The Value Of MRI And ~1H-MRS In Monitoring Globus Pallidus Injury By Bilirubin In Neonate

Objectives1. By a group of healthy neonates' conventional MRI and 1H-MRS of the brain to explore the method of MRI and 1H-MRS detection in newborns with acute bilirubin encephalopathy (ABE).2. To investigate the features of conventional MRI and 1H-MRS of the the globus pallidus in healthy newborns.3. To investigate changes of globus pallidus on conventional MRI and 1H-MRS in newborns with ABE, and whether the inter-brain metabolism had already been abnormal or not before encephalopathy occurred in newborns with hyperbilirubinemia.4. To evaluate correlation among the results of 1H-RS, Neonatal Behavioral Neurological Assessment(NBNA), and TSB crest of newborns with hyperbilirubinemia.Material and method1. PatientsTwenty seven full-term newborns (pregnanted for 37～42 weeks) with hyperbilirubinemia underwent conventional MRI scan of the brain and 1H-MRS scan of the basal ganglia region respectively in Nanfang hospital from March,2010 to February,2011.13 cases (male in 9, female in 4), who were diagnosed as acute bilirubin encephalopathy according to the definition of Shapiro, were incorporated in encephalopathy group. The other 14 patients (male in 9, female in 5) with hyperbilirubinemia and without any evidence of brain injury, were assigned to high risk group. Another 14 healthy newborns which could be precluded by any nervous system diseases (male in 8, female in 6), also underwent conventional MRI scan of the brain and 1H-MRS scan of the basal ganglia region with same parameters, as the control group.Inclusion criteria:Encephalopathy group:Definition of Shapiro was used as a standard, namely pathologic hyperbilirubinemia is proved by laboratory exam, brainstem auditory evoked potential (BAEP) is abnormal in various degree, and more than two kinds of neurological syndrome, including hypersomnia, unwillingness of breast-feedingl, screaming, hypertonia of limbs, opisthotonos, staring eyes and seizure, exist, and exclude hypoxic ischemic encephalopathy (HIE).High risk group:Patients with hyperestrinemia, the BAEP is normal, and without neurological syndrome were included, who were not suffered HIE, intracranial infection or congenital disease of nervous system.Control group:Newborns during the 28 days after birth with normal level of TSB, and without asphyxiation during rinatal period, disease of nervous system, or abnormalities of clinical and laboratory examinations, were included.2. MRI scanning devices and parametersMRI examination adopted the United States, GE signa excite 3.0T superconductive MRI scanners. Conventional MRI scan and 1H-MRS scan in the basal ganglia region used eight-channel head coil and single-channel orthogonal coil specialized for head respectively. Raw data and pictures acquired by 1H-MRS scan were imported into AW 4.3 workstation and were analyzed with its software "Functool".Conventional MRI scan location line was located in the layer of superciliary arch. The sequences included axial and sagittal scan of T1WI, axial scan of T2WI,and axial scan of T2Flair,and all those axial scan sequences set the canthomeatal line as base line.16 to 20 slices,varying according to the practical situation,were scanned from base to roof of skull, with the voxel thickness and spacing being 5mm and 0.5mm(4mm and 0.1mm for sagittal scan). Other parameters were set as follow: T1WISE (TR/TE,500/15), T2WIFSE (TR/TE,5100/130), T2Flair (TR/TE, 9602/114).1H-MRS imaging adopted the two-dimensional multi-voxel spectroscopic imaging acquisition mode, and the point resolved spectroscopy(PRESS) was selected as the favorable pulse sequences, TR/TE:1500/144ms, voxel spacing 20mm, thickness 10mm, NEX 1. The position image for MRS was established axially.Volume of interest (VOI) was located in a layer of lenticular nucleus where the globus pallidus on both side of the brain showed largest, and to include the bilateral temporal lobe and thalamus on both side, the size of VOI was ought to vary personally. Moreover, the region where ventricle cerebrospinal fluid lied in should also be avoided if possible.6 very selective suppressions (VSS) were set around the VOI positioned to reduce the partial effect of surrounding tissues which affects the quality of spectroscopy. Prescan was undertook after positioning on condition of the FWHM<10 and the water suppression (WS)≥98% to enable the MRS scan, or the procedure of shimming should be repeated. The scan time is about 6 minutes.3. Image post-processing and analysisThe data acquired by 1H-MRS scan were imported into AW4.3 workstation.With two-dimensional spectroscopic scanning sequences and position images had been chosen,the interface was converted from "post-processing" to "spectrum post-processing" by pressing the key "Functool", meanwhile four images were obtained,namely Chemical Shift Image (CSI), Spectrum Image (SI), Metabolic Image (MI), and metabolic and anatomical diagram integration map.The post-processing software automatically read all the voxels whose value were set at 31.6mm as the recommended value by computer within the VOI. Delete all the voxels after keeping one in the region of globus pallidus on each side. Activate a voxel on one side to "green" by select it, when what was seen on the spectroscopic image should be Spectrum Image, whose horizontal axis representing chemical shift situs of different metabolites was calculated by ppm,and whose longitudinal axis represented the wave crest with the area under it standing for the concentration of different metabolites.The area under the wave crest of one exact metabolite in its voxel, namely the absolute concentration of this metabolite,and the ratio among different metabolites can be calculated by post-processing software automatically.The major metabolites to be taken into investigation included N-acetylasparate (NAA), choline (Cho), creatine (Cr) and myo-inositol (mI), and the data to be analyzed were NAA/Cr, Cho/Cr, NAA/Cho and mI/Cr.4. Statistical analysisAll data obtained were indicated as the mean±standard deviation (x±s).SPSS 13.0 software was used for statistical analysis. The level of testαwas 0.05,and P<0.05 supported the consideration that the difference was statistically significant.The normality of all measurement data was tested by 1-Sample Kolmogorov-Smirnov TEST prior to statistical comparison.The proportions of gender among encephalopathy, high risk and control group were compared by the statistical mean of Pearsonχ2 test, and those of age, weight were done by mean of one-way ANOVA. Data from two sides of globus pallidus obtained by 1H-MRS scan, namely the four ratios of NAA/Cr,Cho/C,NAA/Cho and ml/Cr, were to paired according to sides and undergo the Paired-Samples T Test to testify their statistical significances. The data which were of no statistical significances between two sides would be averaged and then analyzed among the three groups(encephalopathy, high-risk and control group) by one-way ANOVA, and those which were of statistical significances between two sides would undergo one-way ANOVA respectively according to sides without averaging.If there were statistical significances among the three groups, then multiple comparison should be arranged (LSD for variance being homogeneous and Dunnett's T3 for variance being heterogeneous). Moreover, the correlation between NAA/Cr value with NBNA and TSB, and between NBNA with TSB of 27 newborns with hyperbilirubinaemia would be analyzed respectively (Pearson correlation if data on normal distribution and Spearman correlation on the other hand).Result1. Basic informationThere were no significant differences in age (F=2.790, P=0.074) among encephalopathy, high risk and control group, and the same case in gender(χ2=0.320, P=0.852) and weight (F=1.754, P=0.186)2. Findings of conventional MRI scanOn conventional MRI, the cerebral hemispheres of the 14 newborns from the control group were symmetrical and developed well, the cerebral convolutions of them had formed, and the cortical sulci on the surface of their cerebrums were relatively superficial.Ventricular systems, cisterns of lateral sulcus, basal cisterns and other important structures are all displayed clearly. On T2WI images, most white matter showed lower signal intensity, and grey matter showed higher. However, the white matter in pons, dorso-deutocerebrum, lateral region of abdomen of cerebral ganglion, posterior limb of internal capsule and around the third and fourth ventricle showed significant high signal intensity. On T2WI,white matter showed higher signal intensity and grey matters showed significantly lower.No evident abnormality and difference in comparison with the MRI images of newborns from the control group can be observed in those of 14 newborns from high risk group.The cerebral hemispheres of the 13 patient newborns were also symmetrical and developed well, important structures of them were clearly displyed too. The signal of their cortex and white matter were similer with the control group. While in 5 patients, significantly higher T1, signal intensity can be observed on the region of globus pallidus in contrast with their surrounding brain tissue, which were symmetrical and looked like the word "八". However, this kind of abnormality was invisible on T2WI and T2FLair images.3. Findings of1-MRS scanThe 1H-MRS images of 14 newborns from the control group showed high Cho crest, which was the most significant one, NAA crest, which was the second highest, and Cr crest. Glx and Lac crests were almost low or could not be discernmented. Similar signs could also be observed on those of encephalopathy group, despite a generally lower NAA, which could be significantly lower than Cr crest in some cases. And in high risk group a few cases showed a slightly lower NAA likewise.NAA/Cr values:The NAA/Cr values were 1.062±0.165 in control group, 0.843±0.175 in encephalopathy group, and 1.002±0.189 in high risk group. One-way ANOVA suggested statistically significant differences among these groups (F=5.486, P=0.008). LSD-t test was adopted to make multiple comparison, and showed that there were statistical significances between encephalopathy and control group (P=0.003), and between encephalopathy and high risk group (P=0.025). However, the difference between high risk and control group was not statistically significant(P=0.377).Cho/Cr values:The Cho/Cr values were 2.001±0.417 in control group, 1.997±0.188in encephalopathy group, and 1.801±0.240 in high risk group, and there were no statistical significances among them (F=2.003, P=0.149).NAA/Cho values:The NAA/Cho values were 0.547±0.113 in control group, 0.411±0.069 in encephalopathy group, and 0.570±0.124 in high risk group, and there were statistical significances among them. NAA/Cho value of encephalopathy group was statistically lower than those of other two groups (P=0.002 versus control group, P=0.000 versus high risk group), while the difference between high risk and control group was not statistically significant (P=0.571)ml/Cr value:ml/Cr values were 0.367±0.107 in control group,0.439±0.171 in encephalopathy group, and 0.381±0.095 in high-risk group, and there were no statistical significances among them (F= 1.199, P=0.313).4. Correlation analysis among NBNA,TSB and NAA/CrK-S goodness-of-fit test demonstrate that the three values of NBNA, TSB and NAA/Cr of 27 newborns with hyperbilirubinaemia were all on normal distribution. Pearson correlation was adopted to analyze the correlativity between NAA/Cr value with NBNA and TSB, and between NBNA with TSB. Results suggested that NBNA and TSB were of significant correlation (r=-0.718, P=0.000),while there were no significant correlation between NAA/Cr value with NBNA (r=0.042, P=0.833), or NAA/Cr value with TSB (r=0.087, P=0.665)Conclusion1. On conventional MRI, the signal feature of brian in healthy newborns is different from in adults, by multi-sequence scan of conventional MRI, the shape and signal features of globus pallidus can be showed efficiently. By 2D multi-voxel 1H-MRS imaging, with the TE value as 144, the metabolism condition of chief metabolin in globus pallidus can be displayed well.2. ABE presents characteristic futures on conventional MRI, it would not induce morphological changes of important structures but bring about abnormalities of signal of local brain tissues. The hyperintensity signal of Globus pallidus on T1WI is important signs of ABE.1H-MRS as a non-invasive imaging detection method of metabolic changes in living tissue, has a high value of monitoring bilirubin encephalopathy and to provide more diagnostic evidence at the molecular level. For brain injury in the acute phase of bilirubin encephalopathy, 1H-MRS is more sensitive than conventional MRI, and can find metabolic abnormalities or dysfunctions of neurons and axons much earlier.3. NAA crest of encephalopathy group significantly decreased, but the Cho crest did not change significantly, which indicated that there might be not only neurons -axonal damage and functional changes, but also the retardation of bilirubin deposition to the the development of brain.4. Although 1H-MRS was sensitive in the patholoy of brain damage, it failed to reflect the overall function of the nervous systm conditon. It must be under comprehensive consideration of the clinical manifestations and imaging on the diagnosis and prognosis evaluation of bilirubin encephalopathy.