MR diffusion weighted imaging for quantification of liver fibrosis in patients with chronic viral hepatitis
10.3760/cma.j.issn.1005-1201.2010.01.017
- VernacularTitle:MR扩散加权成像评价慢性病毒性肝炎肝纤维化的临床研究
- Author:
Yu SHI
;
Qiyong GUO
;
Wei LIAO
;
Yue MA
;
Wenxu QI
- Publication Type:Journal Article
- Keywords:
Liver cirrhosis;
Fibrosis;
Diffusion magnetic resonance imaging;
Hepatitis
- From:
Chinese Journal of Radiology
2010;44(1):65-69
- CountryChina
- Language:Chinese
-
Abstract:
Objective The study was to evaluate DWI for quantifying liver fibrosis. Methods A total of 12 volunteers, 47 patients who had chronic HBV or HCV hepatitis and underwent liver biopsy [Scheuer score for fibrosis(S) and inflammation(G)] were enrolled in this study. They were scanned using a 1.5 T MR unit with b value of 0,250,500,750, 1000 s/mm~2. ADCs at b_(250-1000) and b_(500-1000) were the average ADCs of b=250, 500, 750, 1000 s/mm~2 and b=500, 750, 1000 s/mm~2. The studied the correlation between Scbeuer scores and ADC values, and conducted Mann-Whitney U test and Logistic regression to evaluate ADC for prediction of fibrosis scores. Results The average ADCs were (1.41± 0.11),(1.37±0.09), (1.27±0.05), (1.26±0.04), (1.22±0.06) mm~2/s respectively from SO to S4, stage at b=750 s/mm~2 (F=18.31, P<0.01). With the increase of fibrosis score, the average ADC decreased gradually, the two were better negatively correlated at b_(250-1000)(r=-0.727, P<0.01) than other b values. Using b_(750) and the two combined b values, the found significantly lower ADCs in S2 or greater versus S1 or less and in S3 or greater versus S2 or less fibrosis (P<0.01). The best predictor for S2 or greater was b_(750) with the largest AUC of 0.909, sensitivity of 85.7%, and specificity of 100.0% (ADC ≤1.35×10~(-3) mm~2/s). The best predictor for S3 or greater was b_(250-1000) with the largest AUC of 0.864, sensitivity of 69.6%, and specificity of 95.8% (ADC≤1.53×10~(-3) mm~2/s). Conclusion DWI can be a good predictor for scoring liver fibrosis for S2 or S3 stage above, while b_(750) and the combined b values are suitable for evaluation.
