Pore Extractor 2D: An ImageJ toolkit for quantifying cortical pore morphometry on histological bone images, with application to intraskeletal and regional patterning

Abstract

Objectives Cortical porosity is used as a proxy of bone quality, fragility, and remodeling activity in anthropological contexts. Histological quantification is limited by time‐intensive manual annotation. Pore Extractor 2D is an ImageJ toolkit developed for computer‐assisted pore identification and automated pore morphometry. Materials and Methods Toolkit components include: (1) Utilities for cortical border clearing, (2) Image Pre‐Processing: Image contrast enhancement and noise reduction, (3) Pore Extractor: User‐directed options for segmenting, closing, and smoothing pore spaces, (4) Pore Modifier: Utilities to expedite manual correction of extracted pore spaces, and (5) Pore Analyzer: Morphometric analyses by pore type and anatomical region. Pore Extractor 2D was validated against manual annotation in a sample of midshaft mid‐thoracic human ribs (n = 30). Intraskeletal and regional analyses were piloted on matched (n = 9) human midshaft femora, tibiae, and sixth ribs. Results Pore Extractor 2D was statistically consistent with manual annotation of bone area, percent porosity, pore density, and mean pore size. The toolkit significantly (p < .05) reduced the smoothing effect of manual annotation by fitting pore borders to pixel brightness variations. Intraskeletal analyses found that the femur and tibia significantly exceeded the rib in “cortical” type porosity, while the rib predominantly formed “trabecularized” type porosity. Regional analyses determined that pore system expansion was elevated in the anterior femur, the anterior and medial tibia, and the cutaneous cortex of the rib. Discussion This toolkit provides expedited, semi‐automated porosity quantification that replicates manual annotation. Intraskeletal and regional variation in pore morphometry reflect localized strain patterning.

Regional assessment of male murine bone exposes spatial heterogeneity in osteocyte lacunar volume associated with intracortical canals and regulation by VEGF

“…Previously, regional assessment of cortical bone has been used to describe region-specific alterations to the bone cortex utilising both 2D (Cole et al 2022; van Tol et al 2020) and 3D imaging modalities (Núñez et al 2018; Uniyal et al 2021; Schneider et al 2007; Chiba et al 2013). However, the use of manual segmentation for regional selection lacks reproducibility and thus consistency between studies.…”

“…Importantly, when the bone cortex was not regionally analysed, these age effects on the intracortical canals were not observed due to averaging across different cortical regions. While other studies have compartmentalised (partition in different regions, typically quadrants for long bones) the bone cortex for morphometric analysis (Núñez et al 2018; van Tol et al 2020; Uniyal et al 2021; Schneider et al 2007; Cole et al 2022; Chiba et al 2013), manual selection of regions may lead to inter- and intra-observer errors.…”

“…Notably, this distinct population of canal-associated, large lacunae did not form in a murine model of conditional VEGF-knockout at the posterior TFJ. Previously, regional assessment of cortical bone has been used to describe region-specific alterations to the bone cortex utilising both 2D(Cole et al 2022;van Tol et al 2020) and 3D imaging modalities(Núñez et al 2018;Uniyal et al 2021;Schneider et al 2007;Chiba et al 2013). However, the use of manual segmentation for regional selection lacks reproducibility and thus consistency between studies.…”

Automatic Segmentation of Osteonal Microstructure in Human Cortical Bone Using Deep Learning: A Proof of Concept

“…Automated data collection processes have emerged in recent years, including software tools and machine learning-based methods [ 23 , 24 , 25 ]. Approaches such as neural networks or random forests have been applied for population affinity, sex and age estimation [ 24 , 26 ].…”

The impact of vitamin D deficiency on cortical bone area and porosity at the femoral midsection in children from post‐medieval London

“…These canals appear as pores in the bone cross‐section that can vary in size, distribution and number throughout a lifetime (Ramchand & Seeman, 2018). While widely acknowledged that accurate quantification of cortical bone porosity can be challenging (Cole et al, 2022; Zebaze & Seeman, 2015; see Discussion below), previous studies suggest that the amount of porosity can negatively impact bone quality and increase the risk of bone fracture by affecting the load‐bearing area, and ultimately, the amount and distribution of stress placed on bone material (Yeni et al, 1997; Zimmermann et al, 2015). More porous cortical bone has a greater susceptibility for crack propagation and, as a result, has a reduced fracture resistance (Bala et al, 2015; Yeni et al, 1997).…”