基于硅光电倍增管和传统光电倍增管两种机型PET/CT临床成像条件下性能研究

目的 研究新一代硅光电倍增管(SiPM)正电子发射断层扫描与X射线计算机断层成像设备(PET/CT)与传统光电倍增管(PMT) PET/CT空间分辨力及图像质量,探讨不同的PET光电转换器(PMT与SiPM)和Q.Clear算法对PET空间分辨力、定量精度和图像质量的影响。方法 分别使用GE Discovery Elite型PET/CT (PMT PET/CT)和GE Discovery Meaningful Insights (MI)型PET/CT (SiPM PET/CT)对椭圆分辨力模体和NEMA NU2-2018图像质量模体成像。使用OSEM+PSF+TOF (VPFX-S)算法对所有模体原始数据进行图像重建,MI采集的模体数据额外使用Q.Clear算法重建,β取值范围150~550,间隔100。计算分辨力模体中5条线源在3个层面的径向、切向和轴向半高宽(FWHM)并求平均值。计算图像质量模体中小球的恢复系数(RC)、对比度百分比(CRC)、对比噪声比(CNR)、背景变异百分比(PBV)、背景变异系数(BCV)和不同图像层面肺插件残余误差(RE)。结果 MI (VPFX-S)与Elite相比,在径向、切向和轴向FWHM分别降低4.25%~13.58%、7.00%~13.22%和6.02%~36.14%,小球RC_max无明显差异。MI (VPFX-S)提高小球CRC (10.17%~38.89%)与CNR (38.31%~94.95%),降低PBV (26.20%~33.82%)与BCV (31.29%~35.97%)。MI (Q.Clear)与MI (VPFX-S)相比,在径向、切向和轴向FWHM分别降低6.49%~45.02%、7.80%~35.60%和13.31%~36.80%,提高小球CNR (38.31%~94.95%),并降低PBV (26.20%~33.82%)与BCV (6.64%~10.31%),RC无明显差异。MI (Q.Clear)得到的CNR、RE和径向、切向、轴向FWHM均随β增大而增大,RC、CRC、PBV和BCV均随β增大而减小。结论 在使用VPFX-S重建条件下,与PMT PET/CT相比,新一代SiPM PET/CT在空间分辨力、热灶对比度与探测能力和背景噪声方面均有所提高。相较于OSEM,Q.Clear算法改善了空间分辨力、定量精度与图像质量,但与β的选择有关。β值影响Q.Clear算法带来的收益,对小病灶的最大定量值影响最为明显,这点对提高临床诊断能力尤为重要。

Performance evaluation of PET/CT based on silicon photomultiplier and photomultiplier tube under clinical imaging conditions

Objective To investigate the spatial resolution and image quality of positron emission tomography and X-ray computed tomography (PET/CT) based on the next-generation silicon photomultiplier (SiPM) and the conventional photomultiplier tube (PMT) and to explore the effects of different PET photoelectric transducers (PMT and SiPM) and the Q.Clear algorithm on the spatial resolution, quantitative accuracy, and image quality of PET/CT. Methods GE Discovery Elite PET/CT (PMT PET/CT) and GE Discovery Meaningful Insights (MI) PET/CT (SiPM PET/CT) were employed to scan the elliptical resolution phantom and NEMA NU2-2018 image quality phantom. Using the OSEM+PSF+TOF (VPFX-S) algorithm, image reconstruction was performed based on raw data of both phantoms. For the MI-acquired phantom data, additional reconstructions were conducted using the Q.Clear algorithm, with β values ranging from 150 to 550 and an increment of 100. For the elliptical resolution phantom, the radial, tangential, and axial full-width at half-maximum (FWHM) values of five line sources in three slices were calculated and averaged. For the image quality phantom, the recovery coefficient (RC), contrast recovery coefficient (CRC), contrast-to-noise ratio (CNR), percentage of background variability (PBV), background coefficient of variability (BCV) of the spheres, as well as the residual error (RE) of lung inserts at different image slices, were calculated. Results Compared to Elite, MI (VPFX-S) showed decreases in the radial, tangential, and axial FWHM of 4.25%-13.58%, 7.00%-13.22%, and 6.02%-36.14%, respectively; no significant difference in RC_max for the spheres; increases in spheres'' CRC and CNR of 10.17%-38.89% and 38.31%-94.95%, respectively, and decreases in spheres'' PBV and BCV of 26.20%-33.82% and 31.29%-35.97%, respectively. When compared to MI (VPFX-S), MI (Q.Clear) showed decreases in the radial, tangential, and axial FWHM of 6.49%-45.02%, 7.80%-35.60%, and 13.31%-36.80%, respectively; an increase in spheres'' CNR of 38.31%-94.95%; decreases in spheres'' PBV and BCV of 26.20%-33.82% and 6.64%-10.31%, respectively, and no significant difference in spheres'' RC. With an increase in the β value, the CNR, RE, and FWHM in radial, tangential, and axial directions derived from MI (Q.Clear) increased, while RC, CRC, PBV, and BCV decreased. Conclusions Under image reconstruction using VPFX-S, the next-generation SiPM PET/CT exhibits elevated spatial resolution, hot lesion contrast, detectability, and background noise compared to the PMT PET/CT. Compared to OSEM, the Q.Clear algorithm improves spatial resolution, quantitative accuracy, and image quality, with such improvement related to the β value. The β value affects the outcomes of the Q.Clear algorithm, especially the maximum quantitative value of small lesions. This is particularly important for improving clinical diagnostic capabilities.