

Adequate vascularization is essential for the successful progression of bone repair, as it ensures the localized delivery of key biological elements, including growth factors, blood cells, nutrients, osteoblasts and osteoprogenitor cells, minerals, cytokines, inflammatory mediators, collagen, and hormones. As a result, damage to areas with inherently poor or fragile blood supply—particularly those dependent on terminal arterial branches, characterized by a sparse vascular network and minimal collateral circulation—can severely compromise the bone healing process. Such regions are more vulnerable to ischemia following arterial injury, leading to impaired revascularization and an inadequate restoration of tissue perfusion, ultimately reducing the capacity for effective bone regeneration and union. Although substantial evidence and a well-established pathophysiological rationale support the role of vascular supply in bone healing across various fracture sites in the human body, this critical biological factor—namely, arterial and microvascular integrity—has yet to be incorporated into current clinical decision-making for vertebral body fractures (23-36,82-92). Our research group recently conducted an anatomical investigation of the vascular anatomy of the L1 vertebral body in Wistar rats, followed by the development of an animal model of vertebral body fracture and a sequential histological study of the bone healing process. Subsequently, we performed an experimental study and demonstrated that disruption and exclusion of the anterolateral vascular supply significantly delayed the healing of L1 vertebral body fractures, as evidenced by slower phase transitions and impaired bone maturation compared to the control group, which did not undergo vascular intervention. These findings underscore the critical role of vascularization in successful vertebral repair and suggest that it should be considered in the development of future therapeutic strategies (36,93).

Although more than a century has passed since Hermann Kümmel first described post-traumatic vertebral necrosis in 1895, the primary pathophysiological mechanism underlying the condition remains an area of active investigation and has yet to be fully elucidated (23,31). The theory of ischemic or avascular necrosis has been proposed to explain the delayed onset of vertebral body collapse following trauma; however, it remains speculative and lacks definitive evidence. Nevertheless, the prevailing hypothesis currently holds that post-traumatic vertebral necrosis arises from arterial injury disrupting the bone’s blood supply, coupled with insufficient revascularization (23-35,94-97). Any factor that compromises the arterial vessels—whether extraosseous or intraosseous—that supply the vertebral body can induce ischemia, potentially resulting in osteonecrosis and impaired fracture healing (23,82,83,96,98-109). Vertebral body fractures most commonly associated with arterial injury are predominantly compression fractures, especially burst fractures. This is due to the fact that these fractures involve the entire vertebral body, frequently affecting one or both endplates as well as the posterior wall, thereby posing a risk of simultaneous damage to both the penetrating arteries at the anterolateral surfaces and those entering posteriorly (29,30,36).

Given its potential to impair bone regeneration, arterial microvasculature injury following vertebral fracture—which undoubtedly occurs in many cases, particularly in comminuted burst-type injuries—should be integrated into therapeutic decision-making algorithms. Specifically, it should help improve the rationale and standardize the selection between vertebral body-preserving stabilization treatments, which aim to support and promote natural fracture healing, and vertebral body replacement surgeries, where fracture healing is not expected. In the latter scenario, the biological capacity of the fractured vertebral body to heal is deemed insufficient; consequently, the treatment objective shifts from awaiting fracture healing to performing early vertebral body replacement to prevent progression to non-union. Although no current diagnostic method enables the vascular and biological assessment of whether a specific vertebral fracture pattern has caused significant disruption to the intraosseous blood supply sufficient to compromise bone healing and increase the risk of non-union, research in this area is essential for improving the management of severe vertebral body fractures. At present, decisions between initial vertebral body-preserving stabilization and early vertebral body reconstruction remain predominantly guided by the mechanical characteristics of the fracture and dependent on surgeon and institutional preference and experience, with little standardization and minimal practical consideration given to the vascular biological environment. There is a growing need to establish a standardized, clinically relevant approach for assessing the degree of vascular damage in vertebral fractures and for determining whether it significantly compromises bone regenerative capacity (29,30,36). Such a diagnostic strategy should integrate comprehensive anatomical understanding of both extraosseous and intraosseous arterial networks specific to the involved vertebral body with high-resolution imaging techniques capable of accurately depicting the three-dimensional configuration of fracture patterns. The identification of fracture lines intersecting regions known to harbor critical arterioles responsible for maintaining intraosseous perfusion would indicate a substantial risk of vascular compromise. Ideally, this risk should be quantified and systematically investigated to determine its relationship with impaired bone healing. Once future research establishes a threshold of arterial injury within the vertebral body beyond which bone healing is not expected, more precise, clearer, and standardized guidelines may be developed to support appropriate early decision-making regarding vertebral body replacement (interior or total) from the outset. This approach would integrate the biological likelihood of impaired bone regeneration, thereby facilitating a more consistent and tailored therapeutic strategy from the initial stage. In contrast, when vascular injury remains below this critical threshold, the fracture is more likely to heal adequately, supporting a more conservative strategy focused on vertebral preservation, with stabilization procedures or non-surgical management considered sufficient. It is important to emphasize that these concepts remain hypothetical and require validation through extensive investigation, initially in experimental models and subsequently in clinical studies. A key objective will be to establish reliable methods for determining whether the extent of vascular disruption in each vertebral body fracture is sufficient to impair its regenerative capacity. Defining this critical threshold, beyond which bone healing is significantly compromised and vertebral replacement becomes necessary, represents a central challenge in the development of such a therapeutic framework. Ultimately, the implementation of this approach would enable the incorporation of vascular biological parameters into clinical decision-making for vertebral fractures. This would enhance the accuracy and effectiveness of treatment strategies by facilitating early risk stratification and intervention, thereby reducing the incidence of severe complications such as non-union and post-traumatic vertebral necrosis. In the long term, it may support the development of a more robust, biologically grounded treatment algorithm, promoting improved recovery and lower complication rates (23,24,29,30,36).

The present anatomical study, grounded in histological analysis, contributes to the field by providing detailed insights into the intraosseous arterial microvascular architecture along the entire length of the L1 vertebral body—one of the vertebrae most commonly affected by fractures and, consequently, by post-traumatic vertebral osteonecrosis (23,24,30). This work offers crucial foundational knowledge essential for advancing the understanding of the biological vascular component in the diagnostic and therapeutic strategies for vertebral fracture management. Understanding the precise spatial distribution, classification, orientation, and dimensions of the arterial vessels within the L1 vertebral body offers the potential to correlate specific fracture lines and patterns with the likelihood of significant vascular disruption. This, in turn, can inform the development of imaging-based diagnostic criteria for vascular injury and help define thresholds of arterial damage beyond which natural bone healing becomes improbable. Ultimately, such histological insights will be instrumental in developing standardized, biology-driven clinical algorithms that incorporate vascular assessment into the decision-making process, thereby improving patient outcomes by guiding appropriate initial choices between vertebral body preservation and replacement.

Limitations

This study has several limitations, the most significant of which is the analysis being based on only three vertebral bodies. This small sample size may limit the representativeness of the findings for the general population, particularly since the specimens were sourced from donors aged over 60 and 70 years, whose vascularization may already exhibit age-related changes. However, aside from the inherent difficulty in conducting this type of study on younger individuals, as noted in the “Methods” section, donors with conditions potentially affecting vertebral morphology—particularly spinal pathologies, bone metabolic disorders, chronic inflammatory diseases, systemic vascular conditions, among others—were excluded. Nevertheless, the histological examination—performed sequentially along the entire length of each vertebral body and comprising 406 slides corresponding to 406 quadrants and 1,218 analyzed areas, with a total of 13,712 arterioles identified—represents a study of considerable scale and depth. It offers several novel contributions to the existing anatomical literature on the L1 vertebral body, particularly regarding its microvascular architecture. Specifically, it provides detailed information on the classification, localization, and orientation of arterioles, ranging from 1st-order vessels to precapillary arterioles. Even so, larger-scale studies are warranted to enhance the generalizability of these anatomical findings and support their broader clinical applicability. Another inherent limitation is the histological method itself, which relies on two-dimensional axial sections. Unlike other methods for studying vascular structures, such as microarteriography, corrosion casting with fluorescence, or diaphanization, histology does not allow for tracking the trajectories of vessels or capturing their three-dimensional configuration. However, it does permit the detailed identification of very small vessels and enables morphometric analyses that are not feasible with other methods, including precise measurements of luminal diameter, area, and perimeter, as well as accurate localization within the vertebral body. As a result, in cases where arterioles are oriented obliquely, it is possible that multiple sections of the same tortuous vessel may appear in a single histological plane and be mistakenly counted as separate arterioles. This represents an intrinsic limitation of the two-dimensional nature of histological analysis. We also acknowledge that the manual detection of intraosseous arterioles, necessary due to the limitations of the high-resolution automated quantification algorithm (which cannot differentiate arterioles from venules or capillaries as it relies solely on identifying endothelial cell nuclei), may be subject to inherent human error. Although the manual analysis was performed in a systematic and sequential manner, it is possible that it may have led to the under-identification of arterioles in the digitized slides, potentially resulting in an underestimation of their actual number. Additionally, the frequent fragmentation observed in the histological sections may, at times, compromise the accuracy of component quantification, particularly by underestimating the count of intraosseous arterioles. To mitigate these issues, we adopted a strict exclusion threshold: slides were excluded if they exhibited excessive fragmentation, lacked clearly defined cortical boundaries (thus preventing accurate delineation of the PA, IA, and CA areas), or were sectioned too close to the endplate or intervertebral disc, thereby failing to include the full osseous structure of the vertebral body. As a result, 496 slides were excluded from an initial set of 902, yielding a final sample of 406 high-quality slides. This high exclusion rate reflects our rigorous criteria to ensure the inclusion of only those slides with clearly defined anatomical regions and components suitable for accurate analysis. The exclusion of a significant number of slides carries the risk of rendering the available sections unrepresentative of the vertebral body; however, this exclusion occurred consistently across all regions of the vertebral body slices in the three vertebrae, and was not limited to any specific region. Nonetheless, we considered it more important to avoid analyzing slides that could compromise the quality and reliability of the assessment, prioritizing only those sections in optimal condition. Another limitation of this study concerns the differentiation between arterioles and venules in histological sections—a well-known challenge in practical histology that carries an inherent margin of error, even among experienced pathologists. Although standardized and widely accepted morphological criteria were applied—such as wall thickness, lumen diameter, and overall vessel morphology—these features can be significantly influenced by the orientation of the tissue section, potentially resulting in occasional misclassification (73,74). In the present study, the highly variable anatomical course of vessels within the vertebral body resulted in many being sectioned not in true axial planes, but rather in oblique or even longitudinal orientations. Such variations can distort vessel morphology and increase the likelihood of misclassification. Oblique sections, in particular, may cause arterioles to appear artificially wider with thinner walls, while venules may appear more circular with seemingly thicker walls. These distortions complicate the accurate evaluation of wall symmetry and relative layer thickness, thereby increasing the risk of diagnostic error. To mitigate these limitations, multiple vessels were evaluated within each field whenever possible, allowing for direct comparison between adjacent vascular profiles. Particular attention was given to the organization and composition of the vascular wall layers to improve diagnostic accuracy. Additionally, it is generally understood that within trabecular bone, intraosseous venous vessels typically possess thin walls with minimal or absent muscular layers—such as venous plexuses or diploic-type veins—rather than classic muscular veins characterized by a prominent tunica media, which are generally located outside the bone, within adjacent soft tissues or along major venous return pathways. The absence of classic muscular veins within the vertebral body thereby reduces the likelihood of misclassifying them as arterioles, which may otherwise resemble muscular veins due to their similarly thick tunica media. In contrast, venules with little or no muscular layer are generally easier to distinguish histologically. Furthermore, intraosseous veins in this region are not typically expected to contain valves, as valvular structures are more common in larger veins exhibiting unidirectional flow. Venous drainage within the vertebral cancellous bone is generally slow and multidirectional—similar to that observed in venous plexuses—which reduces the physiological necessity for valves. The literature also describes the internal vertebral venous plexus as valveless, supporting the assumption that smaller intraosseous venules likely share this anatomical feature (45,46,110-113). This makes the presence or absence of valves an unreliable criterion for distinguishing between arterioles and venules in this context. In summary, despite the application of multiple morphological criteria and the use of comparative analysis between vessels, a certain degree of subjectivity and variability inevitably remains, which may impact the accuracy of the data obtained from these assessments. This limitation should be acknowledged and considered when interpreting the findings of the present study. However, as previously noted in the “Methods” section, all cases with uncertainty in distinguishing between arterioles and venules were excluded following additional evaluation by two expert pathologists. Only arterial vessels for which there was unanimous consensus were included in the final analysis. Consequently, it is more likely that some arterial vessels were excluded due to diagnostic ambiguity than that venules were inadvertently included. It is also important to highlight that future studies may aim to identify and characterize intraosseous venules to provide a more comprehensive understanding of the vertebral microvascular network. Nevertheless, we anticipate that identifying venules—especially those of smaller caliber—would pose several challenges. Their significantly thinner and more delicate walls, compared to arterioles, not only make detection in histological analysis more difficult but also increase their vulnerability to damage or distortion during tissue preparation.

Nonetheless, despite the aforementioned limitations, we believe that the application of two complementary methods for histological evaluation of intraosseous arterial vessels—namely, high-resolution quantitative tissue analysis using an automated algorithm for detecting endothelial cell nuclei, coupled with manual characterization of arteriole location, orientation, and dimensions—has enabled a robust and comprehensive representation of the arterial microvascular distribution within the L1 vertebral body. Collectively, these methodologies offer reliable and detailed insights into the intraosseous arterial architecture, revealing a level of microvascular detail achievable only through histological analysis. This dual approach ensures both quantitative and qualitative detail, enabling highly informative mapping and an enhanced anatomical understanding of the microvascular organization within the L1 vertebral body.

Conclusions

The present study provides a sequential histological analysis of three adult human L1 vertebral bodies, based on two complementary methods for evaluating intraosseous arterial vessels—namely, high-resolution quantitative tissue analysis using an automated algorithm for detecting endothelial cell nuclei, alongside manual characterization of arteriole topographic location, orientation, and dimensions. A detailed mapping of the intraosseous arterial microvasculature of the L1 vertebral body was performed, allowing for several key conclusions. Within the intraosseous region, only arterial vessels of the arteriole type were identified, with third-order arterioles being the most frequent, followed by second-order subtype B arterioles and precapillaries. A significant predominance of arterioles was observed in the AR, particularly within the ALQ, as well as in the PA and CA areas, mainly concentrated in the superior and middle vertical thirds of the vertebral body. Conversely, the regions with the lowest arteriole density were the PR, the IA, and the lower third of the L1 vertebral body. Although obliquely oriented arterioles were more frequent across all regions, a notably higher number of vertically oriented arterioles was identified in the CA compared to the PA and intermediate IA areas. Conversely, the regions with the lowest arteriole density were the PR the IA, and the lower third of the L1 vertebral body. These findings were compared and successfully integrated with existing three-dimensional anatomical descriptions of arterial microvasculature reported in the literature, enabling a more detailed characterization of previously described arteriole groups. In addition to analyzing intraosseous arterial vasculature, the study included characterization of the vertebral body’s bone tissue, intertrabecular stroma, and bone marrow using a high-resolution automated algorithm designed to detect cell nuclei of specific tissues, thereby expanding the histological understanding of the internal morphology of the L1 vertebral body. The applied methodology offers reliable and detailed insights into the L1 intraosseous arterial vascularization architecture, revealing a level of microvascular detail achievable only through histological analysis. We propose that the present anatomical study provides a foundational framework for advancing research into the vascular contributions to vertebral body fracture biology. A deeper understanding of the pathophysiological impact of intraosseous vascular injury on vertebral bone regeneration holds the potential to revolutionize treatment paradigms for vertebral fractures. By elucidating the precise spatial distribution, classification, orientation, and dimensions of arterial vessels within the L1 vertebral body, this work paves the way for correlating specific fracture lines and patterns with the risk and extent of vascular disruption. Such knowledge can drive the development of imaging-based diagnostic criteria for vascular injury, helping to establish critical thresholds of arterial damage beyond which natural bone healing becomes unlikely. The ability to predict, at an early stage, whether a vertebral body fracture will progress to union or non-union based on the severity of intraosseous vascular compromise could profoundly improve clinical decision-making. This would enable a more standardized, evidence-based approach from the outset, guiding whether to pursue vertebral preservation strategies or opt for vertebral body replacement surgery. Ultimately, this approach promises to enhance the precision and efficacy of treatment by enabling early prediction and prevention, thereby reducing the incidence of non-union and post-traumatic vertebral necrosis. Moreover, it stands to comprehensively refine and improve existing therapeutic algorithms for vertebral body fracture management. In summary, our study provides valuable microvascular anatomical insights that contribute to the growing body of evidence underscoring the critical role of intraosseous blood supply integrity in vertebral fracture healing. Continued research in this area is essential and strongly encouraged, with the ultimate aim of translating these scientific advances into improved clinical outcomes for patients suffering vertebral body fractures.

Acknowledgments

We would like to thank Margarida Marques, Ricardo Moura, Joana Pinheiro Torres, Maria Moura, Alice Moura, and Amélia Moura for support during the experiments.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jss.amegroups.com/article/view/10.21037/jss-2025-aw-179/rc

Data Sharing Statement: Available at https://jss.amegroups.com/article/view/10.21037/jss-2025-aw-179/dss

Peer Review File: Available at https://jss.amegroups.com/article/view/10.21037/jss-2025-aw-179/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jss.amegroups.com/article/view/10.21037/jss-2025-aw-179/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by institutional ethics committees of Nova Medical School, Lisbon, Portugal (No. 135/2019/CEFCM) and Faculty of Medicine, University of Coimbra, Coimbra, Portugal (No. CE-141/2023/FMUC).

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Krüger A, Schmuck M, Noriega DC, et al. Percutaneous Dorsal Instrumentation of Vertebral Burst Fractures: Value of Additional Percutaneous Intravertebral Reposition-Cadaver Study. Biomed Res Int 2015;2015:434873. [Crossref] [PubMed]
2. Verlaan JJ, van Helden WH, Oner FC, et al. Balloon vertebroplasty with calcium phosphate cement augmentation for direct restoration of traumatic thoracolumbar vertebral fractures. Spine (Phila Pa 1976) 2002;27:543-8. [Crossref] [PubMed]
3. Korovessis P, Repantis T, Petsinis G, et al. Direct reduction of thoracolumbar burst fractures by means of balloon kyphoplasty with calcium phosphate and stabilization with pedicle-screw instrumentation and fusion. Spine (Phila Pa 1976) 2008;33:E100-8. [Crossref] [PubMed]
4. Daniaux H. Transpedicular repositioning and spongioplasty in fractures of the vertebral bodies of the lower thoracic and lumbar spine. Unfallchirurg 1986;89:197-213.
5. Marco RA, Meyer BC, Kushwaha VP. Thoracolumbar burst fractures treated with posterior decompression and pedicle screw instrumentation supplemented with balloon-assisted vertebroplasty and calcium phosphate reconstruction. Surgical technique. J Bone Joint Surg Am 2010;92:67-76.
6. Verlaan JJ, Somers I, Dhert WJ, et al. Clinical and radiological results 6 years after treatment of traumatic thoracolumbar burst fractures with pedicle screw instrumentation and balloon assisted endplate reduction. Spine J 2015;15:1172-8. [Crossref] [PubMed]
7. Kramer DL, Rodgers WB, Mansfield FL. Transpedicular instrumentation and short-segment fusion of thoracolumbar fractures: a prospective study using a single instrumentation system. J Orthop Trauma 1995;9:499-506. [Crossref] [PubMed]
8. Parker JW, Lane JR, Karaikovic EE, et al. Successful short-segment instrumentation and fusion for thoracolumbar spine fractures: a consecutive 41/2-year series. Spine (Phila Pa 1976) 2000;25:1157-70. [Crossref] [PubMed]
9. Speth MJ, Oner FC, Kadic MA, et al. Recurrent kyphosis after posterior stabilization of thoracolumbar fractures. 24 cases treated with a Dick internal fixator followed for 1.5-4 years. Acta Orthop Scand 1995;66:406-10.
10. McCormack T, Karaikovic E, Gaines RW. The load sharing classification of spine fractures. Spine (Phila Pa 1976) 1994;19:1741-4. [Crossref] [PubMed]
11. Knop C, Bastian L, Lange U, et al. Complications in surgical treatment of thoracolumbar injuries. Eur Spine J 2002;11:214-26. [Crossref] [PubMed]
12. Verlaan JJ, Dhert WJ, Verbout AJ, et al. Balloon vertebroplasty in combination with pedicle screw instrumentation: a novel technique to treat thoracic and lumbar burst fractures. Spine (Phila Pa 1976) 2005;30:E73-9. [Crossref] [PubMed]
13. Moura DL. The role of kyphoplasty and expandable intravertebral implants in the acute treatment of traumatic thoracolumbar vertebral compression fractures: a systematic review. EFORT Open Rev 2024;9:309-22. [Crossref] [PubMed]
14. Moura DL, Gabriel JP. Expandable Intravertebral Implants: A Narrative Review on the Concept, Biomechanics, and Outcomes in Traumatology. Cureus 2021;13:e17795. [Crossref] [PubMed]
15. Moura DFL, Gabriel JP. Intravertebral expandable implants in thoracolumbar vertebral compression fractures. Acta Ortop Bras 2022;30:e245117. [Crossref] [PubMed]
16. Been HD, Bouma GJ. Comparison of two types of surgery for thoraco-lumbar burst fractures: combined anterior and posterior stabilisation vs. posterior instrumentation only. Acta Neurochir (Wien) 1999;141:349-57.
17. Knop C, Blauth M, Bühren V, et al. Surgical treatment of injuries of the thoracolumbar transition. 2: Operation and roentgenologic findings. Unfallchirurg 2000;103:1032-47.
18. Carl AL, Tranmer BI, Sachs BL. Anterolateral dynamized instrumentation and fusion for unstable thoracolumbar and lumbar burst fractures. Spine (Phila Pa 1976) 1997;22:686-90. [Crossref] [PubMed]
19. Benson DR, Burkus JK, Montesano PX, et al. Unstable thoracolumbar and lumbar burst fractures treated with the AO fixateur interne. J Spinal Disord 1992;5:335-43. [Crossref] [PubMed]
20. McLain RF, Sparling E, Benson DR. Early failure of short-segment pedicle instrumentation for thoracolumbar fractures. A preliminary report. J Bone Joint Surg Am 1993;75:162-7.
21. Verlaan JJ, Diekerhof CH, Buskens E, et al. Surgical treatment of traumatic fractures of the thoracic and lumbar spine: a systematic review of the literature on techniques, complications, and outcome. Spine (Phila Pa 1976) 2004;29:803-14. [Crossref] [PubMed]
22. Palmisani M, Gasbarrini A, Brodano GB, et al. Minimally invasive percutaneous fixation in the treatment of thoracic and lumbar spine fractures. Eur Spine J 2009;18:71-4.
23. Young WF, Brown D, Kendler A, et al. Delayed post-traumatic osteonecrosis of a vertebral body (Kummell's disease). Acta Orthop Belg 2002;68:13-9.
24. Nicholas JJ, Benedek TG, Reece GJ. Delayed traumatic vertebral body compression fracture; part I: clinical features. Semin Arthritis Rheum 1981;10:264-70. [Crossref] [PubMed]
25. Benedek TG, Nicholas JJ. Delayed traumatic vertebral body compression fracture; part II: pathologic features. Semin Arthritis Rheum 1981;10:271-7. [Crossref] [PubMed]
26. Osterhouse MD, Kettner NW. Delayed posttraumatic vertebral collapse with intravertebral vacuum cleft. J Manipulative Physiol Ther 2002;25:270-5. [Crossref] [PubMed]
27. Maldague BE, Noel HM, Malghem JJ. The intravertebral vacuum cleft: a sign of ischemic vertebral collapse. Radiology 1978;129:23-9. [Crossref] [PubMed]
28. Hasegawa K, Homma T, Uchiyama S, et al. Vertebral pseudarthrosis in the osteoporotic spine. Spine (Phila Pa 1976) 1998;23:2201-6. [Crossref] [PubMed]
29. Moura DL, Gabriel JP. Expandable intravertebral implants in post-traumatic vertebral necrosis - new classification suggestion. Acta Ortop Bras 2023;31:e262943. [Crossref] [PubMed]
30. Moura DL, Pais HS, Lourenço P, et al. Surgical treatment of post-traumatic vertebral osteonecrosis including intravertebral expansive implants-clinical, functional and imaging outcomes and a center experience over 7 years. J Spine Surg 2024;10:450-67. [Crossref] [PubMed]
31. Kümmell H. Ueber die traumatischen Erkrankungen der Wirbelsäule. Deutsche Medizinische Wochenschrift 1895;21:180-1.
32. Lane JI, Maus TP, Wald JT, et al. Intravertebral clefts opacified during vertebroplasty: pathogenesis, technical implications, and prognostic significance. AJNR Am J Neuroradiol 2002;23:1642-6.
33. McKiernan F, Jensen R, Faciszewski T. The dynamic mobility of vertebral compression fractures. J Bone Miner Res 2003;18:24-9. [Crossref] [PubMed]
34. Mirovsky Y, Anekstein Y, Shalmon E, et al. Vacuum clefts of the vertebral bodies. AJNR Am J Neuroradiol 2005;26:1634-40.
35. Sevitt S. The healing of fractures of the lower end of the radius. A histological and angiographic study. J Bone Joint Surg Br 1971;53:519-31.
36. Moura DL, Casal D, Alves S, et al. The healing process of vertebral body fracture in Wistar rats: creation of an animal model and demonstration of the impact of anterolateral vascularization disruption on bone healing. J Spine Surg 2025;11:526-53. [Crossref] [PubMed]
37. Mineiro JD. Coluna vertebral humana - alguns aspectos da sua estrutura e vascularização (aspects of the structure and vascularization of the human vertebral column). Lisboa: Sociedade Industrial Grafica; 1965:277.
38. Ratcliffe JF. The arterial anatomy of the adult human lumbar vertebral body: a microarteriographic study. J Anat 1980;131:57-79.
39. De Salis J, Beguiristain JL, Cañadell J. The production of experimental scoliosis by selective arterial ablation. Int Orthop 1980;3:311-5. [Crossref] [PubMed]
40. Konerding MA, Blank M. The vascularization of the vertebral column of rats. Scanning Microsc 1987;1:1727-32.
41. Demmler K. Relationship between marrow and bones in osteoporoses. Med Klin 1975;70:41-8.
42. Ratcliffe JF. Anatomic basis for the pathogenesis and radiologic features of vertebral osteomyelitis and its differentiation from childhood discitis. A microarteriographic investigation. Acta Radiol Diagn (Stockh) 1985;26:137-43.
43. Ratcliffe JF. Arterial changes in the human vertebral body associated with aging. The ratios of peripheral to central arteries and arterial coiling. Spine (Phila Pa 1976) 1986;11:235-40.
44. Breschet G. Essai sur les veines du rachis. Paris: Méquignon-Marvis; 1819:1-16.
45. Batson OV. The function of the vertebral veins and their role in the spread of metastases. Ann Surg 1940;112:138-49. [Crossref] [PubMed]
46. Batson OV. The vertebral vein system. Caldwell lecture, 1956. Am J Roentgenol Radium Ther Nucl Med 1957;78:195-212.
47. Vidus-Vidius. De Anatomie Corporis Humani. Venice: apud Iuntas; 1611:121.
48. Falloppio G. Observationes Anatomicae. Venice: apud Marcum Antonium Ulmum; 1562:222.
49. Portal A. Cours d'Anatomie Medicale ou Elementes de l'Anatomie de l'Homme. Paris: Baudouin; 1803.
50. Crock H, Yoshizawa H. The blood supply of the vertebral column and spinal cord in man. 1st ed. New York: Springer Vienna; 1977.
51. Ferguson WR. Some observations on the circulation in foetal and infant spines. J Bone Joint Surg Am 1950;32-A:640-8.
52. Willis TA. Nutrient arteries of the vertebral bodies. J Bone Joint Surg Am 1949;31A:538-40.
53. Hyrtl J. Die Korrosions-Anatomie und ihre Ergebnisse. Wien; 1873:193-5.
54. Gray H, Lewis WH. Anatomy of the Human Body. 20th Ed. Philadelphia: Lea & Febiger; 1918.
55. Bohmig R. Die Blutgefassversorgung der Wirbelbandscheiben. Arch klin Chir 1930;158:374-424.
56. Wagoner G, Pendergrass EP. Intrinsic Circulation of the Vertebral Body. With Roentgenologic Considerations. American Journal of Roentgenology 1932;27:818-26.
57. Wiley AM, Trueta J. The vascular anatomy of the spine and its relationship to pyogenic vertebral osteomyelitis. J Bone Joint Surg Br 1959;41-B:796-809. [Crossref] [PubMed]
58. Manelfe C, Djindjian R. Angiographic examination of vertebral angiomas. Acta Radiol Diagn (Stockh) 1972;13:818-28.
59. Brookes M, Revell WJ. Blood supply of irregular bones — 1: Vertebral column. In: Blood Supply of Bone. London: Springer; 1998.
60. Fischer LP, Gonon GP, Carret JP, et al. Arterial vascularization of the lumbar vertebrae. Bull Assoc Anat (Nancy) 1976;60:347-55.
61. Crock HV, Yoshizawa H. The blood supply of the lumbar vertebral column. Clin Orthop Relat Res 1976;6-21.
62. Chiras J, Morvan G, Merland JJ. The angiographic appearances of the normal intercostal and lumbar arteries. Analysis and the anatomic correlation of the lateral branches. J Neuroradiol 1979;6:169-96.
63. Chiras J, Morvan G, Merland JJ, et al. Blood supply to the thoracic (dorsal) and lumbar spine. Anat Clin 1982;4:23-31.
64. Rothman RH, Simeone FA. The spine. 2nd ed. Philadelphia: WB Saunders Company; 1982.
65. Hajnovic L, Sefranek V, Schütz L. Influence of blood supply on fracture healing of vertebral bodies. Eur J Orthop Surg Traumatol 2018;28:373-80. [Crossref] [PubMed]
66. Forssman G, Petrén T. Die arterielle Versorgung der Brustwirbelkörper. Anat Anz 1938;88:167.
67. Stojanovic J, Kovac V. Diagnosis of ischemic vertebral collapse using selective spinal angiography. Rofo 1981;135:326-9. [Crossref] [PubMed]
68. Hanson R. Some anomalies, deformities and diseased conditions of the vertebrae during their different stages of development, elucidated by anatomical and radiological findings. Acta Chir Scand 1926;60:309.
69. Paturet G. Traité d’anatomie humaine. Paris: Masson; 1958:202-4.
70. Fischer AH, Jacobson KA, Rose J, et al. Hematoxylin and eosin staining of tissue and cell sections. CSH Protoc 2008;2008:pdb.prot4986.
71. Foot NC. The Masson trichrome staining methods in routine laboratory use. Stain Technol 1933;8:101-10.
72. Pusztaszeri MP, Seelentag W, Bosman FT. Immunohistochemical expression of endothelial markers CD31, CD34, von Willebrand factor, and Fli-1 in normal human tissues. J Histochem Cytochem 2006;54:385-95. [Crossref] [PubMed]
73. Aharinejad SH, Lametschwandtner A. Identification and interpretation of cast vessel structures. In: Aharinejad SH, Lametschwandtner A, editors. Microvascular Corrosion Casting in Scanning Electron Microscopy: Techniques and Applications. 1st ed. New York: Springer-Verlag; 1992:103-15.
74. Rahman M, Siddik AB. Anatomy, Arterioles. Treasure Island, FL, USA: StatPearls Publishing; 2025.
75. Carpenter K, Decater T, Iwanaga J, et al. Revisiting the Vertebral Venous Plexus-A Comprehensive Review of the Literature. World Neurosurg 2021;145:381-95. [Crossref] [PubMed]
76. Schröder G, Denkert K, Hiepe L, et al. Histomorphometric analysis of osteocyte density and trabecular structure of 92 vertebral bodies of different ages and genders. Ann Anat 2023;246:152022. [Crossref] [PubMed]
77. Ma YL, Dai RC, Sheng ZF, et al. Quantitative associations between osteocyte density and biomechanics, microcrack and microstructure in OVX rats vertebral trabeculae. J Biomech 2008;41:1324-32. [Crossref] [PubMed]
78. Auger JD, Frings N, Wu Y, et al. Trabecular Architecture and Mechanical Heterogeneity Effects on Vertebral Body Strength. Curr Osteoporos Rep 2020;18:716-26. [Crossref] [PubMed]
79. Nouh MR, Eid AF. Magnetic resonance imaging of the spinal marrow: Basic understanding of the normal marrow pattern and its variant. World J Radiol 2015;7:448-58. [Crossref] [PubMed]
80. Shah LM, Hanrahan CJ. MRI of spinal bone marrow: part I, techniques and normal age-related appearances. AJR Am J Roentgenol 2011;197:1298-308. [Crossref] [PubMed]
81. Frost HM. Wolff's Law and bone's structural adaptations to mechanical usage: an overview for clinicians. Angle Orthod 1994;64:175-88. [Crossref] [PubMed]
82. Modena V, Maiocco I, Bosio C, et al. Intravertebral vacuum cleft: notes on five cases. Clin Exp Rheumatol 1985;3:23-7.
83. Chou LH, Knight RQ. Idiopathic avascular necrosis of a vertebral body. Case report and literature review. Spine (Phila Pa 1976) 1997;22:1928-32.
84. Lu C, Miclau T, Hu D, et al. Ischemia leads to delayed union during fracture healing: a mouse model. J Orthop Res 2007;25:51-61. [Crossref] [PubMed]
85. Miedel E, Dishowitz MI, Myers MH, et al. Disruption of thrombospondin-2 accelerates ischemic fracture healing. J Orthop Res 2013;31:935-43. [Crossref] [PubMed]
86. Lu C, Hansen E, Sapozhnikova A, et al. Effect of age on vascularization during fracture repair. J Orthop Res 2008;26:1384-9. [Crossref] [PubMed]
87. Ma L, Zheng LW, Sham MH, et al. Uncoupled angiogenesis and osteogenesis in nicotine-compromised bone healing. J Bone Miner Res 2010;25:1305-13. [Crossref] [PubMed]
88. Zheng LW, Ma L, Cheung LK. Changes in blood perfusion and bone healing induced by nicotine during distraction osteogenesis. Bone 2008;43:355-61. [Crossref] [PubMed]
89. Street J, Bao M, deGuzman L, et al. Vascular endothelial growth factor stimulates bone repair by promoting angiogenesis and bone turnover. Proc Natl Acad Sci U S A 2002;99:9656-61. [Crossref] [PubMed]
90. Sharmin F, O'Sullivan M, Malinowski S, et al. Large scale segmental bone defect healing through the combined delivery of VEGF and BMP-2 from biofunctionalized cortical allografts. J Biomed Mater Res B Appl Biomater 2019;107:1002-10. [Crossref] [PubMed]
91. Schlickewei C, Klatte TO, Wildermuth Y, et al. A bioactive nano-calcium phosphate paste for in-situ transfection of BMP-7 and VEGF-A in a rabbit critical-size bone defect: results of an in vivo study. J Mater Sci Mater Med 2019;30:15. [Crossref] [PubMed]
92. Donneys A, Blough JT, Nelson NS, et al. Translational treatment paradigm for managing non-unions secondary to radiation injury utilizing adipose derived stem cells and angiogenic therapy. Head Neck 2016;38:E837-43. [Crossref] [PubMed]
93. Moura DL, Casal D, Alves S, et al. Vascular anatomy of L1 vertebral body in Wistar rat-corrosion-fluorescence, diaphanization and histological analysis, comparison to humans, and importance in blood supply-related investigation. J Spine Surg 2025;11:554-79. [Crossref] [PubMed]
94. Steel HH. Kümmell's disease. Am J Surg 1951;81:161-7. [Crossref] [PubMed]
95. Brower AC, Downey EF Jr. Kümmell disease: report of a case with serial radiographs. Radiology 1981;141:363-4. [Crossref] [PubMed]
96. Hermann G, Goldblatt J, Desnick RJ. Kümmell disease: delayed collapse of the traumatised spine in a patient with Gaucher type 1 disease. Br J Radiol 1984;57:833-5. [Crossref] [PubMed]
97. Libicher M, Appelt A, Berger I, et al. The intravertebral vacuum phenomen as specific sign of osteonecrosis in vertebral compression fractures: results from a radiological and histological study. Eur Radiol 2007;17:2248-52. [Crossref] [PubMed]
98. Kim YC, Kim YH, Ha KY. Pathomechanism of intravertebral clefts in osteoporotic compression fractures of the spine. Spine J 2014;14:659-66. [Crossref] [PubMed]
99. Allen BL Jr, Jinkins WJ 3rd. Vertebral osteonecrosis associated with pancreatitis in a child. A case report. J Bone Joint Surg Am 1978;60:985-7.
100. Queinnec JY, de Bray JM, Audran M, et al. Vertebral osteonecrosis. Apropos of 3 cases, 1 complicated by spinal cord compression and 2 associated with another necrotic localization. Rev Rhum Mal Osteoartic 1985;52:1-5.
101. Laloux P, Lefebvre S, Esselinckx W, et al. Spinal cord compression secondary to vertebral aseptic osteonecrosis. Spine (Phila Pa 1976) 1991;16:480-1.
102. Kaneda K, Asano S, Hashimoto T, et al. The treatment of osteoporotic-posttraumatic vertebral collapse using the Kaneda device and a bioactive ceramic vertebral prosthesis. Spine (Phila Pa 1976) 1992;17:S295-303. [Crossref] [PubMed]
103. Bhalla S, Reinus WR. The linear intravertebral vacuum: a sign of benign vertebral collapse. AJR Am J Roentgenol 1998;170:1563-9. [Crossref] [PubMed]
104. Kanchiku T, Taguchi T, Toyoda K, et al. Dynamic contrast-enhanced magnetic resonance imaging of osteoporotic vertebral fracture. Spine (Phila Pa 1976) 2003;28:2522-6; discussion 2. [Crossref] [PubMed]
105. Naul LG, Peet GJ, Maupin WB. Avascular necrosis of the vertebral body: MR imaging. Radiology 1989;172:219-22. [Crossref] [PubMed]
106. Baba T, Shitoto K, Yoshioka C, et al. Pathological fracture due to vertebral osteonecrosis associated with pancreatitis. Arch Orthop Trauma Surg 2011;131:11-4. [Crossref] [PubMed]
107. Javier RM, Moser T, Dietemann JL, et al. Multiple vertebral osteonecrosis. Joint Bone Spine 2008;75:341-4. [Crossref] [PubMed]
108. Malzer U, Pfeiffer M, Griss P. Posttraumatic vertebral body necrosis (Kümmell-Verneuil disease). Case report of a rare orthopaedic disease with histopathological correlation. Eur Spine J 1992;1:55-9.
109. Saito S, Inoue A, Ono K. Intramedullary haemorrhage as a possible cause of avascular necrosis of the femoral head. The histology of 16 femoral heads at the silent stage. J Bone Joint Surg Br 1987;69:346-51.
110. Théron J, Moret J. Lumbar phlebography. In: Théron J, Moret J, editors. Spinal Phlebography. Lumbar and Cervical Techniques. Berlin: Springer-Verlag; 1978:27-42.
111. Anderson R. Diodrast studies of the vertebral and cranial venous systems to show their probable role in cerebral metastases. J Neurosurg 1951;8:411-22. [Crossref] [PubMed]
112. Stringer MD, Restieaux M, Fisher AL, et al. The vertebral venous plexuses: the internal veins are muscular and external veins have valves. Clin Anat 2012;25:609-18. [Crossref] [PubMed]
113. Groen RJ, Groenewegen HJ, van Alphen HA, et al. Morphology of the human internal vertebral venous plexus: a cadaver study after intravenous Araldite CY 221 injection. Anat Rec 1997;249:285-94. [Crossref] [PubMed]