Can hepatocellular carcinoma tumor grade and microscopic vascular invasion be predicted with an artificial neural network?

Abstract

Background and aims. Recurrence after curative treatments of hepatocellular carcinoma on cirrhosis, strongly depends on tumor pathological features such as nuclear grade and microscopic vascular invasion but these data are accurately assessable only on the final surgical specimens. Aim of the present study was to develop an artificial neural network (ANN, a tool showing its greatest strength in case of non linear relationships), able to predict tumor grade and microvascular invasion on the basis of preoperative variables.