ABSTRACT
Osteoporosis is a multifactorial disease, and bone removal differs widely from patient to patient with considerable consequences for the respective bone mechanics. The general features of osteoporotic abnormalities include alterations in cancellous bone such as fenestration of bone plates, and thinning and reduction of bone trabeculae. Alterations in cortical bone such as thinning of its width due to endosteal bone resorption, and increased porosity of the Haversian canals
also occur. Dual-energy X-ray absorptiometry (DEXA) became so popular because most of the osteoporotic
changes lead to a reduction of bone mass. DEXA is based on projectional images and provides the mineral content or the areal density of the examined bone. It cannot differentiate between cortical and cancellous bone loss, and it is insensitive to structural changes within cancellous bone, such as reduction of bony elements and compensatory thickening of the remaining elements, transformation of plates to rods, or alterations in connectivity and anisotropy. Projectional images are not useless per se to analyze changes in cancellous bone structure. The proximal femur, for example, has a distinctive pattern of trabecular architecture, which is disturbed in the course of osteoporosis and can be described semiquantitatively with the Singh index. The Singh index
distinguishes six grades of osteoporosis by comparing radiographs of a patient with a set of standard radiographs or with charts. The procedure proved to be useful for epidemiological studies, but was not sufficiently reliable for diagnostic purposes. Considerable efforts have been made to analyze the trabecular pattern from radiographs of the radius and the calcaneus
with the help of image analysis tools such as histogram analysis, run length methods, and fractual dimensions, just to name a few. Plain film radiography, however, does not record the real architecture of bone, but an apparent one which is created by superposition of the X-ray shadows produced by calcified tissues.
