Original ArticlesValidation of quantitative backscattered electron imaging for the measurement of mineral density distribution in human bone biopsies
Introduction
The measurement of bone mineral density (BMD) is a commonly used technique in the diagnosis and management of osteoporosis.14 This in vivo method can be used to monitor the mineral content in a given portion of bone, but it cannot differentiate between changes in bone volume or in degree of mineralization of the bone matrix. Hence, it is possible that a patient has a high bone mass, while the bone is characterized by poor biomechanical stability (e.g., because of abnormal mineralization). In contrast to BMD, the bone mineral density distribution (BMDD), as measured by quantitative backscattered electron imaging (qBEI), is able to measure differences in degrees of mineralization.9, 23, 2631 The drawback is that the determination of BMDD requires undecalcified, plastic-embedded bone samples (e.g., from biopsies). The advantage is that the BMDD can be measured directly on the blocks of resin-embedded tissue routinely used for preparing sections for bone histomorphometry.7 Hence, qBEI provides important additional information without interfering with the routinely applied investigation methods or needing new preparation techniques.
The qBEI method is based on the detection of electrons backscattered in a region close to the surface of the specimen struck by the primary electron beam of a scanning electron microscope (SEM). In general, the fraction of the electrons backscattered increases with the increase of the atomic number Z (Z contrast) of the atoms hit by the beam.12 In the case of bone tissue, where organic matrix (H, C, N, O, P, S) and mineral (Ca, P, O, H, C, Mg) are the essential components, the concentration of calcium, the constituent with the highest atomic number (Z = 20), dominantly influences the intensity of the backscattered electrons (BE). Comparable with the more common microradiographic method,1, 10, 24, 29 but with a much better resolution, relative changes of local mineral concentrations can be visualized within cortical and cancellous bone. While qBEI images are taken in about a 0.5-μm-thick surface layer of the specimen, microradiography is based on the absorption of X-rays by 30–100-μm-thick bone ground sections. The small sampling thickness is a main advantage of qBEI because it avoids most of the projection effect errors that may occur in the investigation of trabecular bone by microradiography.2
Until now, only a few quantitative studies related to clinical problems, with emphasis on age and gender changes6, 8, 19 or treatment strategies for osteoporosis,22 were conducted by means of qBEI. Some of the major problems in the qBEI method are the stability of the SEM, the calibration of the BE signal, and the standardization of the BE signal for mineral concentration values. A Z calibration of the BE gray level using pure elements as references is the technique usually applied. The idea was first proposed by Reid and Boyde19 and Robinson.20 This methodology was further enhanced by Bloebaum et al.3, 4 and Boyce et al.5 Boyce et al.5 were the first to make the gray-level values consistent within a single SEM session. This was further improved by Vajda et al.28 toward consistency between multiple SEM sessions. They used aluminum together with magnesium were applied as calibration standards. Skedros et al.25, 26 were working on the correlation between mineral content and BE signal using simulated bone tissue of known mineral content. Roschger et al.23 correlated BE gray levels of bone with calcium content (given in weight percent Ca) based on the Ca Kα-line intensities detected from identical bone areas. In addition, they calibrated the BE signal by carbon and aluminum, thus covering the whole Z range of bone between osteoid and a theoretically hypermineralized bone with 34.9 wt% Ca. Vajda et al.27 could confirm the high correlation between BE gray level and wt% Ca measured with EDX as found by our group.23 Furthermore, they tried to validate the EDX measurements by traditional ash weight measurements.
In particular, when the BE technique is used to measure the distribution of mineral density within a given biopsy (e.g., for diagnostic purposes), all factors that may increase the width of this distribution have to be evaluated carefully. First, there may be contributions due to the technique itself, which can be determined by repeatedly measuring the BMDD of the same bone area during a given session (intraassay variance) or in different sessions (interassay variance). Moreover, one has to know how BMDD varies between different areas of the same bone (intraindividual variation) or when specimens are taken from different individuals (interindividual variation).
To estimate all the influences just mentioned, and to validate the method for routine clinical studies, we investigated the BMDD of human biopsies and necropsies from 20 individuals by variance analysis. Furthermore, the Ca standardization of the BE signal using the X-ray fluorescence method23 was replaced by a technique using only the BE Z contrast, and a new method of evaluating BMDD histograms was developed. Finally, a BMDD histogram from a patient with a metabolic bone disease was measured and compared with the mean BMDD histogram evaluated from 20 bone-healthy individuals.
Section snippets
Materials and methods
Transiliac biopsies or necropsies from 20 individuals with accidental death (13 females, 7 males, ages 30–85 years) were used in this study. The medical history of the patients as well as the histomorphology of these bones showed no evidence of metabolic bone disease.18 In addition, one transiliac biopsy from a patient with osteomalacia due to celiac disease was added to the study.
Results
To use qBEI as an additional quantitative method for the clinical evaluation of bone biopsies, we tested the reliability of the Z calibration and the Ca standardization, and developed a new method for the evaluation of the BMDD histograms.
The calibration of the BE signal was performed by the measurement of BE gray levels of C and Al as references, thus establishing a linear relation between BE gray level and atomic number Z. For determining the linearity and accuracy of the calibration line
Discussion
The degree of mineralization and its topographical distribution are major determinants of the biomechanical quality of bone. SEM techniques9, 23, 26 have been developed to visualize and measure the amount and distribution of mineral in undecalcified bone samples. In this study, we have modified the calibration and standardization procedures, which reliably relate gray levels in the SEM images to wt% Ca of the mineralized matrix. In addition, we developed a new way to evaluate the BMDD
Acknowledgements
The authors thank Dr. S. Sieghart, Second Medical Department, Kaiserin Elisabeth Spital, Vienna, who referred the patient with osteomalacia to our institution for bone biopsy and G. Dinst and J. Thorvig for technical assistance.
References (30)
- et al.
Determining mineral content variations in bone using backscattered electron imaging
Bone
(1997) - et al.
Stereology and histogram analysis of backscattered electron imagesAge changes in bone
Bone
(1993) - et al.
Mineral density quantitation of human cortical iliac crest by backscattered electron image analysisVariations with age, sex, and degree of osteoarthritis
Bone
(1995) - et al.
Mineralization of cancellous bone after alendronate and sodium fluoride treatmentA quantitative backscattered electron imaging study on minipig ribs
Bone
(1997) - et al.
Studies on X-ray absorption and diffraction of bone tissue
Acta Anat
(1952) - et al.
Projection effect errors in biomaterial and bone research
Cells Mater
(1992) - et al.
Backscattered electron imagingThe role in calcified tissue and implant analysis
J Biomater Appl
(1990) - et al.
Reproducible method for calibrating the backscattered electron signal for quantitative assessment of mineral content in bone
Scann Microsc
(1990) - et al.
Scanning electron microscopy of boneInstrument, specimen, issues
Microsc Res Technol
(1996) - Boyde, A., Howell, P. G. T., Bromage, T., Elliot, J. C., Riggs, C. M., Bell, L. S., Kneissel, M., Reid, S. A.,...