PET/MR的图像融合精度测试研究

Research on accuracy of image fusion with hybrid PET/MR

目的:运用专用测试体模和Na源在模拟实际临床状态下测试PET/MR一体机(Philips Ingenuity TF PET/MR扫描仪)全身3D采集方式的图像融合精度。材料和方法:使用Philips公司提供的测试体模及其配套的Na源,先将Na源分别放置于模型的规定位置对采集和扫描的数据进行融合校准,再将相同的6个Na源分别固定于检查床两测,在检查床负重10 kg和60 kg时分别采集PET/MR图像数据。重建图像后,分别在PET和MR图像的横断层面勾画Na源的轮廓,计算出Na源在PET和MR上的重心位置。对两个位置进行对比,算出差值。结果:检查床负重10 kg时最大偏差出现在上段右侧,为3.01 mm;检查床负重60 kg时最大偏差出现在中段右侧,为4.66 mm。结论:本研究方法可以比较精确地测量PET/MR图像融合精度,并有效地指导PET/MR仪器的验收和质量控制等。

objective： To evaluate the accuracy of image fusion between PET and MR data by 3D whole body PET/MR scanner with the phantoms and Na point sources simulating clinical practice situation. Materials and Methods： Phantoms and Na point sources were provided by Philips company. The six Na point sources were placed at the fixed positions of the phan- toms for image fusion coregistration, and then were fixed on the two sides of the bed for simulating different positions of the patient. The PET and MR images were acquired at 2 different loading conditions： （1）uniformly loading phantoms of 10 kg on the table; （2）uniformly loading phantoms of 60 kg on the table. The Na point sources were manually outlined on axial MR and PET images. The difference between the point sources in PET and MR was obtained by calculating the distance of the re- spective center of gravities （COGs）. Results： The maximal difference was 3.01 mm in the upper and right part of the table, with loading phantoms of 10 kg. The maximal difference was 4.66 mm on the center and right part of the table, with loading phantoms of 60 kg. Conclusions： The accuracy of image fusion can be measured more objectively with these methods, and it can be used for the acceptance test and routine quality control of PET/MR scanner.