Lamellar Thickness Measurements in Normal and Osteogenesis Imperfecta Human Bone, with development of a method of automated thickness averaging to simplify quantitation.

Abstract

PurposeLamellar bone that forms in moderate and severe osteogenesis imperfecta (OI) is often composed of structurally irregular lamellae compared to those in normal bone. Polarization light microscopy (PLM) demonstrates lamellar bone well but has rarely been used for quantitative studies; information available on normal bone lamellae tends to be variable and studies specifically assessing OI bone lamellae have not been done. We report on PLM histomorphometry quantifying bright and dark lamellar thicknesses in normal and OI bone. Manual measurements of individual lamellar thicknesses have been made on histologic sections using the cellSens image analysis system; in an effort to augment the number of measurements we also developed a method of automated thickness averaging in quantifying regions of lamellae. MethodsFemoral and tibial cortical bone fragments from 5 individuals 5 - 26 years old (without molecular bone disorders) and 8 individuals 5 - 16 years old with progressively deforming (Sillence III) OI were obtained. The fragments were decalcified, infiltrated in JB4 solution, embedded in JB4 plastic, sectioned at 5 thickness and stained with 1% toluidine blue for light and polarizing microscopy. Manual measurements: Strict criteria for measurement, primarily to eliminate oblique lamellae, included accumulations of 16-20 bright and dark lamellae under PLM with a relatively narrow range of thicknesses, flattened elliptical osteocytes along the longitudinal axis of the lamellae and canaliculi passing from the walls of the osteocyte lacunae at right angles to the lamellae. Histomorphometric measurements of bright and dark lamellae by PLM were made at 20X magnification. Automated measurements: A script for automated measurement of average lamellar thicknesses from PLM images was developed in MATLAB (Mathworks, Natick, MA) to make measurement faster and less subjective. The script isolates a region from an image for measurement and marks each pixel as either bright or dark based on a local average intensity threshold. It then takes multiple pixel measurements along the length of the lamellae in the image and returns the average thickness of each in m. Results1. OI bone mean lamellar thickness values are always less than those in normal bone. The mean value for all OI bright and dark lamellae combined is 1.80 {+/-} 0.72 m and the value in normal bone is 2.54 {+/-} 0.92 m. 2. Mean value for the bright lamellae is less than that for the dark lamellae in both normal and OI bone. The mean value for bright lamellae in OI is 1.47 {+/-} 0.53 m and for dark lamellae 2.18 {+/-} 0.72 m; in normal bone the mean value for bright lamellae is 2.06 {+/-} 0.54 m and for dark lamellae 3.07 {+/-} 0.96 m. The differences are statistically significant: between groups of normal and OI lamellae (p<0.001), normal and OI light bands (p<0.001), and normal and OI dark bands (p<0.001). 3. Ratio of mean values for bright/dark lamellar thicknesses is the same in OI and normal bone. The ratio in OI bone is 0.67 (range: 0.54 - 0.83) and in normal bone 0.67 (range: 0.60 - 0.88). 4. Validation of automated vs. manual datasets: For each lamella in the validation dataset, the percent difference between the automated and manual measurements was calculated. The mean of the absolute values of these percent differences was 18.9%, a statistically non-significant difference (p = 0.0518). Discussion and conclusionsLamellar bone that forms in moderate and severe OI is composed of thinner and less regular lamellae than those in normal bone. i) PLM histomorphometry shows mean lamellar thicknesses (bright and dark merged) are statistically significantly decreased in OI compared to normal bone as are bright and dark lamellar thicknesses measured independently. ii) The automated method can be adapted readily to the assessment process for lamellar thicknesses and is, most likely, more accurate since it averages a greatly increased number of measurements per individual lamella. iii) Lamellar thickness measurements can be helpful in assessing the effect of specific collagen mutations on OI bone synthesis and warrant inclusion in both research and clinical histomorphometric assessments.