Ishii Shirou, Shishido Fumio, Miyajima Masayuki, Sakuma Koutarou, Shigihara Takeshi, Tameta Tadanobu, Miyazaki Makoto, Kuroda Hiroshi
Department of Radiology, Fukusuhima Medical University Hospital, 1 Hikarigaoka, Fukushima, 960-1295, Japan.
Ann Nucl Med. 2009 Jan;23(1):25-31. doi: 10.1007/s12149-008-0210-8. Epub 2009 Feb 11.
We investigated the ability to discriminate between Alzheimer's disease (AD) and vascular dementia (VaD), and between AD and non-dementia using the program "easy Z score imaging system" (eZIS) developed by Matsuda et al., for the diagnosis of very early AD.
Of 201 patients, we investigated 12 patients with AD, 10 with VaD, and 9 with non-dementia, who underwent brain perfusion single-photon emission computed tomography by technetium-99m ethyl cysteinate dimer (99mTc-ECD) between February 2005 and September 2006. The sensitivity and specificity of the indicators of specific volume of interest (VOI) analysis, namely, severity, extent, and ratio were evaluated for the distinction of AD from VaD and non-dementia.
There was a significant difference in all the criteria for severity, extent, and ratio between AD and non-dementia cases and in the ratio between AD and VaD. Between AD and non-dementia, the sensitivity and specificity of severity were 100% and 45%, respectively, using the cutoff value of 1.19. When using the cutoff value of 14.2 for extent, the sensitivity and specificity were both 100%. Using the cutoff value of 2.22 for ratio, the sensitivity of 42% and specificity of 100% were demonstrated. When comparing AD with VaD, using the cutoff value of 2.22 for ratio, the sensitivity and specificity were 42% and 100%, respectively. Using the cutoff value of 1.5 for ratio, the sensitivity and specificity between AD and VaD were 92% and 80%, respectively, thereby showing the best results.
The specific VOI analysis program of AD using specific voxel-based Z score maps is not influenced by interobserver differences among radiologists and is useful to discriminate AD from VaD and non-dementia. However, the setting of the cutoff value at each institution and comparison with original and eZIS images are suggested to distinguish better AD from VaD.
我们使用松田等人开发的“简易Z评分成像系统”(eZIS)程序,研究区分阿尔茨海默病(AD)与血管性痴呆(VaD)以及AD与非痴呆的能力,用于极早期AD的诊断。
在201例患者中,我们研究了2005年2月至2006年9月期间接受锝-99m乙半胱氨酸二聚体(99mTc-ECD)脑灌注单光子发射计算机断层扫描的12例AD患者、10例VaD患者和9例非痴呆患者。评估了感兴趣区(VOI)分析指标,即严重程度、范围和比率,用于区分AD与VaD和非痴呆的敏感性和特异性。
AD与非痴呆病例在严重程度、范围和比率的所有标准上以及AD与VaD在比率上存在显著差异。在AD与非痴呆之间,使用截断值1.19时,严重程度的敏感性和特异性分别为100%和45%。使用范围的截断值14.2时,敏感性和特异性均为100%。使用比率的截断值2.22时,敏感性为42%,特异性为100%。将AD与VaD比较时,使用比率的截断值2.22时,敏感性和特异性分别为42%和100%。使用比率的截断值1.5时,AD与VaD之间的敏感性和特异性分别为92%和80%,从而显示出最佳结果。
使用基于体素的特定Z评分图的AD特异性VOI分析程序不受放射科医生之间观察者差异的影响,有助于区分AD与VaD和非痴呆。然而,建议各机构设定截断值,并将原始图像与eZIS图像进行比较,以更好地区分AD与VaD。