Three-dimensional computational reconstruction of mixed anatomical tissues following histological preparation

doi:10.1016/S1350-4533(99)00032-6

Medical Engineering & Physics

Volume 21, Issue 2, March 1999, Pages 111-117

https://doi.org/10.1016/S1350-4533(99)00032-6 Get rights and content

Abstract

The creation of geometrically accurate computer models of anatomical structures with complex shape and mixed tissue types can be difficult. A method for shape reconstruction based on digital images of polymer embedded, serially sectioned specimens is presented. The distortion of bone and soft tissue specimens during all stages of histological preparation was measured. Serial sections of one specimen were stained with common histological stains to enhance the contrast between different tissue types. High-resolution digital images of these sections were then processed into a three-dimensional solid model using commercial software. Preparations containing bone and cartilaginous tissues were dimensionally stable following fixation, dehydration and embedding (shrinkage/expansion less than 2%). Staining was necessary to identify anatomical features that otherwise could not be differentiated from their surroundings. Although time consuming, this method provides cross-section images of a higher resolution than those obtained from CT or MRI scanning, and with better soft tissue visualisation.

Introduction

In the field of biomechanics, it is necessary to create geometrically accurate computer models of anatomical objects with a higher resolution and better soft-tissue differentiation than those available with current techniques. Accurate anatomical data of objects with complex shape or composition forms the basis of computational models used in kinematic analysis and finite element analysis (FEA). Whole-joint models, which incorporate both bone and soft tissues, could be used to study the interaction of anatomical structures, for example the effect of the glenoid labrum on the stability and range of motion of the shoulder joint. Three-dimensional reconstruction from serial cross-section images is the most common technique for creating such a model. For most FEA studies based on bone anatomy, computed tomography (CT) scan datasets are used to define the model geometry. Three-dimensional models based on CT scan data have been used extensively to study the femur, pelvis and spine 1, 2, 3, 4, 5. High resolution images of trabecular bone structure have been obtained using microscopic CT [6], but at present this method is limited to small specimens. The primary advantages of the CT method are: minimal distortion of the bone cross-section on the CT image, automated image processing for shape extraction, and in vivo cross-sectional data. However, it is less desirable to use CT data as the basis for models, which include structures other than bone, due to the poor resolution of soft tissues. Magnetic resonance imaging (MRI) can be used to obtain cross-sectional images with better differentiation of soft tissues [7], but the resolution of MR images is limited, typically 256×256 pixels. In many cases, discrete anatomical structures, such as the acetabular labrum and adjoining articular cartilage, cannot be delineated due to the ambiguous relationship between signal intensity and tissue composition [8].

Serial images of milled frozen specimens can provide high resolution images with good soft tissue visualisation [9]. However, handling and sectioning frozen specimens requires specialised equipment to ensure that the specimen remains frozen throughout the process, and that a satisfactorily cut surface is obtained. Also, milling of frozen specimens unavoidably destroys the specimen. Additionally, images of the cut surface are not truly two-dimensional, as the underlying structures are visible, making evaluation more difficult. Finally, it is not possible to differentiate between similar soft tissue structures, such as the glenoid labrum and adjacent articular cartilage, the meniscus and the articular surfaces of the knee, or the acetabular cartilage layers and the acetabular labrum, due to their similar appearance on the digital images. Ideally, digital images should be obtained using a method which enhances the contrast between neighbouring soft tissue structures.

Polymer embedded specimens are routinely used in histological and gross anatomy studies to investigate the morphological properties of anatomical structures 10, 11. To make accurate morphometric measurements of tissue, it is necessary to understand the effect that specimen processing has on the dimensional distortion of the tissue. There have been numerous studies of the dimensional changes of particular tissues during histological preparation, many with conflicting results 12, 13, 14, 15, 16. For example, small specimens of liver and kidney tissue shrink by up to 17% (linear) following fixation, dehydration and paraffin embedding [12], while large cancellous bone specimens may shrink by up to 7% [13]. A thickness reduction in cartilage specimens of up to 50% following histological preparation has been reported [14], but others have shown that small specimens of cartilage and subchondral bone experienced an area shrinkage of only 10% (approximately 5% linear shrinkage) following ethanol dehydration [15]. Block plastination, followed by section staining to enhance tissue contrast, has been proposed as a suitable method to study soft tissue components, such as the larynx, in their undisturbed state [16], but it is unclear whether such a procedure would be appropriate for larger specimens with a mixture of hard and soft tissues. It is to be expected that the dimensional changes of the soft tissues in such specimens would be limited by the reinforcing effect of the underlying bone. Thus, serially-sectioned, polymer-embedded specimens could provide accurate geometrical data for the three-dimensional reconstruction of hard and soft tissue anatomy.

The goal of this study was to develop a suitable method for serial specimen imaging in the special case where clear differentiation of visually similar, but compositionally different, collagenous tissues is required for large, whole-joint specimens. Preparation for imaging should not affect the dimensions of the specimen. Using this technique, it would be possible to construct accurate three-dimensional models of joints which are composed of several different tissue types, each with their own specific properties and function. This technique is demonstrated on a specimen taken from the acetabular rim, which was chosen for its variety of tissue types and composition.

Section snippets

Dimensional changes

Specimens were cut from the acetabular rims of two fresh-frozen porcine hip joints using a conventional band saw (Bizerba SPA, Milan, Italy). Ovine and porcine soft tissue have mechanical properties similar to human tissue [17], and so should demonstrate similar dimensional changes during processing. Porcine specimens were selected, as they were readily available fresh from the local slaughterhouse. Specimens were taken from five different locations around the circumference of the acetabular

Results

During the measurements, some tissue expansion and shrinkage were noticed. Considering linear dimension measurements, the unfixed soft tissue expanded, on average, by 3.5% following alcohol dehydration, and shrunk to 99.7% (±7.8%) of its original size following polymer embedding (Fig. 3). The fixed soft tissue expanded by 5.4% following fixation, expanded a further 1.4% following dehydration, and shrunk to 100.8% (±6.5%) of its original size following polymer embedding (Fig. 4). The dimensions

Discussion

The dimensional changes of substructures within polymer-embedded specimens have been measured. Fixation appears to limit the extent of dimensional changes in the subsequent dehydration step. While others have reported significant shrinkage of soft tissues following dehydration and embedding, we found very little effect of the specimen preparation on final specimen size. A one-sample test to determine statistically significant differences in the dimensions of the embedded specimens relative to

Conclusions

A method has been presented for three-dimensional shape reconstruction based on digital images of polymer-embedded, serially-sectioned anatomical structures. Large preparations containing bone and cartilaginous tissues are dimensionally stable throughout fixation, dehydration and embedding, with linear shrinkage or expansion of less than 2% on average. Before using this method with other tissue types, the dimensional changes would have to be re-evaluated for each specific tissue type. The

Acknowledgements

Supported in part by the AO Foundation. The authors wish to thank the following individuals for their contributions to this study: Prof. R. Ganz, I. Keller and J.-P. Imken.

References (19)

Y.K Kang et al.
Three dimensional shape reconstruction and finite element analysis of femur before and after the cementless type of total hip replacement
J Biomed Eng
(1993)
J.H Keyak et al.
Validation of an automated method of three-dimensional finite element modelling of bone
J Biomed Eng
(1993)
J.S.S Wu et al.
Clarification of the mechanical behaviour of spinal motion segments through a three-dimensional poroelastic mixed finite element model
Med Eng Phys
(1996)
S Aritan et al.
Program for generation of three-dimensional finite element mesh from magnetic resonance imaging scans of human limbs
Med Eng Phys
(1997)
J Lane et al.
Changes in dimensions of large cancellous bone specimens during histological preparation as measured on slabs from humar femoral heads
Calcif Tissue Int
(1983)
B Merz et al.
Automated finite element analysis of excised human femora based on precision-QCT
J Biomech Eng
(1996)
Pfleiderer M. Micromotion of cementless acetabular cups in the pelvis (German). Ph.D. thesis, University of...
J.L Kuhn et al.
Evaluation of a microcomputed tomography system to study trabecular bone structure
J Orthop Res
(1990)
P.A Kaplan et al.
MR imaging of the normal shoulder: variants and pitfalls
Radiology
(1992)

There are more references available in the full text version of this article.

Cited by (23)

Calcified cartilage in patients with osteoarthritis of the hip compared to that of healthy subjects. A design-based histological study
2021, Bone
Citation Excerpt :
In the present work, correction for tissue shrinkage was not performed. However, it has been demonstrated that cartilage and bone specimens from porcine hips embedded in methylmethacrylate shrink less than 2%, which can be regarded as negligible [45]. This study is the first to estimate and compare the volume and surface of the calcified cartilage in entire human femoral heads from patients with OA and healthy subjects.
Calcified cartilage is suggested to be involved in the pathogenesis of osteoarthritis (OA) by facilitating endochondral ossification at the bone-cartilage unit. Therefore, the objective was to quantify the volume and surface area of the calcified cartilage in the femoral head in OA patients and healthy subjects.
We used design-based stereological principles, i.e., systematic uniform random sampling and vertical uniform random sections of the entire femoral head. We investigated the articular and calcified cartilage and femoral head surface area and volume, excluding fovea capitis and marginal osteophytes, in 20 patients with OA and 15 healthy subjects.
The volume of the calcified cartilage was significantly larger for the patients with OA compared with the healthy subjects (mean difference [95% CI]) (284 [110,457] mm³, p = 0.002). The upper and lower surface area of the calcified cartilage, i.e. the tidemark and cement line, were both significantly larger for OA patients compared with the healthy subjects (17.8 [8.4,27.3] cm², p < 0.001) and (38.7 [20.8,56.7] cm², p = 0.002), respectively. The volume of the calcified cartilage and the volume of the femoral head were significantly correlated for the patients with OA (Spearman's ρ = 0.51, p = 0.021), but not for the healthy subjects (ρ = 0.41, p = 0.123).
Patients with OA had a larger femoral head surface area and more calcified cartilage compared to healthy subjects. The volume of the calcified cartilage correlated positively with the volume of the femoral head for patients with OA, but not for healthy subjects. This strongly supports the existing view that bone growth in OA is associated with endochondral ossification.
Age-related histological changes in calcified cartilage and subchondral bone in femoral heads from healthy humans
2019, Bone
Citation Excerpt :
The sections were stained with either May-Grünwald toluidine blue or Masson-Goldner trichrome. Shrinkage is low in plastic embedded sections, and therefore no correction for tissue shrinkage was performed [24]. The caput was macroscopically and microscopically defined by the shape and the extent of the articular cartilage.
Age is the most important risk factor for osteoarthritis (OA). It is suggested that changes in subchondral bone and calcified cartilage may occur in early OA. Therefore, the aim was to investigate age-related changes in the femoral head composition. We hypothesise that the thickness of the subchondral bone plate decreases with age, while the thickness of the calcified cartilage increases with age as seen in early-stage OA.
Femoral heads from 29 women (20–74 years) and 32 men (23–78 years), who had died suddenly and unexpectedly, were obtained at autopsy. Individuals with bone or joint diseases or macroscopic abnormal cartilage were excluded. Using design-based stereology, femoral head volume as well as thickness and volume of the calcified cartilage and subchondral bone plate were estimated and correlated to sex and age.
The thickness and volume of the subchondral bone plate were not correlated with age. Calcified cartilage thickness and volume correlated positively with age in women, while the femoral head volume was correlated positively with age in men.
In human femoral heads obtained from a cross-sectional population without macroscopic OA changes, the thickness of the subchondral bone plate did not change with age, which differs from the thinning seen in early OA. Surprisingly, the age-related changes of the volume and thickness of the calcified cartilage and of the volume of the femoral head were different for women and men. This indicate that cartilage and bone metabolism is sex-specific, which may influence ageing of the hip joint.
Inter-lamellar shear resistance confers compressive stiffness in the intervertebral disc: An image-based modelling study on the bovine caudal disc
2015, Journal of Biomechanics
Citation Excerpt :
Since the degree of shrinkage was unknown, we did not attempt any geometric normalization of disc compressive stiffness, although we note that such approaches have been usefully employed for comparing different sized specimens by other authors (Beckstein et al., 2008; Showalter et al., 2012). However, we did attempt to minimize shrinkage by performing the dehydration and embedding en bloc with the disc still attached to the neighbouring vertebrae (Ferguson et al., 1999). Fig. 1 shows evidence of shrinkage-induced delamination between lamellar layers, and careful inspection of Fig. 1 shows that there are some regions of distinct colour (e.g. between lamellae 6 and 7) which could either be another thin lamella or a partial delamination between layers (where the interference colour is distinct due to the reduced collagen density in the partially delaminated layer).
The intervertebral disc withstands large compressive loads (up to nine times bodyweight in humans) while providing flexibility to the spinal column. At a microstructural level, the outer sheath of the disc (the annulus fibrosus) comprises 12–20 annular layers of alternately crisscrossed collagen fibres embedded in a soft ground matrix. The centre of the disc (the nucleus pulposus) consists of a hydrated gel rich in proteoglycans. The disc is the largest avascular structure in the body and is of much interest biomechanically due to the high societal burden of disc degeneration and back pain. Although the disc has been well characterized at the whole joint scale, it is not clear how the disc tissue microstructure confers its overall mechanical properties. In particular, there have been conflicting reports regarding the level of attachment between adjacent lamellae in the annulus, and the importance of these interfaces to the overall integrity of the disc is unknown. We used a polarized light micrograph of the bovine tail disc in transverse cross-section to develop an image-based finite element model incorporating sliding and separation between layers of the annulus, and subjected the model to axial compressive loading. Validation experiments were also performed on four bovine caudal discs. Interlamellar shear resistance had a strong effect on disc compressive stiffness, with a 40% drop in stiffness when the interface shear resistance was changed from fully bonded to freely sliding. By contrast, interlamellar cohesion had no appreciable effect on overall disc mechanics. We conclude that shear resistance between lamellae confers disc mechanical resistance to compression, and degradation of the interlamellar interface structure may be a precursor to macroscopic disc degeneration.
Volume shrinkage of bone, brain and muscle tissue in sample preparation for micro-CT and light sheet fluorescence microscopy (LSFM)
2014, Microscopy and Microanalysis
Light sheet microscopy of the gerbil cochlea
2021, Journal of Comparative Neurology
Effect of a formalin-based fixation method on bone mineral content in human ex-vivo specimens
2021, Acta Odontologica Scandinavica

View all citing articles on Scopus

View full text

Technical NoteThree-dimensional computational reconstruction of mixed anatomical tissues following histological preparation

Abstract

Introduction

Section snippets

Dimensional changes

Results

Discussion

Conclusions

Acknowledgements

J Biomed Eng

J Biomed Eng

Med Eng Phys

Med Eng Phys

Calcif Tissue Int

Automated finite element analysis of excised human femora based on precision-QCT

J Biomech Eng

Evaluation of a microcomputed tomography system to study trabecular bone structure

J Orthop Res

MR imaging of the normal shoulder: variants and pitfalls

Radiology

Technical Note
Three-dimensional computational reconstruction of mixed anatomical tissues following histological preparation