Nutzung von Dual-Energy Computertomographie zur Bestimmung von Knochenmineraldichte unter Verwendung einer 3-Material-Zerlegung

Introduction: Osteoporosis is a systemic bone disease responsible for loss of bone strength, endangering life and life quality of especially the elderly. For osteoporosis screening, the density of calcium-hydroxylapatite within the bone is evaluated. This density is commonly named bone mineral density (BMD) and forms the primary parameter for bone strength evaluation. The two standard procedures based on x-ray imaging for measuring BMD are dual energy x-ray absorptiometry (DXA) and quantitative CT (qCT). Both procedures have their respective benefits and drawbacks. DXA is the gold standard. It is low-dose, and well researched with large international databases and established guidelines by the World Health Organization. The two-dimensional nature of DXA images is susceptible to errors by e.g. aortic calcifications or osteophytes. DXA only detects areal bone mineral density (aBMD). Use of qCT as a three-dimensional imaging method circumvents those problems. qCT yields volumetric bone mineral density (vBMD) and allows for separate examination of trabecular and cortical bone. Drawbacks of qCT include a higher radiation dose and the lack of material decomposition or internationally established evaluation guidelines. Until recently, a qCT-scan required the presence of a special calibration phantom on each scan. Considering the rising numbers of osteoporosis cases worldwide, research for improvements on the established screening procedures is worthwhile. Objectives: The objective of this work is the evaluation of a radiographic test procedure which avoids the problems of DXA and qCT as far as possible by augmenting the qCT procedure with a three-material iiidecomposition based on Dual Energy CT (DECT). Accuracy and precision have to be at least comparable with the gold standard. The possibilities of phantomless qCT and scan protocol and parameter independence need to be explored. These have the potential to enable opportunistic evaluation of CT data for BMD. The effect of the increasing number of contrast-enhanced DECT scans requires investigation. Material and Methods: DECT enables the adaption of a decomposition algorithm into the three primary components of trabecular bone: Red and yellow bone marrow and the bone mineral itself. For this purpose, a DXA calibration phantom (ESP, ”European Spine Phantom”, QRM GmbH, Möhrendorf, Germany) has been used in combination with state- of-the-art CT scanners (SOMATOM Force and SOMATOM Definition Flash, Siemens Healthcare GmbH, Forchheim, Germany). All DECT beam energy combinations have been calibrated and the effects of scan parameters for DECT and standard single energy CT (SECT) were analysed. The acquired data facilitated development of the necessary beam hardening correction, a material decomposition algorithm into three materials, and conversion functions for CT-values into vBMD. For validation, 29 patients who had DECT data and concurrent DXA scans were evaluated using this enhanced dual energy qCT procedure. Six of these patients were scanned with standard phantom-based qCT data available as well. The effect of intravenously (IV) administered contrast agent (CA) and of CA in the gastrointestinal tract were analyzed using 9 patient datasets each. Lacking a ground truth by bone ash spectroscopy, DXA was used as standard of reference. Results: A DECT-based BMD acquisition procedure (DE-qCT) was developed. Accuracy and precision errors are comparable to DXA ranging from < 1% on the latest CT device SOMATOM Force and optimal parameters to 2 − 3% during opportunistic screening on older hardware. aBMD and vBMD are inherently incomparable. This has been addressed by evaluating the medical indication classifications of healthy bone, osteopenia and osteoporosis yielded by DE-qCT and DXA. The highest Area-under-curve correlation of the BMD measurement procedures introduced in this work is just over 0.90 and significant. Another procedure is more strictly oriented on WHO guidelines and has a lower, statistically significant correlation of 0.73. CT data containing CA showed that IV-administered CA prevents meaningful evaluation for BMD, gastrointestinal CA does not show this effect. DE-qCT yielded at least equal, if not better results than qCT, when compared to the gold standard. During development it was confirmed that the qCT calibration phantom is no longer necessary on modern CT devices. Evaluation of CT data obtained fpr a different clinical purpose for BMD is possible. Conclusion: The described limitations to DXA and qCT were overcome using the DE-qCT procedure. Our study is limited by the low number of patients evaluated and the lack of a real ground truth. The low number of patients is also evident for the CA evaluations and the qCT comparison. It can be concluded that DE-qCT has a high, significant agreement with DXA even at low patient numbers concerning medical indication. It is necessary to conduct a study including the required statistically significant number of patients of different ages, sexes, bone health status, fracture history and with a significant number of bone ash spectroscopies included.