Image analysis of liver biopsy samples measures fibrosis and predicts clinical outcome

Background & Aims Histopathological scoring of liver fibrosis mainly measures architectural abnormalities and requires a minimum biopsy size (⩾10 mm). Liver collagen quantification may allow use of small size biopsies and improve the prediction of clinical outcomes. This study evaluated the ability of the collagen proportional area (CPA) measurement to predict clinical outcomes. Methods Clinical outcomes were determined using population based data-linkage for chronic hepatitis C (CHC) patients from 1992 to 2012. Quantitative digital image analysis of liver biopsies was used for CPA measurement. Results 533 patients with a biopsy size ⩾5 mm were included. Median follow up was 10.5 years. 26 developed hepatocellular carcinoma (HCC), 39 developed liver decompensation and 33 had liver related death. 453 had Metavir F0-F2 and 80 had F3-F4. CPA ranged from 1.3% to 44.6%. CPA and Metavir stage were independently associated with liver related death. Metavir stage, CPA stage and age were independently associated with HCC. CPA stage (C1: 0%–5%, C2: 5%–10%, C3: 10%–20%, C4: >20%) stratified risk and a significant difference in outcomes was present between all CPA stages for HCC and between C2-C3 and C3-C4 for decompensation and liver related death. The 15 year composite endpoint-free survival was 97% for C1, 89% for C2, 60% for C3, 7% for C4. C4 had significantly worse survival than ⩽C3 ( p