Quantitative analysis using the star volume method applied to skeleton patterns extracted with a morphological filter

Abstract

In this study, a morphological filter was combined with star volume analysis and applied to digital images to determine its potential usefulness in assessing trabecular structure. Three digital "geometric" test patterns (square, rectangle, circle) were created on a CRT (cathode ray tube). Each shape was arranged into five groups by size to yield 15 final "skeletal" patterns that were subsequently assessed with star volume analysis. Also, three digital X-ray images (background, soft tissue, bone block) were processed with a morphological filter to create three sets of 11 skeletal patterns each. These patterns were also assessed with star volume analysis and the ratio of extracted skeletal elements (in pixel numbers) to total pixel numbers was expressed as the pixel percentage. Star volume analysis was then applied to these digital skeletal images to yield the volume of extracted "skeletal" trabecular elements (Vsk) and the volume of nonskeletal (marrow space) elements (Vsp). The Vsk and Vsp were compared for all the different skeletal patterns. The pixel percentages were then compared to the star volume results for the X-ray test patterns. The Vsk decreased and Vsp increased as the number of operations (n) increased for both digital X-ray images and the geometric test patterns when the X-ray images were depicted by pixel percentages. Also, all true bone test patterns were clearly different both visually and quantitatively when compared to the noise skeletons extracted from background and soft tissue. Therefore, as Vsk was increased, so was connectivity. It can be concluded that the application of morphological filters and star volume analysis may be a useful tool in quantitatively determining the characteristics and continuity of trabecular skeletal structures. Further studies involving a larger number of bone samples and using models to compare measurements of calculated versus actual volume should reveal the true potential of this method for evaluating bone structure and its relationship to bone strength and also increase the tools available for evaluating bone diseases such as osteoporosis.