Arana E, Martínez-Granados B, Marti-Bonmati L, Martínez-Bisbal M C, Gil A, Blasco C, Celda B
Servicio de Radiologia, Hospital Quirón, Valencia.
Neurologia. 2007 Jun;22(5):267-74.
The objective is analyze the complementarity between 1H magnetic resonance spectroscopy (MRS) and magnetic resonance (MR) imaging in the global diagnosis of Alzheimer's disease (AD) or vascular dementia (VD).
We studied 168 patients with cognitive impairment from AD, VD, mild cognitive impairment (MCI) and major depression. All patients were evaluated by brain MR imaging and MRS using two sample volumes localized at right medial temporal gyrus and posterior parietal gyrus. Metabolites analyzed were N-acetylaspartate (NAA), myo-Inositol (mI), Choline (Cho) and creatine (Cr), as standard references for obtaining the Co/Cr, mI/Cr and NAA/Cr ratios. Imaging and spectroscopy alterations were graded from 0 to 4 and the average of both was used to draw ROC and SROC curves. Area under ROC curve (Az) was used as a measure of discriminative ability.
Combination of MR imaging and MRS significantly improved AD diagnosis (Global Az: 0.722 vs. MR imaging Az: 0.624; p: 0.003). However, the combination of MR imaging and MRS did not improve VD diagnosis. SROC curve obtained for the diagnosis of global dementia was Az: 0.6658 with 0.67 sensitivity and 0.65 specificity.
Combination of both MR techniques significantly improved AD diagnosis versus MR imaging alone. More studies are needed to enhance VD classification. Metabolic data found by MRS can be useful to differentiate cognitive impairment
目的是分析氢质子磁共振波谱(MRS)与磁共振(MR)成像在阿尔茨海默病(AD)或血管性痴呆(VD)整体诊断中的互补性。
我们研究了168例患有认知障碍的患者,这些患者分别患有AD、VD、轻度认知障碍(MCI)和重度抑郁症。所有患者均通过脑MR成像和MRS进行评估,使用两个位于右侧颞中回和顶叶后回的样本体积。分析的代谢物为N-乙酰天门冬氨酸(NAA)、肌醇(mI)、胆碱(Cho)和肌酸(Cr),作为获得Co/Cr、mI/Cr和NAA/Cr比值的标准参考物。成像和波谱改变从0到4进行分级,两者的平均值用于绘制ROC和SROC曲线。ROC曲线下面积(Az)用作判别能力的度量。
MR成像和MRS的联合显著改善了AD的诊断(整体Az:0.722 vs. MR成像Az:0.624;p:0.003)。然而,MR成像和MRS的联合并未改善VD的诊断。用于整体痴呆诊断的SROC曲线的Az为0.6658,敏感性为0.67,特异性为0.65。
与单独的MR成像相比,两种MR技术的联合显著改善了AD的诊断。需要更多的研究来加强VD的分类。MRS发现的代谢数据可能有助于区分认知障碍