1H-MR spectroscopy, magnetization transfer, and diffusion-weighted imaging in alcoholic and nonalcoholic patients with cirrhosis with hepatic encephalopathy

Abstract

Purpose: Mild swelling of astrocytes is proposed as a key event in the pathogenesis of hepatic encephalopathy. Proton MR spectroscopy ((1)H-MR spectroscopy), diffusion-weighted imaging (DWI), and magnetization transfer imaging were performed in patients with alcoholic and nonalcoholic liver cirrhosis and correlated with different clinical stages of hepatic encephalopathy to assess alterations in cerebral water metabolism in different subgroups of patients with cirrhosis.

Material and methods: Forty-five patients (26 alcoholics, 19 nonalcoholics [due to hepatitis C (n = 9), hemochromatosis (n = 2), primary chronic cholangitis (n = 2), hepatitis B (n = 1), Wilson disease (n = 1), cryptogenic cirrhosis (n = 4)]) and 18 controls underwent (1)H-MR spectroscopy, magnetization transfer imaging, and DWI of the basal ganglia and normally appearing occipital white matter (NAWM). N-acetylaspartate (NAA), choline (Cho), myo-inositol (mIns), and glutamine/glutamate (Glx) relative to creatine (Cr), the apparent diffusion coefficients (ADC), and the magnetization transfer ratios (MTR) were correlated to the neuropsychologic status, which was assessed by computerized psychometry and mental state grading, according to the West Haven criteria.

Results: Compared with controls, nonalcoholic subjects exhibited a gradual increase of Glx/Cr in the basal ganglia and NAWM; a decrease in mIns/Cr; a significant decrease of MTR in the thalamus, the putamen, the pallidum, and NAWM; and an increase in the ADC of the NAWM with increasing hepatic encephalopathy severity. In alcoholics, mIns/Cr of the basal ganglia and the NAWM, Cho/Cr of the basal ganglia, and MTR of all assessed regions were decreased. Glx/Cr of the basal ganglia and of the NAWM was increased, compared with that of controls; but no correlation to the clinical hepatic encephalopathy grading was found. ADC did not change significantly between the groups.

Conclusions: Apart from a typical pattern of (1)H-MR spectroscopy alterations in hepatic encephalopathy, a gradual decrease in MTR and an increase of ADC was found correlating to clinical grading of hepatic encephalopathy in nonalcoholic patients with cirrhosis. In alcoholic patients with hepatic encephalopathy, there was no such correlation. Abnormalities detected by MR imaging may hint at different pathways of brain damage in alcohol-induced liver disease.

Figures

Fig 1.

Graph shows hepatic encephalopathy grading and occipital white matter Glx/Cr in mean values; whiskers show SD. White matter Glx of alcoholics, graded mHE, HE 1, and HE 2 and white matter Glx of nonalcoholics with overt hepatic encephalopathy (HEO) are elevated, compared with that of controls (Co) (P < 0.05; Mann-Whitney U test).

Fig 2.

Graph shows hepatic encephalopathy-grading and occipital white matter MTR in mean values; whiskers show SD. White matter MTR of alcoholics, graded any hepatic encephalopathy, and white matter MTR of nonalcoholics with overt hepatic encephalopathy (HRO) are reduced, compared with that of controls (C0) (P < 0.05; Mann-Whitney U test).

Fig 3.

A, Nonalcoholics: occipital white matter Glx/Cr and MTR. Graph shows linear regression analysis (r2=0.48) and 95% predictive interval of means. B, Alcoholics: occipital white matter Glx/Cr and MTR. Graph shows linear regression analysis (r2 = 0.07) and 95% predictive interval of means.

Fig 4.

A, Nonalcoholics: occipital white matter Glx/Cr and mIns/Cr. Graph shows linear regression analysis (r2 = 0.34) and 95% predictive interval of means. B, Alcoholics: occipital white matter Glx/Cr and mIns/Cr. Graph shows linear regression analysis (r2=0.00) and 95% predictive interval of means.

Fig 5.

MR Spectra of posterior NAWM in a healthy control (A, patient 11), a nonalcoholic patient with cirrhosis (B, patient 55) with overt hepatic encephalopathy, and an alcoholic patient with cirrhosis (C, patient 59). Arrows indicate alterations in the patient’s spectrum, compared with that of the control group. Control (A): mIns/Cr, 0.926; NAA/Cr, 1.952; Cho/Cr, 0.301; Glx/Cr, 1.442. Nonalcoholic patient with cirrhosis (B): mIns/Cr, 0.038; NAA/Cr, 2.129; Cho/Cr, 0.255; Glx/Cr, 3.307. Alcoholic patient with cirrhosis (C): mIns/Cr, 0.260; NAA/Cr, 1.722; Cho/Cr, 0.249; Glx/Cr, 1.784.

Fig 6.

MTR map (width, 100%; center, 50%). A, Healthy volunteer (subject 8). B, Nonalcoholic patient with cirrhosis with HE 2 (subject 57). C, Alcoholic patient with cirrhosis with HE 2 (subject 59). MTR of volunteer (A): thalamus, 39.1%; pallidum, 36.4%; putamen, 37.3%; caudate, 37.5%. The following data refer to means of right and left side of the brain. MTR of nonalcoholic patients with cirrhosis (B): thalamus, 35.7%; pallidum, 27.0%; putamen, 31.1%; caudate, 27.7%. MTR of alcoholic patients with cirrhosis (C): thalamus, 34.8%; pallidum, 29.7%; putamen, 28.4%; caudate, 27.5%

Fig 7.

ADC map (width, 400%; center, 200%). A, Healthy volunteer (subject 14). B, Nonalcoholic patient with cirrhosis with HE 2 (subject 47). C, Alcoholic patient with cirrhosis and HE 2 (subject 50). ADC of volunteer (A): thalamus, 73.6; NAWM, 83.3. The following data refer to means of right and left side of the brain. ADC of nonalcoholic patients with cirrhosis (B): thalamus, 87.9; NAWM, 101.4. ADC of alcoholic patients with cirrhosis (C): thalamus, 80.6; NAWM, 108.