Evaluation of postprocessing dual-energy quantitative computed tomography

CT scanners can be used to provide quantitative information on body composition. Its main application is for bone mineral content estimation within the lumbar vertebral body. This is usually done with a single-energy technique. The estimates obtained with this technique are influenced by the intravertebral fat content, which varies interindividually and with disease. Dual-energy techniques have been proposed to solve the fat-error, but their exact value is unknown. The aim of the studies presented in this thesis, is to evaluate different postprocessing dual-energy methods for quantitative computed tomography for bone mineral and fat content analysis within the trabecular region of the vertebral body. Comparison of these methods by transforming them to a standard set of equations, reveals that only two out of five methods would give optimal results (chapter 3). This is confirmed in phantom studies (chapter 4). In the phantom studies, two major problems for performing postprocessing DEQCT are encountered: 1) the accuracy of the tissue equivalence of reference materials or accuracy of tissue description and 2) the effective energy difference between the site of the vertebral body and the reference device. Using a patient simulation model, the influence of these disturbing factors on the accuracy of bone mineral and fat content estimation is evaluated for different clinical conditions (chapters 5 and 6). The precision of the bone mineral and fat content determination with postprocessing DEQCT is evaluated in chapter 7. In addition to an in vitro experiment, the precision is estimated using the patient simulation model. The results from chapters 5 through 7, show that postprocessing DEQCT can be of value in clinical practice. Postprocessing DEQCT should be used in longitudinal studies on bone mineral content changes, if significant changes in the bone marrow composition are anticipated. Furthermore, postprocessing DEQCT can be used for evaluation of fat content differences between groups of patients. Striking differences in the fat content estimates are seen in a number of patients with different metabolic disorders (chapter 8). It is concluded that the theoretical superb accuracy of bone mineral measurements obtained with postprocessing DEQCT, can be eliminated by practical problems such as improper tissue equivalence of the reference materials and energy differences between the region of interest and the reference device. More research is necessary to obtain an exact knowledge of the elemental compositions and mass densities of the various anatomical structures within the vertebral body, especially intravertebral fat tissue. Beam hardening corrections, implemented in the CT scanner, should be evaluated for their effect on OCT