Department of Diagnostic Radiology, Yale University, New Haven, Connecticut 06520, USA.
Med Phys. 2011 May;38(5):2715-23. doi: 10.1118/1.3582692.
We present a method to correct respiratory motion blurring in PET/CT imaging using internal-external (INTEX) motion correlation. The internal motion of a known tumor is derived from respiratory-gated PET images; this internal motion is then correlated with external respiratory signals to determine the complete information of tumor motion during the scan.
For each PET/CT data, PET listmode data were phase-gated into five bins and reconstructed. The centroid of a targeted tumor in each bin was determined and correlated with the corresponding mean displacement of externally monitored respiratory motion signal. Based on this correlation, the external motion signal was converted into internal tumor motion information in the superior-inferior direction. Then, the PET listmode data were binned sequentially to multiple 1-s sinograms. According to the converted internal tumor motion signal, each 1-s sinogram was registered to a reference frame, which best matched the helical CT attenuation map based on consistency conditions. The registered sinograms were summed and reconstructed to form an image, corrected for the motion of the specific tumor. In this study, the proposed INTEX method was evaluated with phantom and patient studies in terms of tracer concentration and volume.
The INTEX method effectively recovered the tracer concentration to the level of the stationary scan data in the phantom experiment. In the patient study, the INTEX method yielded a (17 +/- 22)% tumor volume decrease and a (10 +/- 10)% tumor SUVmax increase compared to non-gated images.
The proposed INTEX method reduces respiratory motion degradation of PET tumor quantification and delineation in an effective manner. This can be used to improve the assessment of response to therapy for a known tumor by minimizing residual motion and matching the attenuation correction, without increasing image noise.
我们提出了一种使用内外（INTEX）运动相关来校正 PET/CT 成像中呼吸运动模糊的方法。已知肿瘤的内部运动是从呼吸门控 PET 图像中得出的；然后将这种内部运动与外部呼吸信号相关联，以确定在扫描过程中肿瘤运动的完整信息。
对于每个 PET/CT 数据，将 PET 列表模式数据相位门控到五个bins 并进行重建。确定每个 bin 中目标肿瘤的质心，并与外部监测呼吸运动信号的相应平均位移相关联。基于这种相关性，将外部运动信号转换为上下方向的内部肿瘤运动信息。然后，将 PET 列表模式数据按顺序分为多个 1 秒正弦图。根据转换后的内部肿瘤运动信号，将每个 1 秒正弦图注册到参考帧，该参考帧根据一致性条件最佳匹配螺旋 CT 衰减图。注册的正弦图被求和并重建以形成图像，以校正特定肿瘤的运动。在这项研究中，通过在体模和患者研究中评估示踪剂浓度和体积来评估 INTEX 方法。
在体模实验中，INTEX 方法有效地将示踪剂浓度恢复到静止扫描数据的水平。在患者研究中，与非门控图像相比，INTEX 方法导致肿瘤体积减少（17 +/- 22）%，肿瘤 SUVmax 增加（10 +/- 10）%。
所提出的 INTEX 方法有效地减少了 PET 肿瘤定量和勾画中的呼吸运动退化。这可以通过最小化残留运动和匹配衰减校正来改善对已知肿瘤的治疗反应评估，而不会增加图像噪声。
