Combining quantitative susceptibility mapping to the morphometric index in differentiating between progressive supranuclear palsy and Parkinson's disease
Combining quantitative susceptibility mapping to the morphometric index in differentiating between progressive supranuclear palsy and Parkinson's disease
复制标题
DOI：
10.1016/j.jns.2019.116443
复制
发表时间：
2019-11-15
影响因子：
4.4
通讯作者：
Wang, Yi
中科院分区：
文献类型：
--
作者：
Azuma, Minako;Hirai, Toshinori;Wang, Yi
目的：探讨定量磁化率图(QSM)灰质深层磁化率值是否可作为鉴别进行性核上性瘫痪(PSP)和帕金森病(PD)的形态计量学指标的补充指标。材料与方法：PSP-(n=8)和PD患者(n=18)和18例年龄匹配的健康对照组接受QSM和3D磁化准备快速梯度回波(MPRAGE)序列检查。由两位神经放射科医师测量QSM图像上深部灰质结构的平均感受值(MSVS)和3D MPRAGE图像上中脑区域的平均感受度指数(MI)。采用方差分析、Scheffe检验和受试者工作特征(ROC)分析，以MSVS和MI来评估PSP、PD和对照组之间的差异和区别。结果：PSP组苍白球(GP)和黑质(SN)的MSVS显著高于PD组和对照组(P<0.05)。经ROC分析(PSP与PD)，GP的AUC值最大(0.903)。PSP组的MI显著小于PD组和对照组(P<；0.05)；AUC(PSP组和PD组)为0.917。以GP的MSV值为244 ppb，MI的截断值为74.0 mm(2)的决策树可以完全区分PSP和PD。结论：QSM图像上GP的MSV增加了MI对区分PSP和PD的价值。
Purpose: To determine whether the susceptibility value in the deep gray matter obtained by quantitative susceptibility mapping (QSM) provides additive value to the morphometric index for differentiating progressive supranuclear palsy (PSP) from Parkinson's disease (PD).Materials and methods: PSP- (n = 8) and PD patients (n = 18) and 18 age-matched healthy controls who underwent QSM and 3D magnetization-prepared rapid gradient echo (MPRAGE) sequences. The mean susceptibility values (MSVs) of the deep gray matter structures on QSM- and areas of the midbrain (morphometric index, MI) on 3D MPRAGE images were measured by two neuroradiologists. Analysis of variance, the Scheffe test and receiver operating characteristic (ROC) analysis were conducted to assess differences and discriminate among PSP, PD and controls by the MSVs and the MI. Using the MSV of a structure with the best area under the curve (AUC) and the MI, we created a decision tree to differentiate between PSP and PD.Results: The MSVs of the globus pallidus (GP) and substantia nigra (SN) were significantly higher in PSP than PD and the controls (p < .05). By ROC analysis (PSP vs PD), AUC was greatest (0.903) for the GP. The MI was significantly smaller in PSP than PD and the controls (p < .05); AUC (PSP vs PD) was 0.917. The decision tree using cutoff values of 244 parts per billion for MSV of the GP and 74.0 mm(2) for MI served to completely differentiate between PSP and PD.Conclusion: The MSV in the GP on QSM images adds value to the MI for differentiating PSP from PD.