论文部分内容阅读
目的研究一种基于标准MRI和扩散张量成像(DTI)的、用于鉴别多系统萎缩(MSA)和帕金森病(PD)的决策树。方法 26例PD和13例MSA病人行3T脑MRI和DTI检查。感兴趣区(ROI)为壳核、黑质、脑桥、小脑中脚(MCP)和小脑。测量线性、容量和DTI参数[各向异性分数(FA)和平均扩散系数(MD)]。建立一个三节点的决策树,设计第1节点目标为100%特异度,第2节点为100%敏感度,第3节点为敏感度和特异度的最佳组合。结果 MSA组和PD组间的9个参数(小脑中脚的平均宽度、FA、MD;脑桥前后径;小脑的FA值和体积;脑桥和平均壳核体积;黑质致密部平均FA值)差异具有统计学意义(P<0.05),选择小脑中脚平均宽度、脑桥前后径和小脑中脚的平均FA值这3个参数进入决策树。阈值分别为14.6mm、21.8mm、0.55。决策树的总体性能表现为敏感度92%、特异度96%、阳性预测值92%和阴性预测值96%。13例MSA病人中12例被准确区分。结论使用这些参数形成的决策树可描述性、前瞻性地区别MSA和PD。要点①MR成像能区分PD和MSA。②联合应用常规MRI和DTI可提高诊断准确度。③决策树可描述性、前瞻性地区别不同临床疾病。④决策树能可靠地区分PD和MSA。
Objective To study a decision tree based on standard MRI and diffusion tensor imaging (DTI) for the identification of multiple system atrophy (MSA) and Parkinson’s disease (PD). Methods 26 cases of PD and 13 cases of MSA patients underwent 3T brain MRI and DTI examination. Areas of interest (ROI) are putamen, substantia nigra, pons, middle cerebellum (MCP) and cerebellum. Linearity, volume, and DTI parameters [anisotropic fraction (FA) and average diffusivity (MD)] were measured. A three-node decision tree is constructed. The first node is designed to have 100% specificity, the second node is 100% sensitivity and the third node is the best combination of sensitivity and specificity. Results The differences of 9 parameters (average width of midfoot in cerebellum, FA, MD; anterior and posterior pons of cerebellum; FA value and volume of cerebellum; pachymetry and average putamen volume; mean FA of substantia nigra pars compacta) between MSA group and PD group (P <0.05). The three parameters of the mean width of the midbrain in the cerebellum, the mean anterior and posterior pons of cerebellum and the mean FA in the cerebellum were entered into the decision tree. Thresholds were 14.6mm, 21.8mm, 0.55 respectively. The overall performance of the decision tree is 92% sensitivity, 96% specificity, 92% positive predictive value and 96% negative predictive value. Thirteen of 13 MSA patients were accurately distinguished. Conclusion The decision tree formed using these parameters allows a descriptive and prospective distinction between MSA and PD. Points ① MR imaging can distinguish between PD and MSA. ② combined with conventional MRI and DTI can improve diagnostic accuracy. ③ decision tree descriptive, prospective to distinguish between different clinical diseases. Decision tree can reliably distinguish between PD and MSA.