A morphometric model to minimize subjectivity in the histological assessment of hepatocellular carcinoma and its precursors in cirrhosis

Objective: To explore how morphometry can minimize subjectivity in the assessment of liver nodules in cirrhosis using a novel classification tool. Study design: Ten hepatocellular carcinoma (HCC), 6 large regenerative nodules (LRN), and 34 regenerative (cirrhotic) nodules (RN), obtained from cirrhotic explant livers, were analyzed using a Kontron-Zeiss KS400 image analyzer. We generated a morphometric model based on the analysis of volume fractions occupied by hepatocyte nuclei/cytoplasm, sinusoidal endothelium and lumen, neoplastic acini, fibrosis, centrilobular veins, portal arteries, veins and bile ducts, individual lesional arteries (smooth muscle actin), and capillarized sinusoids (CD34), and on surface fraction occupied by reticulin, and number in unit volume and size distribution of hepatocyte nuclei, and mean hepatocyte nucleus diameter and volume. Results: Volume fraction of capillarized sinusoids and of individual lesional arteries were more prominent in HCC and LRNs, when compared with RN, whereas surface fraction of reticulin was markedly decreased in HCC. The morphometric values of these three features were integrated into our classification tool to construct a hybrid system, which reclassified the nodules in the same categories. Conclusion: Our novel hybrid classification tool may minimize subjectivity in the histological assessment of nodular lesions in cirrhosis