Ahangari Sahar, Beck Olin Anders, Kinggård Federspiel Marianne, Jakoby Bjoern, Andersen Thomas Lund, Hansen Adam Espe, Fischer Barbara Malene, Littrup Andersen Flemming
Department of Clinical Physiology, Nuclear Medicine, and PET, Rigshospitalet, Copenhagen, Denmark.
Siemens Healthcare GmbH, Erlangen, Germany.
EJNMMI Phys. 2022 Aug 17;9(1):55. doi: 10.1186/s40658-022-00486-8.
Deep convolutional neural networks have demonstrated robust and reliable PET attenuation correction (AC) as an alternative to conventional AC methods in integrated PET/MRI systems. However, its whole-body implementation is still challenging due to anatomical variations and the limited MRI field of view. The aim of this study is to investigate a deep learning (DL) method to generate voxel-based synthetic CT (sCT) from Dixon MRI and use it as a whole-body solution for PET AC in a PET/MRI system.
Fifteen patients underwent PET/CT followed by PET/MRI with whole-body coverage from skull to feet. We performed MRI truncation correction and employed co-registered MRI and CT images for training and leave-one-out cross-validation. The network was pretrained with region-specific images. The accuracy of the AC maps and reconstructed PET images were assessed by performing a voxel-wise analysis and calculating the quantification error in SUV obtained using DL-based sCT (PET) and a vendor-provided atlas-based method (PET), with the CT-based reconstruction (PET) serving as the reference. In addition, region-specific analysis was performed to compare the performances of the methods in brain, lung, liver, spine, pelvic bone, and aorta.
Our DL-based method resulted in better estimates of AC maps with a mean absolute error of 62 HU, compared to 109 HU for the atlas-based method. We found an excellent voxel-by-voxel correlation between PET and PET (R = 0.98). The absolute percentage difference in PET quantification for the entire image was 6.1% for PET and 11.2% for PET. The regional analysis showed that the average errors and the variability for PET were lower than PET in all regions. The largest errors were observed in the lung, while the smallest biases were observed in the brain and liver.
Experimental results demonstrated that a DL approach for whole-body PET AC in PET/MRI is feasible and allows for more accurate results compared with conventional methods. Further evaluation using a larger training cohort is required for more accurate and robust performance.
在集成式PET/MRI系统中，深度卷积神经网络已证明可作为传统PET衰减校正（AC）方法的替代方案，实现强大且可靠的PET衰减校正。然而，由于解剖结构的差异以及MRI视野有限，其全身应用仍具有挑战性。本研究旨在探讨一种深度学习（DL）方法，从迪克森MRI生成基于体素的合成CT（sCT），并将其用作PET/MRI系统中PET AC的全身解决方案。
15例患者接受了PET/CT检查，随后进行了从头部到足部的全身PET/MRI检查。我们进行了MRI截断校正，并使用配准后的MRI和CT图像进行训练和留一法交叉验证。该网络使用特定区域的图像进行预训练。通过进行体素级分析并计算使用基于DL的sCT（PET）和供应商提供的基于图谱的方法（PET）获得的SUV中的定量误差，以基于CT的重建（PET）作为参考，评估AC图和重建PET图像的准确性。此外，还进行了特定区域分析，以比较这些方法在脑、肺、肝、脊柱、骨盆骨和主动脉中的性能。
我们基于DL的方法对AC图的估计效果更好，平均绝对误差为62 HU，而基于图谱的方法为109 HU。我们发现PET与PET之间在体素水平上具有极佳的相关性（R = 0.98）。整个图像的PET定量绝对百分比差异，PET为6.1%，PET为11.2%。区域分析表明，PET在所有区域的平均误差和变异性均低于PET。最大误差出现在肺部，而最小偏差出现在脑和肝脏。
实验结果表明，PET/MRI中用于全身PET AC的DL方法是可行的，与传统方法相比能获得更准确的结果。为了获得更准确和稳健的性能，需要使用更大的训练队列进行进一步评估。
