Department of Radiology, Leiden University Medical Center, Albinusdreef 2, 2333 ZA Leiden, The Netherlands.
Med Phys. 2010 Feb;37(2):620-8. doi: 10.1118/1.3276777.
Metal prostheses cause artifacts in computed tomography (CT) images. The purpose of this work was to design an efficient and accurate metal segmentation in raw data to achieve artifact suppression and to improve CT image quality for patients with metal hip or shoulder prostheses.
The artifact suppression technique incorporates two steps: metal object segmentation in raw data and replacement of the segmented region by new values using an interpolation scheme, followed by addition of the scaled metal signal intensity. Segmentation of metal is performed directly in sinograms, making it efficient and different from current methods that perform segmentation in reconstructed images in combination with Radon transformations. Metal signal segmentation is achieved by using a Markov random field model (MRF). Three interpolation methods are applied and investigated. To provide a proof of concept, CT data of five patients with metal implants were included in the study, as well as CT data of a PMMA phantom with Teflon, PVC, and titanium inserts. Accuracy was determined quantitatively by comparing mean Hounsfield (HU) values and standard deviation (SD) as a measure of distortion in phantom images with titanium (original and suppressed) and without titanium insert. Qualitative improvement was assessed by comparing uncorrected clinical images with artifact suppressed images.
Artifacts in CT data of a phantom and five patients were automatically suppressed. The general visibility of structures clearly improved. In phantom images, the technique showed reduced SD close to the SD for the case where titanium was not inserted, indicating improved image quality. HU values in corrected images were different from expected values for all interpolation methods. Subtle differences between interpolation methods were found.
The new artifact suppression design is efficient, for instance, in terms of preserving spatial resolution, as it is applied directly to original raw data. It successfully reduced artifacts in CT images of five patients and in phantom images. Sophisticated interpolation methods are needed to obtain reliable HU values close to the prosthesis.
金属假体在计算机断层扫描（CT）图像中会产生伪影。本研究旨在设计一种高效、准确的金属原始数据分割方法，以抑制伪影，提高金属髋关节或肩关节假体患者的 CT 图像质量。
该抑制伪影技术包括两个步骤：在原始数据中进行金属物体分割，并使用插值方案替换分割区域的新值，然后添加经缩放的金属信号强度。金属分割直接在正弦图中进行，与当前在重建图像中结合使用 Radon 变换进行分割的方法不同，这种方法效率更高。金属信号分割是通过使用马尔可夫随机场模型（MRF）来实现的。应用并研究了三种插值方法。为了提供概念验证，本研究纳入了五名金属植入患者的 CT 数据，以及一个带有聚甲基丙烯酸甲酯（PMMA）、特氟龙、聚氯乙烯和钛插件的 phantom CT 数据。通过比较钛（原始和抑制后）和无钛插件 phantom 图像的平均亨氏单位（HU）值和标准偏差（SD）作为失真的度量，以及比较未校正的临床图像和抑制后的图像，对准确性和定性改善进行了定量和定性评估。
phantom 和五名患者的 CT 数据中的伪影被自动抑制。结构的整体可见度明显改善。在 phantom 图像中，该技术的 SD 接近未插入钛的情况下的 SD，表明图像质量得到了改善。校正后图像的 HU 值与所有插值方法的预期值不同。在插值方法之间发现了细微的差异。
新的抑制伪影设计效率高，例如，在保持空间分辨率方面，因为它直接应用于原始原始数据。它成功地减少了五名患者和 phantom 图像中的 CT 图像伪影。需要使用复杂的插值方法来获得接近假体的可靠 HU 值。
