Methods and theory in bone modeling drift: comparing spatial analyses of primary bone distributions in the human humerus

Abstract

This study compares two novel methods quantifying bone shaft tissue distributions, and relates observations on human humeral growth patterns for applications in anthropological and anatomical research. Microstructural variation in compact bone occurs due to developmental and mechanically adaptive circumstances that are 'recorded' by forming bone and are important for interpretations of growth, health, physical activity, adaptation, and identity in the past and present. Those interpretations hinge on a detailed understanding of the modeling process by which bones achieve their diametric shape, diaphyseal curvature, and general position relative to other elements. Bone modeling is a complex aspect of growth, potentially causing the shaft to drift transversely through formation and resorption on opposing cortices. Unfortunately, the specifics of modeling drift are largely unknown for most skeletal elements. Moreover, bone modeling has seen little quantitative methodological development compared with secondary bone processes, such as intracortical remodeling. The techniques proposed here, starburst point-count and 45° cross-polarization hand-drawn histomorphometry, permit the statistical and populational analysis of human primary tissue distributions and provide similar results despite being suitable for different applications. This analysis of a pooled archaeological and modern skeletal sample confirms the importance of extreme asymmetry in bone modeling as a major determinant of microstructural variation in diaphyses. Specifically, humeral drift is posteromedial in the human humerus, accompanied by a significant rotational trend. In general, results encourage the usage of endocortical primary bone distributions as an indicator and summary of bone modeling drift, enabling quantitative analysis by direction and proportion in other elements and populations.

Keywords: anthropology; bioarchaeology; bone; growth; histology; histomorphometry; method; modeling; primary bone.

Figure 1

Micrograph of a complete humeral cross‐section generated using digital stitching in autopano giga ^®. Each image was captured using a 530‐nm compensated, cross‐polarized 45° photographic technique, augmenting the primary tissue contrast and orientation. In this illustrative image the endosteal lamellar pocket (ELP) (A) was photographed with lamellar bone at one 45° angle, rendering it blue; Haversian tissue (B) show up ‘mottled’, displaying both warm and cool hues; periosteal tissue (C) was photographed at the other 45° angle, lending it a golden hue. Black‐lined arrows denote large, radially oriented ‘primary Volkmann's’ canals responsible for initial vascular supply of endosteal tissue. White solid arrows mark secondary remodeling events or osteons beginning to reorganize ELP primary tissue, and black solid arrows show the same in periosteal tissue. Unlike the endocortex, the pericortex has a primary vascularization that is longitudinal (made up of primary canals, sometimes called primary osteons if they contain lamellae). White‐lined arrows mark trabecularization of once highly remodeled periosteal origin compact tissue. These features combined, along with the position of the ELP, indicate the repositioning of the medullary cavity due to a net posteromedial drift. The insert details differences in orientation between old (i) and more recent (ii) endocortex, and old (iii) and more recent (iv) pericortex. Left humerus from the Xoclan cemetery, Merida, Mexico, male 35 years old at death. Unlike this illustrative image, analytical imaging for this project did not photograph endosteal and periosteal bone at different angles (in those images, all primary tissue was blue vs. mottled secondary tissues).

Figure 2

Schematic representation of the starburst sampling pattern superimposed on a left humerus. Four sampling axes construct eight regions of interest (ROIs) corresponding to the anterior, anterolateral, lateral, posterolateral, posterior, posteromedial, medial, and anteromedial aspects of the cross‐section. Counts were made from the outside field of view (FOV), inward. This array increases the likelihood of sampling distributed tissues in biomechanically informative axes while keeping efficient total times of analysis. Endosteal sampled area shown in dark‐shaded FOVs. Endosteal hits there weighted a vector (large arrow), defining the point‐count ELP position as the angular deviation from the posterior ROI.

Figure 3

(A) Right‐angle stitched micrograph demonstrating the interference artifact typically discussed in secondary osteons, the ‘Maltese cross’ artifact, disrupts visibility of tissue even more dramatically in primary tissue. (B) Using the 45° stitching technique, however, ensures primary tissue is only photographed at its brightest orientation. Upon reconstruction, the artifact is absent from the primary tissue, but is importantly retained in secondary tissue due its much smaller diameter. This has the combined effect of increasing visibility of primary tissue in whole cross‐section microphotography. Scale bar: 1 mm.

Figure 4

(A) Overlay generated to permit the geometric measurement of drift histology, including: (i) the total cross‐sectional scan, (ii) the 45° stitched cross‐polarized micrograph, (iii) the hand‐drawn mask defining the ELP and its centroid, (iv) the imagej generated geometric data including the cross‐sectional centroid (~7‐year‐old left humerus from an individual from Xcambó). (B) Schematic representing the angle recording ELP position (HEAn), defined in degrees deviation between the posterior aspect and the line connecting both centroids.

Figure 5

(A) Circular distribution of point‐count endosteal and periosteal angles (PEAn and PPAn) and hand‐drawn endosteal angles (HEAn), as well as their means, and standard deviation. Note the predictable general position of the ELP in the anterolateral aspect of the cross‐section and the near direct opposition of periosteal tissue (which has a slightly smaller range and is less variable). (B) The same data plotted for pairwise comparisons shows that both techniques identify similar trends in ELP position, despite a few differences approaching 50°.

Figure 6

Pairwise comparisons of endosteal lamellar area shows that although reporting the same trends, PEAr and HEAr, as measured in this study, do not equally report endosteal area.