复发性视神经脊髓炎脑扩散张量成像研究

Diffusion tensor imaging of brain in relapsing neuromyelitis optica

目的利用扩散张量成像(DTI)探讨复发性视神经脊髓炎(RNMO)患者是否存在隐匿性脑组织损伤及其发生机制。方法对16例 RNMO 患者和16例性别和年龄匹配的正常志愿者脑组织的平均扩散系数(MD)和部分各向异性(FA)进行直方图和感兴趣区(ROI)分析,以 P≤0.005为差异有统计学意义。结果与志愿者比较,RNMO 患者脑的平均 MD 升高[脑组织:RNMO 患者(0.95±0.02)×10^(-3)mm^2/s,志愿者(0.91±0.03)×10^(-3)mm^2/s,t=3.940,P<0.001;脑白质:RNMO患者(0.82±0.02)×10^(-3)mm^2/s,志愿者(0.80±0.02)×10^(-3)mm^2/s,t=3.117,P=0.004;脑灰质:RNMO 患者(1.06±0.04)×10^(-3)mm^2/s,志愿者(0.88±0.05)×10^(-3)mm^2/s,t=4.031,P<0.001]、脑白质的 FA 直方图峰高抬高[RNMO 患者(2.61±0.18)‰,志愿者(2.38±0.18)‰,t=3.627,P=0.001]及脑灰质的 MD 直方图峰高降低[RNMO 患者(7.37±0.89)‰,志愿者(8.91±1.71)‰,t=3.210,P=0.003]和峰位置抬高[RNMO 患者(0.83±0.02)×10^(-3)mm^2/s,志愿者(0.81±0.02)×10^(-3)mm^2/s,t=4.373,P<0.001];与脊髓和视神经有直接联系 ROI 的平均 MD 升高[延髓:RNMO患者(1.27±0.11)×10^(-3)mm^2/s,志愿者(1.11±0.10)×10^(-3)mm^2/s,t=4.260,P<0.001;大脑脚:RNMO 患者(1.01±0.11)×10^(-3)mm^2/s,志愿者(0.87±0.05)×10^(-3)mm^2/s,t=4.391,P<0.001;内囊:RNMO 患者(0.74±0.01)×10^(-3)mm^2/s,志愿者(0.72±0.01)×10^(-3)mm^2/s,t=4.683,P<0.001;视放射:RNMO 患者(0.88±0.03)×10^(-3)mm^2/s,志愿者(0.82±0.03)×10^(-3)mm^2/s,t=4.619,P<0.001)、平均 FA 降低(延髓:RNMO 患者0.27±0.01,志愿者0.29±0.03,t=2.996,P=0.005;大脑脚:RNMO 患者0.49±0.04,志愿者0.54±0.03,t=4.280,P<0.001;内囊:RNMO 患者0.66±0.02,志愿者0.69±0.02,t=3.953,P<0.001;视放射:RNMO 患者0.53±0.04,志愿者0.59±0.03,t=4.705,P<0.001);而与脊髓和视神经无直接联系的胼胝体的平均 MD[膝部:RNMO患者(0.76±0.04)×10^(-3)mm^2/s,志愿者(0.73±0.03)×10^(-3)mm^2/s;压部:RNMO 患者(0.77±0.05)×10^(-3)mm^2/s,志愿者(0.73±0.04)×10^(-3)mm^2/s]和 FA 值(膝部:RNMO 患者0.82±0.03,志愿者0.82±0.03;压部:RNMO 患者0.83±0.03,志愿者0.83±0.02)差异无统计学意义(P 值均>0.005)。结论 RNMO 患者存在隐匿性脑组织损伤,这可能与继发于脊髓和视神经病灶的顺行和逆行性变性有关。

Objective To investigate the presence of occult brain tissue damage in patients with relapsing neuromyelitis optica （RNMO） and its possible mechanism by using diffusion tensor imaging （DTI）.Methods DTI scans were performed in 16 patients with RNMO and 16 sex- and age-matched healthy controls. Histogram analysis of mean diffusivity （MD） and fractional anisotropy （FA） was performed in brain tissue （BT）, white matter （WM） and gray matter （GM） to detect the presence of occult brain tissue damage in RNMO patients. Region of interest （ROI） analysis of MD and FA was also performed in 6 dedicated regions with or without direct connection with spinal cord or optic nerve to determine the relationship between occult brain tissue damage and the damage of spinal cord and optic nerve. Results Patients with RNMO had a significantly higher average MD of the BT [ RNMO （ 0. 95 ± 0. 02 ） ×10^-3mm^2/s,controls（0.91 ±0.03） ×10^-3mm^2/s,t = 3.940, P〈0. 001], WM [RNMO（0.82 ± 0.02） ×10^-3mm^2/s,controls（0. 80 ±0.02） ×10^-3mm^2/s,t =3. 117, P = 0.004] and GM [RNMO （1.06 ±0.04） ×10^-3mm^2/s,controls（0. 88 ±0. 05） ×10^-3mm^2/s,t = 4.031, P 〈 0. 001 ], a lower MD peak height [ RNMO（7. 37 ±0. 89）‰,controls（8.91 ± 1.71 ）‰,t = 3. 210, P =0. 003] and a higher MDpeak location [RNMO（0.83 ±0.02） ×10^-3mm^2/s, controls （ 0. 81 ±0.02） ×10^-3mm^2/s,t = 4. 373, P 〈 0. 001 ] of the GM, and a higher FA peak height [ RNMO（2. 61 ± 0. 18）‰, controls （2. 38 s 0. 18）‰,t = 3. 627, P =0. 001 ] of the WM than did control subjects. In ROIs in direct connection with spinal cord or optic nerve, RNMO patients had a significantly higher average MD [ medulla oblongata： RNMO（1.27 ±0.11） ×10^-3mm^2/s, controls （1. 11 ±0.10） ×10^-3mm^2/s,t = 4.260, P〈 0.001; cerebral peduncle： RNMO（1.01 ±0.11） ×10^-3mm^2/s, controls （ 0. 87 ±0.05） ×10^-3mm^2/s,t = 4.391, P〈 0.001; internal capsule： RNMO（0.74 ±0.01） ×10^-3mm^2/s,eontrols（0.72 ±0.01） ×10^-3mm^2/s,t = 4.683, P 〈 0.001; and optic radiation： RNMO（0.88 ±0.03） ×10^-3mm^2/s,eontrols （0.82±0.03） ×10^-3mm^2/s, t = 4.619, P 〈 0.001] than did controls; RNMO patients had a significantly lower average FA （ medulla oblongata ： RNMO 0. 27 s 0. 01, controls 0. 29 ± 0. 03, t = 2. 996, P= 0.005; cerebral peduncle： RNMO 0.49 ± 0. 04, controls 0.54 ±0.03,t = 4. 280, P 〈 0.001; internal capsule： RNMO 0. 66 ±0. 02,controls 0. 69 ±0. 02,t = 3. 953, P 〈 0. 001 ; and optic radiation： RNMO 0. 53 ±0. 04,controls 0.59 ±0.03 ,t = 4. 705, P 〈 0. 001 than did controls. In ROIs without direct connection with spinal cord or optic nerve, however, no significant difference was found in average MD [ genu of corpus eallosum ： RNMO （0. 76 ± 0. 04） ×10^-3mm^2/s, controls （0. 73 ± 0.03 ） ×10^-3mm^2/s, t = 2. 804, P = 0. 009 and splenium of corpus eallosum： RNMO（0. 77 ± 0. 05） ×10^-3mm^2/s, controls （0. 73 ±0. 04） ×10^-3mm^2/s,t = 2. 234, P = 0. 033] and FA [genu of corpus eallosum： RNMO 0. 82 ± 0. 03,controls 0. 82 ±0. 03,t = 0. 196, P = 0. 846 and splenium of corpus callosum： RNMO 0. 83 ±0. 03, controls 0. 83 ±0. 02,t = 0. 333, P = 0. 741 ] between RNMO patients and controls. Conclusion RNMO patients have occult brain tissue damage, which might be related to the antegrade and retrograde degeneration secondary to lesions in the spinal cord and optic nerve.