Sugiura Akie, Onoguchi Masahisa, Shibutani Takayuki, Kouno Yasuhisa
Kariya Toyota General Hospital.
Kanazawa University, Japan.
J Nucl Med Technol. 2022 Jun 30. doi: 10.2967/jnmt.122.264058.
The counts per pixel of brain perfusion single photon emission computed tomography (SPECT) images depend on the administration dose, acquisition time or patient condition, and they sometimes become poor acquisition counts in daily clinical study. The aim of this study was to evaluate the effect of different acquisition counts on qualitative images and statistical imaging analysis and to determine the minimum acquisition counts necessary for accurate examinations. We performed a brain phantom experiment simulating normal accumulation of Tc -ethyl-cysteinate dimer (Tc-ECD) as a brain uptake of 5.5 %. The SPECT data were acquired in a continuous repetitive rotation. Ten types of SPECT images with different acquisition counts were created by varying the addition of the number of rotations. We used the normalized mean squared error (NMSE) and visual analysis. For the clinical study, we used 25 patients acquired in a continuous repetitive rotation, and created six brain images with different acquisition counts by varying the number of rotations added from 1 to 6. The contrast-to-noise ratio (CNR) was calculated from the mean counts with ROIs in gray and white matter. In addition, the severity, extent and ratio of disease-specific regions were evaluated as indices of statistical imaging analysis. For the phantom study, the curve of NMSE showed a tendency of convergence from approximately 23.6 counts/pixel. Furthermore, the visual score showed that images with 23.6 counts/pixel or more were barely diagnosable. For the clinical study, the CNR was significantly decreased at 11.5 counts/pixel or less. Severity and extent tended to increase with decreasing acquisition counts, and a significant increase was shown at 5.9 counts/pixel. On the other hand, there was no significant difference in ratio values among defferent acquisition counts. Based on comprehensive assessment of phantom and clinical studies, we suggested that 23.6 counts/pixel or more were necessary to keep image quality of qualitative images and to accurately calculate indices of statistical imaging analysis.
脑灌注单光子发射计算机断层扫描（SPECT）图像的每像素计数取决于给药剂量、采集时间或患者状况，在日常临床研究中有时会出现采集计数不佳的情况。本研究的目的是评估不同采集计数对定性图像和统计成像分析的影响，并确定准确检查所需的最低采集计数。我们进行了一项脑模体实验，模拟锝-乙基半胱氨酸二聚体（Tc-ECD）的正常蓄积，其脑摄取率为5.5%。SPECT数据通过连续重复旋转采集。通过改变旋转次数的增加量，创建了十种具有不同采集计数的SPECT图像。我们使用了归一化均方误差（NMSE）和视觉分析。对于临床研究，我们使用了25例通过连续重复旋转采集的患者，并通过将增加的旋转次数从1增加到6，创建了六幅具有不同采集计数的脑图像。从灰质和白质中感兴趣区域的平均计数计算对比度噪声比（CNR）。此外，评估疾病特异性区域的严重程度、范围和比例作为统计成像分析的指标。对于模体研究，NMSE曲线显示从约23.6计数/像素开始有收敛趋势。此外，视觉评分显示，计数为23.6计数/像素或更高的图像几乎无法诊断。对于临床研究，CNR在11.5计数/像素或更低时显著降低。严重程度和范围倾向于随着采集计数的减少而增加，在5.9计数/像素时显示出显著增加。另一方面，不同采集计数之间的比值没有显著差异。基于对模体和临床研究的综合评估，我们建议计数为23.6计数/像素或更高对于保持定性图像的质量和准确计算统计成像分析指标是必要的。
