Fibrotic human liver slices remained viable for 24 hours and the gene expression of the fibrosis markers was stable up to 24 hours. As shown before, Imatinib inhibited in healthy and fibrotic rat PCLS the gene expression of Hsp47 (more than 50%) and Pcol1A1 (more than 80%), the protein expression of collagen I was inhibited by more than 40 %. In both healthy and fibrotic human PCLS imatinib did TAM Receptor inhibitor not have an effect on the gene expression

of fibrosis markers. In healthy human PCLS imatinib did not influence the protein expression of collagen I. In conclusion, clear species differences in the antifibrotic effect of imatinib were apparent. These results may explain why imatinib has not Src inhibitor reached the market as effective antifibrotic

drug. PCLS from human (fibrotic) liver tissue are a promising tool to study the efficacy of antifibrotic compounds in the early and end stage of liver fibrosis and are useful to reveal species differences in antifibrotic efficacy. Disclosures: The following people have nothing to disclose: Inge M. Westra, Dorenda Oosterhuis, Rick Mutsaers, Geny M. Groothuis, Peter Olinga Introduction. Determining the severity of liver fibrosis is important for care management in chronic liver diseases. Classical fibrosis stagings on biopsies are hampered by poor observer reproducibility. Our main aim was to develop a precise fibrosis classification based on automated morphometric diagnosis to avoid variability and increase diagnostic accuracy and precision compared to available fibrosis staging. Methods. 834 patients with chronic hepatitis C were included: 549 in the derivation population and 285 in the validation population. The pathological reference was Metavir fibrosis staging

by consensus between 2 experts. Automated morphometric analysis included the area of portoseptal fibrosis (Modern Pathology 2014) and 43 other descriptors providing scores for clinically significant fibrosis (CSF score by 5 descriptors) and cirrhosis (FM4 score by 6 descriptors). Different fibrosis classifications were derived from these scores according to published statistical merging. In the multicentric validation population, different Metavir stagings were available: this website initial local pathologists, 2 central experts and their consensus. Results. Accuracy (correctly classified patients) in the derivation population was: CSF classification (7 classes from CSF score): 94.2%; and FM4 classification (8 classes from FM4 score): 95.3%. The CSF/ FM4 classification was derived by combining the two previous classifications. We obtained 6 classes roughly reflecting the 5 Metavir stages with cirrhosis distinguished as early or definitive. CSF/FM4 classification combined the advantages of the individual classifications with an accuracy of 96.2%. The classification reproducibility was very high with intra-class correlation coefficient =0.988.