Lindemann Maike E, Oehmigen Mark, Lanz Titus, Grafe Hong, Bruckmann Nils Martin, Umutlu Lale, Quick Harald H
High-Field and Hybrid MR Imaging, University Hospital Essen, University of Duisburg-Essen, Essen, Germany.
RAPID Biomedical GmbH, Rimpar, Germany.
Med Phys. 2021 Nov;48(11):6696-6709. doi: 10.1002/mp.15284. Epub 2021 Oct 27.
Aim of this study was to evaluate the use of computer-aided design (CAD) models for attenuation correction (AC) of hardware components in positron emission tomography/magnetic resonance (PET/MR) imaging.
The technical feasibility and quantitative impact of CAD-AC compared to computer tomography (CT)-based AC (reference) was investigated on a modular phantom consisting of 19 different material samples (plastics and metals arranged around a cylindrical emission phantom) typically used in phantoms, patient tables, and radiofrequency (RF) coils in PET/MR. The clinical applicability of the CAD-AC method was then evaluated on a 16-channel RF breast coil in a PET/MR patient study. The RF breast coil in this study was specifically designed PET compatible. Using this RF breast coil, the impact on clinical PET/MR breast imaging was systematically evaluated in breast phantom measurements and, furthermore, in n = 10 PET/MR patients with breast cancer. PET data were reconstructed three times: (1) no AC (NAC), (2) established CT-AC, and (3) CAD-AC. For both phantom measurements, a scan without attenuating hardware components (material probes or RF breast coil) was acquired serving as reference. Relative differences in PET data were calculated for all experiments.
In all phantom and patient measurements, significant gains in PET signal compared to NAC data were measurable with CT and CAD-AC. In initial phantom experiments, mean relative differences of -0.2% for CT-AC and 0.2% for CAD-AC were calculated compared to reference measurements without the material probes. The application to a RF breast coil depicts that CAD-AC results in significant gains compared to NAC data (10%) and a slight underestimation in PET signal of -1.3% in comparison to the no-coil reference measurement. In the patient study, a total of 15 congruent lesions in all 10 patients with a mean relative difference of 14% (CT and CAD-AC) in standardized uptake value compared to NAC data could be detected.
To ensure best possible PET image quality and accurate PET quantification in PET/MR imaging, the AC of hardware components such as phantoms and RF coils is important. In initial phantom experiments and in clinical application to an RF breast coil, it was found that CAD-based AC results in significant gains in PET signal compared to NAC data and provides comparably good results to the established method of CT-based AC.
本研究旨在评估计算机辅助设计（CAD）模型在正电子发射断层扫描/磁共振成像（PET/MR）中对硬件组件进行衰减校正（AC）的应用。
在一个模块化体模上研究了CAD-AC与基于计算机断层扫描（CT）的AC（参考标准）相比的技术可行性和定量影响，该体模由19种不同的材料样本（塑料和金属围绕圆柱形发射体模排列）组成，这些材料常用于PET/MR中的体模、检查床和射频（RF）线圈。然后在一项PET/MR患者研究中，对一个16通道RF乳腺线圈评估了CAD-AC方法的临床适用性。本研究中的RF乳腺线圈是专门设计为与PET兼容的。使用该RF乳腺线圈，在乳腺体模测量中系统评估了对临床PET/MR乳腺成像的影响，此外，还在n = 10例患有乳腺癌的PET/MR患者中进行了评估。PET数据重建了三次：（1）无AC（NAC），（2）既定的CT-AC，以及（3）CAD-AC。对于这两种体模测量，均采集了一次不包含衰减硬件组件（材料探头或RF乳腺线圈）的扫描作为参考。计算了所有实验中PET数据的相对差异。
在所有体模和患者测量中，与NAC数据相比，使用CT和CAD-AC可测量到PET信号有显著增加。在最初的体模实验中，与没有材料探头的参考测量相比，CT-AC的平均相对差异为-0.2%，CAD-AC为0.2%。应用于RF乳腺线圈表明，与NAC数据相比，CAD-AC导致PET信号显著增加（10%），与无线圈参考测量相比，PET信号略有低估，为-1.3%。在患者研究中，在所有10例患者中共检测到15个一致的病变，与NAC数据相比，标准化摄取值的平均相对差异为14%（CT和CAD-AC）。
为确保PET/MR成像中获得尽可能好的PET图像质量和准确的PET定量，对体模和RF线圈等硬件组件进行AC很重要。在最初的体模实验和对RF乳腺线圈的临床应用中，发现基于CAD的AC与NAC数据相比可使PET信号显著增加，并且与既定的基于CT的AC方法提供的结果相当。
