| ObjectivesTo determine normative values of DTI parameters in healthy population, and analyze their distribution, variation and influencing factors. To investigate the changes of the DTI parameters in patients with cervical intraspinal tumors,and study their association with spinal cord function and prognosis. To explore the correlation between SEP indicators and DTI parameters.Materials and MethodsConventional T1 WI, T2 WI and sagittal DTI scan of cervical spinal cord were conducted in 36 healthy adult volunteers(21 males and 15 females, aged 18 to 77 years, averaged 43.6 years), who had no neurological symptoms or positive signs, history of neurological diseases or MRI contraindications. DTI was performed with single-shot sensitivity-encoding echo planar imaging sequence(b value=1000s/mm2, number of diffusion gradient directions=12) by using 3.0T SIEMENS MR TRIOTIM. Through post-processing workstation provided by SIEMENS Company, regions of interest on sagittal views were placed in defferent levels of cervical spinal cord to measure FA and ADC values by a qualified radiologist. White matter fiber bundle was reconstructed for observation on its traveling and integrity.The following study was performed with study group and control group. The inclusion criteria for study group were adult patients hospitalized for suffering from cervical intraspinal tumors. Exclusion criteria were association with intracranial tumors or spinal cord tumor beyond cervical segment, pathological diagnosis of non-neoplastic diseases, without surgery, with MRI contraindications or failing follow-up.Twenty-two patients were enrolled finally into study group. In the meanwhile, twenty-two age-matched volunteers were selected as control group. Cervical DTI scan was performed in two groups three days before surgery, and study group was reviewed 1 month posteroperatively. Spinal cord function was assessed by Mc Cormick grading. The spinal cord was divided into area rostral to tumor, tumor area and area caudal to tumor. Cervical FA and ADC values of each level and area were measured and calculated. Differences between these values and their relationship with Mc Cormick grading were comparatively analyzed.Somatosensory evoked potential(SEP)detection was performed in two groups three days preoperatively and 1 month posteroperatively. By wrist median nerve short latency somatosensory evoked potentials, spinal cord sensory conduction function was assessed. The N9-N20 interpeak latency was set as SEP index. Compared the relationship between N9-N20 interpeak latency and Mc Cormick grading and the difference of SEP index value between two groups. Comparative analysis of discrepancy between preoperative and posteroperative value was made. Study group was futher divided into delayed latency group(lat+ group) and normal latency group(lat- group), and the relationship between SEP index and DTI parameters was explored.All data analyses were performed using SPSS 21.0 analysis sofeware.ResultsFeatures of ADC map: the cerebrospinal fluid was white, and normal spinal cord presented gray signal. Features of color-coded FA map: cerebrospinal fluid was red and normal spinal cord was uniform blue. Diffusion tensor tracking(DTT) could clearly show the white matter fiber bundle presented as rostral-caudal direction in cervical spinal cord.Normal spinal cord fibers confirmed by T2 WI images had full and complete structure. DTT image possessed 3D effect, and fiber bundles could be observed at any angle of rotation. As quantitative parameters of DTI, ADC and FA values showed a significant negative correlation( r =-0.559, P<0.001).There was no significant difference between whole cervical FA and ADC values of the female and male subjects(P>0.05). But there was a significantly negative correlation of FA(r =-0.801, P<0.001) and a significantly positive correlation of ADC(r = 0.426, P=0.010) with age within normal spinal cord. There was no significant difference of FA and ADC values in various levels of cervical spinal cord(P>0.05).Applied to study group, DTI was able to distinctly show tractography full or sparse, intact or interrupted.Through this technique, spacial relationship(displacement or encasement) between the tumor and fiber bundles could be also observed.Pattern of tractography correlated with the prognosis, though unrelated with Mc Cormick grading(P> 0.05). Spinal cord function of patients with intact form in tractography was better than interrupt form one month operatively(P<0.05). In study group, FA values in area rostral to tumor, tumor area and area caudal to tumor were significantly lower than control group(P<0.05),and ADC values significantly higher than control group(P<0.05).There wassignificant regions difference of FA and ADC values. FA values in area rostral to tumor were lower than area caudal to tumor(P=0.006), and ADC values in area rostral to tumor were higher than area caudal to tumor(P=0.019). In control group, there was not prominent regions difference of FA and ADC values. Postoperative whole spinal FA values significantly increased(P<0.05) and ADC values significantly decreased(P<0.05).One month postoperative FA values in area rostral to tumor and tumor area were both significantly higher than preoperative values(P<0.05), and the degree of recovery in area rostral to tumor was larger than the tumor area(P<0.05). In area caudal to tumor, there was also a certain degree of recovery in FA values, but the difference was not statistically significant(P>0.05). Similar recovery occurred in postoperative ADC values.Preoperative FA and ADC values in study group did not correlate with Mc Cormick grading(P>0.05). In study group, improvement rate of patients with increasing FA values of area rostral to tumor was 93.8%, significantly higher than that of patients with decreasing FA values, which was 33.3%(P = 0.009).The change of postoperative FA values in area rostral to tumor was a sensitive prognostic factor to spinal cord function for patients with cervical intraspinal tumor, whose sensitivity,specificity, positive predictive value and negative predictive value was 88.2%,80.0%, 93.8% and 66.7% respectively.SEP was detected in two groups. Preoperative N9-N20 interpeak latency of study group was significantly longer than control group(P<0.05), but did not correlate with Mc Cormick grading(P>0.05). Postoperative N9-N20 interpeak latency of study group was significantly shorter than preoperative value(P<0.05). In term of preoperative fiber tract morphology of study group, the proportion of delayed N9-N20 interpeak latency in patients with interrupt type was higher than intact type(P=0.026). FA values of lat+ and lat- group in area rostral to tumor, tumor area and area caudal to tumor were significantly lower than control group(P<0.05). And FA values of lat+ group in these three regions were significantly lower than lat- group(P<0.05).ConclusionsThe values of FA and ADC are sensitive indices for quantitative analysis of cervical spinal cord. Form of white matter fiber bundles visualized through DTT can be used to determine the prognosis of cervical intraspinal tumors. Cervical intraspinal tumors are characterized by significant change of DTI indicators over the full range of spinal cord.These changes postoperatively recover with various degrees in different regions of the cervical spinal cord, which is directly related to the improvement of spinal cord function.In particular, postoperative change of FA value in area rostral to tumor is highly sensitive prognostic predictor for spinal cord function. DTI metrics correlate with SEP measures. |
PDF Full Text Request