Artificial Neural Network and Tissue Genotyping of Hepatocellular Carcinoma in Liver-Transplant Recipients: Prediction of Recurrence


    loading  Checking for direct PDF access through Ovid

Abstract

Background.Liver transplantation (LT) is the treatment of choice for early stage hepatocellular carcinoma (HCC) with excellent 5-year survival, with a recurrence rate after LT of 3.4%. An artificial neural network (ANN), combined with genotyping for microsatellite mutations/deletions (TM-GTP), was designed at the University of Pittsburgh to predict tumor recurrence with a discriminatory power of 85%. This study aims to validate the ANN/TM-GTP model on patients receiving transplants in a single center.Methods.Nineteen patients with HCC underwent LT at our center between 1999 and 2002 (mean follow-up of 49.3 months). The ANN/TM-GTP analysis was performed blindly to prognosticate the risk of HCC recurrence, which was then validated against the actual clinical outcomes.Results.Nineteen patients received transplants. The primary diagnosis was hepatitis C (n=16), cryptogenic cirrhosis (n=2), and autoimmune hepatitis (n=1). ANN/TM-GTP was applied to all patients. The combination of ANN/TM-GTP predicted three patients to suffer recurrence of HCC. All three had HCC recurrence within 39 months (11, 23, and 39 months) postLT and died. Fourteen patients were predicted not to have HCC recurrence, and none did. Two patients could not be classified and were termed indeterminate for recurrence.Conclusion.ANN/TM-GTP had a high discriminatory power (17/19, 89.5%) in our cohort, accurately predicting HCC recurrence.

    loading  Loading Related Articles