Pattern Recognition Lab, Friedrich-Alexander-University Erlangen-Nürnberg, Erlangen, Germany.
Invest Radiol. 2013 May;48(5):323-32. doi: 10.1097/RLI.0b013e318283292f.
Attenuation correction of positron emission tomographic (PET) data is critical in providing accurate and quantitative PET volumes. Deriving an attenuation map (μ-map) from magnetic resonance (MR) volumes is a challenge in PET/MR hybrid imaging. The difficulty lies in differentiating cortical bone from air from standard MR sequences because both these classes yield little to no MR signal and thus shows no distinguishable information. The objective of this contribution is 2-fold: (1) to generate and evaluate a continuous valued computed tomography (CT)-like attenuation map (μ-map) with continuous density values from dedicated MR sequences and (2) to compare its PET quantification accuracy with respect to a CT-based attenuation map as the criterion standard and other segmentation-based attenuation maps for studies of the head.
Three-dimensional Dixon-volume interpolated breath-hold examination and ultrashort echo time sequences were acquired for each patient on a Siemens 3-T Biograph mMR PET/MR hybrid system and the corresponding patient CT on a Siemens Biograph 64. A pseudo-CT training was done using the epsilon-insensitive support vector regression ([Latin Small Letter Open E]-SVR) technique on 5 patients who had CT/MR/PET triplets, and the generated model was evaluated on 5 additional patients who were not included in the training process. Four μ-maps were compared, and 3 of them derived from CT: scaled CT (μ-map CT), 3-class segmented CT without cortical bone (μ-map no bone), 4-class segmented CT with cortical bone (μ-map bone), and 1 from MR sequences via [Latin Small Letter Open E]-SVR technique previously mentioned (ie, MR predicted [μ-map MR]). Positron emission tomographic volumes with each of the previously mentioned μ-maps were reconstructed, and relative difference images were calculated with respect to μ-map CT as the criterion standard.
For PET quantification, the proposed method yields a mean (SD) absolute error of 2.40% (3.69%) and 2.16% (1.77%) for the complete brain and the regions close to the cortical bone, respectively. In contrast, PET using μ-map no bone yielded 10.15% (3.31%) and 11.03 (2.26%) for the same, although PET using μ-map bone resulted in errors of 3.96% (3.71%) and 4.22% (3.91%). Furthermore, it is shown that the model can be extended to predict pseudo-CTs for other anatomical regions on the basis of only MR information.
In this study, the generation of continuous valued attenuation maps from MR sequences is demonstrated and its effect on PET quantification is evaluated in comparison with segmentation-based μ-maps. A less-than-2-minute acquisition time makes the proposed approach promising for a clinical application for studies of the head. However, further experiments are required to validate and evaluate this technique for attenuation correction in other regions of the body.
正电子发射断层扫描（PET）数据的衰减校正对于提供准确和定量的 PET 容积至关重要。从磁共振（MR）容积中得出衰减图（μ-图）在 PET/MR 混合成像中是一个挑战。困难在于区分标准 MR 序列中的皮质骨和空气，因为这两个类别都几乎没有或没有 MR 信号，因此没有可区分的信息。本研究的目的有两个：（1）从专用 MR 序列生成并评估具有连续密度值的连续值计算机断层扫描（CT）样衰减图（μ-图）；（2）将其 PET 量化准确性与基于 CT 的衰减图进行比较，作为标准，并与其他基于分割的衰减图进行比较，以研究头部。
在西门子 3-T 西门子 Biograph mMR PET/MR 混合系统上对每位患者采集三维 Dixon 容积插值屏气检查和超短回波时间序列，在西门子 Biograph 64 上对每位患者采集相应的患者 CT。在 5 名同时具有 CT/MR/PET 三联体的患者上使用 epsilon-insensitive 支持向量回归（[拉丁小写字母开 E]-SVR）技术进行伪 CT 训练，并在 5 名未包含在训练过程中的额外患者上评估生成的模型。比较了 4 种 μ-图，其中 3 种来自 CT：缩放 CT（μ-图 CT）、不包括皮质骨的 3 类分割 CT（μ-图无骨）、包括皮质骨的 4 类分割 CT（μ-图骨），以及 1 种来自之前提到的 MR 序列的[拉丁小写字母开 E]-SVR 技术（即 MR 预测[μ-图 MR]）。使用之前提到的每个 μ-图重建 PET 容积，并计算相对于 μ-图 CT 的相对差异图像作为标准。
对于 PET 量化，与使用 μ-图无骨相比，所提出的方法分别产生了 2.40%（3.69%）和 2.16%（1.77%）的完整大脑和靠近皮质骨的区域的平均（SD）绝对误差。然而，尽管使用 μ-图骨的 PET 产生了 3.96%（3.71%）和 4.22%（3.91%）的误差，但使用 μ-图无骨的 PET 产生了 10.15%（3.31%）和 11.03%（2.26%）。此外，结果表明，该模型可以基于仅有的 MR 信息扩展到预测其他解剖区域的伪 CT。
本研究演示了从 MR 序列生成连续值衰减图，并在与基于分割的 μ-图进行比较的情况下评估了其对 PET 量化的影响。不到 2 分钟的采集时间使该方法有望在头部研究的临床应用中得到应用。然而，还需要进一步的实验来验证和评估该技术在身体其他部位的衰减校正中的应用。
