Thomas Edward Dufresne

Three-dimensional microimaging (MRμI and μCT), finite element modeling, and rapid prototyping provide unique insights into bone architecture in osteoporosis

With the proportion of elderly people increasing in many countries, osteoporosis has become a growing public health problem, with rising medical, social, and economic consequences. It is well recognized that a combination of low bone mass and the deterioration of the trabecular architecture underlies osteoporotic fractures. A comprehensive understanding of the relationships between bone mass, the three‐dimensional (3D) architecture of bone and bone function is fundamental to the study of new and existing therapies for osteoporosis. Detailed analysis of 3D trabecular architecture, using high‐resolution digital imaging techniques such as magnetic resonance microimaging (MRμI), micro‐computed tomography (μCT), and direct image analysis, has become feasible only recently. Rapid prototyping technology is used to replicate the complex trabecular architecture on a macroscopic scale for visual or biomechanical analysis. Further, a complete set of 3D image data provides a basis for finite element modeling (FEM) to predict mechanical properties. The goal of this paper is to describe how we can integrate three‐dimensional microimaging and image analysis techniques for quantitation of trabecular bone architecture, FEM for virtual biomechanics, and rapid prototyping for enhanced visualization. The integration of these techniques provide us with an unique ability to investigate the role of bone architecture in osteoporotic fractures and to support the development of new therapies. Anat Rec (New Anat) 265:101–110, 2001.

Risedronate Preserves Trabecular Architecture and Increases Bone Strength in Vertebra of Ovariectomized Minipigs as Measured by Three‐Dimensional Microcomputed Tomography

Risedronate reduces the risk of new vertebral fractures up to 70% within 1 year of treatment in patients with osteoporosis. Both increases in bone mass and preservation of bone architecture are thought to contribute to antifracture effects. Our objectives were to determine the effects of risedronate on trabecular bone mass and architecture and to determine the relative contributions of mass and architecture to strength in the vertebra of ovariectomized (OVX) minipigs. The minipigs were OVX at 18 months of age and were treated daily for 18 months with either vehicle or risedronate at doses of 0.5 mg/kg per day or 2.5 mg/kg per day. The three‐dimensional (3D) bone architecture of the L4 vertebral cores of Sinclair S1 minipigs was evaluated by 3D microcomputed tomography (μCT). Compared with the OVX control, the vertebral bone volume (bone volume/tissue volume [BV/TV]) was higher in both treated groups (p < 0.05). The architectural changes were more significant at the 2.5‐mg/kg dose and were more prevalent at the cranial‐caudal ends compared with the midsection. At the higher dose, the trabecular thickness (Tb.Th), trabecular number (Tb.N), and connectivity were higher, and marrow star volume (Ma.St.V) and trabecular separation (Tb.Sp) were lower (p < 0.05). The trabecular separation variation index(TSVI), a new measure to approximate structural variations, was smaller in the 2.5‐mg/kg‐treated group (p < 0.05). In this group, a significant preservation of trabeculae orthogonal to the cranial‐caudal axis was confirmed by a decrease in the degree of anisotropy (DA) and an increase in the percent Cross‐strut (%Cross‐strut; p < 0.05). Both normalized maximum load (strength) and normalized stiffness of the same vertebral cores were higher in the 2.5‐mg/kg risedronate group compared with the OVX group (p < 0.05). BV/TV alone could explain 76% of the variability of the bone strength. The combination of bone volume and architectural variables explained >90% of the strength. The study showed that risedronate preserved trabecular architecture in the vertebra of OVX minipigs, and that bone strength is tightly coupled to bone mass and architecture.

Advances in automated 3-D image analysis of cell populations imaged by confocal microscopy

Automated three-dimensional (3-D) image analysis methods are presented for rapid and effective analysis of populations of fluorescently labeled cells or nuclei in thick tissue sections that have been imaged three dimensionally using a confocal microscope. The methods presented here greatly improve upon our earlier work (Roysam et al.:J Microsc 173: 115-126, 1994). The principal advances reported are: algorithms for efficient data pre-processing and adaptive segmentation, effective handling of image anisotrophy, and fast 3-D morphological algorithms for separating overlapping or connected clusters utilizing image gradient information whenever available. A particular feature of this method is its ability to separate densely packed and connected clusters of cell nuclei. Some of the challenges overcome in this work include the efficient and effective handling of imaging noise, anisotrophy, and large variations in image parameters such as intensity, object size, and shape. The method is able to handle significant inter-cell, intra-cell, inter-image, and intra-image variations. Studies indicate that this method is rapid, robust, and adaptable. Examples were presented to illustrate the applicability of this approach to analyzing images of nuclei from densely packed regions in thick sections of rat liver, and brain that were labeled with a fluorescent Schiff reagent.

An improved watershed algorithm for counting objects in noisy, anisotropic 3-D biological images

Effective 3-D image processing algorithms are presented for automatic counting and analysis of cells in anisotropic 3-D biological images that are collected by laser-scanning confocal microscopes. In these instruments, the x-y resolution is much better than the resolution along the z axis, hence the voxels (pixels in 3-D) are anisotropic. In this work, the images are pre-processed by a 3-D extension of an anisotropic diffusion algorithm, and the resulting images are binarized by a clustering based segmentation algorithm. As a result of binary segmentation, some regions consist of individual objects while others are multi-object clusters. An extension of Vincent and Soilles watershed algorithm (1991) to anisotropic 3D spaces is used to separate such cell clusters. The watershed algorithm is applied on marker functions that are generated using a combination of 3-D morphological inverse distance functions and 3-D image gradients. Cell measurements, such as volume, average intensity and locations, are calculated on the result of watershed segmentation. This algorithm has been successfully applied to the automated analysis of cell populations from a variety of biological studies involving large numbers of tissue samples.

Architecture Is One of the Determinants of Bone Strength

In our recent publication in JBMR, we assessed the effects of risedronate on trabecular architecture and compressive strength of lumbar vertebrae from ovariectomized (OVX) minipigs. Regression analysis showed that trabecular bone volume (BV/TV) explained 76% of the variability of compressive strength. Adding architectural parameters to BV/TV in a three-parameter multiple linear regression model increased the R value to 91%. We concluded that risedronate improves the three-dimensional trabecular architecture in the vertebra of OVX minipigs in a way that contributes to an increase in bone strength. In their letter to the editor, Drs Kiebzak and Miller commented that our data lacked critical information, namely, ash weight or mineral content. Our decision not to include bone mineral content (BMC) or ash weight was based on data from a previous study on minipig vertebral cores in which there was a strong correlation between BV/TV and BMC (R 0.82). For completeness, however, we now report dual-energy X-ray absorptiometry (DXA) data that we collected from the minipig vertebral cores before microcomputed tomography measurements. After 18 months of treatment, risedronate significantly increased bone mineral density (BMD) compared with OVX controls (0.303 0.040 vs. 0.248 0.037 g/cm). Trabecular BV/TV (measured by microcomputed tomography) and BMD (measured by DXA) correlated strongly (R 0.96). In a simple linear regression model, BMD correlated strongly (R 0.80) with strength. Adding the architectural parameters TSVI and Tb.N to BMD in a three-parameter multiple linear regression model increased R to 0.91. The conclusion, therefore, is that although BV/TV and BMD correlate strongly with compressive strength, neither explains all the variance in bone strength. Although the contribution of trabecular architecture to bone strength, as indicated by change in R, was less than 15% in these minipigs, trabecular bone volume was high ( 25–33%). Architecture is likely to contribute more to bone strength and fracture risk in low bone mass conditions, and this is supported by data from several studies. Key architectural parameters have an exponential relationship with BV/TV. Furthermore, there were larger changes in architectural parameters when BV/TV was less than 11%, a value that has been suggested as threshold for spontaneous vertebral fracture. For trabecular bone in human biopsy specimens from different skeletal sites, BV/TV explained only 37–67% of the variance in elastic constants, especially at low bone volume. The prediction of the mechanical properties was significantly improved when architectural indices were included with bone volume. These data show that the relative contribution of architecture to strength increases as the bone mass decreases. It was not our intention to downplay the importance of BMD. Low BMD is clearly associated with a high risk of fragility fracture. However, there is now strong clinical evidence that an increase in BMD does not fully explain the fracture benefits of antiresorptive agents. It is appropriate, therefore, that we revisit the paradigm that BMD alone is an adequate surrogate for bone strength. We agree with Drs Kiebzak and Miller that the effectiveness of antiresorptive therapies for osteoporosis may be related to increase in BMD as well as to changes in one or several other factors that contribute to better bone quality. Our focus on the role of architecture does not preclude the importance of other components such as decreased bone turnover and osteocyte viability. Continued investigation into the effects of treatments on bone quality will improve our understanding of the pathophysiology and treatment of bone diseases.

Quantification of regional fat volume in rat MRI

Multiple initiatives in the pharmaceutical and beauty care industries are directed at identifying therapies for weight management. Body composition measurements are critical for such initiatives. Imaging technologies that can be used to measure body composition noninvasively include DXA (dual energy x-ray absorptiometry) and MRI (magnetic resonance imaging). Unlike other approaches, MRI provides the ability to perform localized measurements of fat distribution. Several factors complicate the automatic delineation of fat regions and quantification of fat volumes. These include motion artifacts, field non-uniformity, brightness and contrast variations, chemical shift misregistration, and ambiguity in delineating anatomical structures. We have developed an approach to deal practically with those challenges. The approach is implemented in a package, the Fat Volume Tool, for automatic detection of fat tissue in MR images of the rat abdomen, including automatic discrimination between abdominal and subcutaneous regions. We suppress motion artifacts using masking based on detection of implicit landmarks in the images. Adaptive object extraction is used to compensate for intensity variations. This approach enables us to perform fat tissue detection and quantification in a fully automated manner. The package can also operate in manual mode, which can be used for verification of the automatic analysis or for performing supervised segmentation. In supervised segmentation, the operator has the ability to interact with the automatic segmentation procedures to touch-up or completely overwrite intermediate segmentation steps. The operators interventions steer the automatic segmentation steps that follow. This improves the efficiency and quality of the final segmentation. Semi-automatic segmentation tools (interactive region growing, live-wire, etc.) improve both the accuracy and throughput of the operator when working in manual mode. The quality of automatic segmentation has been evaluated by comparing the results of fully automated analysis to manual analysis of the same images. The comparison shows a high degree of correlation that validates the quality of the automatic segmentation approach.

Characterizing three dimensional open cell structures without segmentation

Foam cells, particle conglomerates, biological tissue slices and colloidal suspensions are just a few examples of collections that create an image with multiple touching or overlapping regions. The characterization of the open cell size of such a continuous structure is tedious and computationally intensive for large 3D data sets. Typically, it is accomplished by segmenting the cells with a watershed technique and aggregating the statistics of all regions found. This paper provides the mathematical foundation for a newly discovered relationship between the average pixel value of a Euclidean Distance Map (EDM) and the radius of a conic section. The implementation of this relationship allows for a computationally simple and accurate characterization of the aggregate diameter associated with these open cell structures without segmentation.