Assessment of liver cirrhosis for patients with Child's A classification before hepatectomy using dynamic contrast-enhanced MRI

AIM: To determine the feasibility of semi-quantitative haemodynamic parameters derived from dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) to assess liver fibrosis. MATERIALS AND METHODS: Seventy-five patients with Child's A classification (males/females=24/51; average age, 58 years; range, 30–80 years) received DCE-MRI 3 days prior to hepatectomy. Semi-quantitative haemodynamic parameters, including the wash-in slope, wash-out slope, and time-to-peak, were calculated from DCE-MRI data. Liver fibrosis of the resected non-tumour liver was graded pathologically from F0 (no fibrosis) to F6 (cirrhosis) in the regions corresponding to those assessed by DCE-MRI. RESULTS: The wash-out slope showed higher interobserver and intra-observer reliabilities than the wash-in slope and time-to-peak. There was a significant positive correlation between the wash-out slope and pathological grade of fibrosis (Spearman's correlation coefficient: r=0.5331, p<0.0001). The area under the receiver operating characteristic curve was 0.8066 when using the wash-out slope to differentiate cirrhosis (grade F6) from non-cirrhosis (grades F0–5). Using the cut-off point that maximised specificity, the sensitivity was 62.07%, specificity was 91.30%, positive predictive value was 81.81%, negative predictive value was 79.25%, and accuracy was 80%. CONCLUSIONS: The wash-out slope derived from DCE-MRI might be potentially useful in assessing liver cirrhosis in patients with Child's A classification before hepatectomy.