Brain Metabolic And Functional Changes In Patients With Hepatic Cirrhosis: Magnetic Resonance Imaging Study

Objective: To evaluate whether the patients with hepatic cirrhosis have metabolite changes of cingulate cortex using MRS, and whether MRS changes of basal ganglia and cingulate cortex have significant differences, and to assess the correlation between the metabolite ratio of MRS, venous ammonia, and neuropsychiatric tests;to evaluate whether the patients with hepatic cirrhosis have abnormal cerebral changes using magnetization transfer imaging (MTI), and whether the changes have a correlation with venous ammonia, GPI, and neuropsychiatric tests;to explore neural basis of cognitive control deficient in patients with hepatic cirrhosis using functional magnetic resonance imaging (fMRI).Part I MR spectroscopy study of brain biometabolite in cirrhosis patientsMaterials and Methods: 52 patients with hepatic cirrhosis proven by pathology, biopsy, and clinical examinations and 30 healthy volunteers were included in this study. All subjects performed three kinds of neuropsychiatric tests, that is, number-connection test type A (NCT-A), digital symbol test (DST), and symbol digital test (SDT) before MR examinations. MRS data in cingulate cortexand basal ganglia were acquired by point-resolved echo spin spectroscopy sequence of single voxel 1-hydrogen MRS. The software automatically completed metabolite content measurement, giving the ratio of all metabolites using Cr as a reference, including NAA/Cr, Cho/Cr, mlns/Cr, Glx/Cr, and the ratio of (Cho+mlns/NAA) was computed. Venous ammonia was measured in 44 cases.Results: (pCompared with controls, the patients had longer NCT-A reaction time (r=8.278, P=0.005), lower scores of DST (/=4.592, P=0.036) and SDT (r=4.529, P=0.029). ?Decreased Cho/Cr (/=4.470, PO.001), mlns/Cr (/=8.940, PO.001), (Cho+mIns)/NAA (f=4.509, PO.001), and increased Glx/Cr (/=-3.591, PO.001) in cingulate cortex were detected in the patients with hepatic cirrhosis;and for basal ganglia, decreased Cho/Cr (f=-6.688, PO.001), mlns/Cr (7=4.562, PO.001), (Cho+mIns)/NAA (/=4.821, PO.001) were found. ?mlns/Cr of cingulate cortex correlated negatively with Child-Pugh scale (r=-0.496, P<0.00\) and HE degree (r=-0.313, PO.05), implying it is a sensitive marker to perdict HE;Cho/Cr (r=-0.497, PO.001), mlns/Cr (r=-0.341, PO.05), and (Cho+mIns)/NAA (r=-0.276, PO.05) of basal ganglia correlated negatively with Child-Pugh scale, and Glx/Cr positively with HE degree (r=0.385, PO.05), suggesting the role of hyperammonia in basal ganglia changes. (4)MRS in cingulate cortex and basal ganglia can differentiate Child A type from B and C types, HE degree. ?Venous ammonia had a significant correlation with Cho/Cr (r=-0.329, PO.05), mlns/Cr (r=-0.347, .PO.05), and Glx/Cr (r=0.306, P<0.05) in cingulate cortex;and with (Cho+mIns)/NAA (r=-0.337, PO.05) and Cho/Cr (r=-0.457, PO.05) in basal ganglia;Cho/Cr and mlns/Cr correlated positively in cingulate cortex (r=0.340, P=0.014);hinting that hyperammonia had a role in brain changes for patients with cirrhosis. ?NCT-A had a correlation with Glx/Cr of cingulate cortex (r=0.366, P=0.028), suggesting cingulate cortex was a sensitive location to define brain changes of the patients with hepatic cirrhosis.Part II Brain changes in patients with hepatic cirrhosis: evaluation with magnetization transfer imagingMaterials and Methods: 48 patients with hepatic cirrhosis and 30 healthy volunteers were included into this study. All subjects performed axial Ti WI, T2WI, MTI with and without MTC. Signal intensities of putamen and globus pallidus were measured on TiWI to compute GPL In MTI, ROIs were placed on the white matter of parietal lobe, frontal lobe, caudate head, putamen, globus pallidus, thalamus, and occipital lobe. MTR of all locations were calculated according to the equation.Results: ?MTR of every location (the white matter of parietal lobe, frontal lobe, caudate head, putamen, globus pallidus, thalamus, and occipital lobe) for patients with hepatic cirrhosis were lower than those for controls (P<0.001), MTR of lobus pallidus decreased most, about 9%. (2)MTR of white matter of parietal lobe, frontal lobe, putamen, globus pallidus, and occipital lobe had a significant difference among various groups, and no significant difference in caudate head and thalamus. MTR of white matter of parietal lobe, frontal lobe, putamen, globus pallidus, and occipital lobe had a significant difference between Child-Pugh class A and C, class A and class B, and no significance between class B and class C. No significant difference for HE degree was found among three patient groups. (3)GPI and MTR of white matter of parietal lobe, frontal lobe, putamen, globus pallidus, and occipital lobe negatively correlated with Child-Pugh scale (P<0.05), only MTR of globus pallidus correlated negatively with HE degree (r=-0.291, P=0.045). ?MTR of white matter of parietal lobe, frontal lobe, putamen, globus pallidus, and occipital lobe had no correlation with venous ammonia, and correlated significantly with all neuropsychiatric tests performed (PO.05), whichpartial correlation test did not display their correlation with neuropsychiatric tests. Part III Cognitive control in patents with hepatic cirrhosis: an fMRI studyMaterials and Methods: 14 patients with hepatic cirrhosis and 14 healthy volunteers were included in this study. Modified Stroop task in Chinese character was used as target stimulus, block-design fMRI was used to acquire resource data, including 4 stimulus blocks and 5 control blocks, each block was alternatively present. Image analysis was performed using statistical parameter mapping 99. A P value of less than or equal to 0.01 and continuous activating voxels of more than or equal to 10 were defined as significant activating areas. One-sample / test was used for word-reading and color-naming task in patients with hepatic cirrhosis and controls, and two-sample / test was also used for inter-group analysis of word-reading and color-naming task in patients with hepatic cirrhosis and controls. After fMRI examinations were over, behavior tests of Stroop interference were performed for all subjects. Overall reaction time and error numbers were recorded, respectively. SPSS 10.0 version was used to perform statistical calculations for behavioral data.Results: (T)Both healthy volunteers and the patients with hepatic cirrhosis performed faster incongruous word-reading task than incongruous color-naming task (/=-8.188, PO.001;/=-12.447, PO.001), and the patients with hepatic cirrhosis performed more slowly both incongruous word-reading task and incongruous color-naming task than healthy volunteers (f=3.897, P^O.001;r=5.83, PO.001). Both healthy volunteers had similar errors for performing both tasks(f=-1.031, P=0.320);and the patients with hepatic cirrhosis had more errors in performing incongruous color-naming task than doing incongruous word-reading task (f=-7.089, PO.001), and the patients with hepatic cirrhosis hadmore errors in performing incongruous color-naming task than healthy volunteers (/=5.998, PO.001), and without differences for performing incongruous word-reading task (/=1.913, P=0.066). ?Healthy volunteers activated more brain areas when they performed incongruous color-naming task than incongruous word-reading task, and activation intensity was strengthened for the brain areas appeared in the incongruous word-reading task. When they performed incongruous color-naming task, bilateral middle frontal gyrus, bilateral superior frontal gyrus, bilateral inferior frontal gyrus, bilateral medial frontal lobe (extending to bilateral anterior cingulate cortex), bilateral parietal lobes, and bilateral temporal fusiform gyrus were activacted. ?Compared with controls, the patients with hepatic cirrhosis had more activation of bilateral prefrontal cortex and parietal cortex when performing incongruous word-reading task. With increased conflict, the activation of anterior cingulate cortex, bilateral prefrontal cortex, parietal lobe and temporal fusiform gyrus were decreased when the patients with hepatic cirrhosis performed incongruous color-naming task.Conclusions: ?Cingulate cortex and basal ganglia have an abnormal metabolite change for the patients with hepatic cirrhosis, presenting with decreased mIns/Cr, Cho/Cr, and (Cho+mIns)/NAA, and increased Glx/Cr. MRS of cingulate cortex appears to be the most sensitive marker to detect brain metabolite abnormality in patients with hepatic cirrhosis;it has a significant correlation with NCT-A and venous ammonia, suggesting the role of hyperammonia in pathogenesis of HE. (2)mIns/Cr and Cho/Cr of cingulate cortex in patients with hepatic cirrhosis are related to Child-Pugh scale of hepatic cirrhosis, and Glx/Cr of basal ganglia correlates positively with the degree of hepatic encephalopathy. ?The patients with hepatic cirrhosis have mild decreased MTR and increased GPI in the brain, which implies the mild brain edema. ?MTR and GPI correlatesignificantly with Child-Pugh classification and the degree of hepatic encepholapathy;however, they cannot differentiate different stage of HE. ?All of data support the hypothesis that hyperammonia has an important role in pathogenesis of HE, and the abnormal metabolite and MTR changes result from hyperammonia. ?Behavioral data and fMRI results have demonstrated that the patients with hepatic cirrhosis have cognitive control deficient, the abnormal brain network of cingulate cortex- prefrontal cortex-parietal lobe- fusiform gyrus is the neural basis of cognitive control impairment of hepatic cirrhosis patients.