Feasibility of diffusion tensor imaging in lumbosacral neuropathy / 中华放射学杂志

ABSTRACT

Objective To study the feasibility of diffusion tensor imaging(DTI) in assessment of lumbosacral neuropathy. Methods Eighteen patients with lumbosacral neuropathy (experimental group) and 20 healthy volunteers (control group) were enrolled prospectively in the neurological department by clinical symptoms and neuroelectrophysiology. The experimental group were divided into experimental subgroup A(demyelination group, 6 cases),subgroup B(axonal injury with demyelination group, 12 cases) based on electrophysiology.All of L4-SN nerves of the experiment group and the control group received DTI examination."Sowing points"as ROIs were placed in the proximal, middle and distal of bilateral L4-SN nerves by two experienced physicians.The average of the DTI parameters of 3 ROIs at each nerve were used as its final DTI parameters.Independent sample t test was used to evaluate the difference of the L4-SN DTI parameters between the experimental group and the control group.The differences of DTI parameters among the control group, the experimental subgroup A and the experimental subgroup B were compared by the nonparametric independent samples Kruskal-Wallis H test and the Nemenyi test. The ROC analysis was performed on the FA values and λ⊥values of the control and experimental groups.Results (1)The L4-SN FA values in the experimental group was significantly lower than those in the control group (t range from 4.11 to 6.36, P<0.05). The L4 and S1-SN ADC values of the experimental group were significantly higher than those of the control group (t range from - 4.77 to - 1.17, P<0.05). The L4-SN λ ⊥ values in the experimental group were significantly higher than those in the control group (t range from -5.30 to -2.57，P< 0.05). (2) L4-SN FA values (H values ranged from 18.5 to 30.6, P<0.05), λ⊥ values (H values ranged from 6.8 to 29.2, P<0.05) and L4, S1-SN ADC values (H values ranged from 6.8 to 19.6, P<0.05) were significantly different among control group, experimental subgroup A and experimental subgroup B. Nemenyi test showed there were significant differences in FA values (H values ranged from 18.4 to 30.1,P<0.05) and λ⊥ values (H values ranged from 6.2 to 29.0, P<0.05)) of L4-SN and in ADC values (H values ranged from 6.4 to 19.54, P<0.05) of L4，S1-SN between control group and experimental subgroup B. (3) The analysis of ROC of the experimental group and control group showed the AUC values of L4-SN FA values were 0.834, 0.745, 0.860, 0.772, 0.811 respectively, the sensitivity were 66.7%, 72.2%, 77.8%, 83.3% and 66.7% respectively, and the specificity were 91.7%,69.4%,80.6%,61.1% and 88.9% respectively. The AUC values of L4-SN λ⊥ values were 0.796, 0.656, 0.791, 0.701, 0.843 respectively, and the sensitivity were 55.6%, 50.0%, 66.7%, 97.2% and 88.9% respectively,and the specificity were 91.7% , 77.85% , 86.1% , 36.1% and 66.7% respectively. Conclusion DTI has the feasibility to evaluate lumbosacral neuropathy and it can identify patients with axonal injury with demyelination and healthy subjects.