Artificial Neural Networks for the Prediction of Response to Interferon Plus Ribavirin Treatment in Patients with Chronic Hepatitis C

Author(s): P. A. Maiellaro, R. Cozzolongo, P. Marino

Journal Name: Current Pharmaceutical Design

Volume 10 , Issue 17 , 2004

Become EABM
Become Reviewer
Call for Editor


Combined therapy using Interferon alfa (IFN) and Ribavirin (RIB) represents the standard treatment in patients with chronic hepatitis C. However, the percentage of responders to this regimen is still low, while its cost and side effects are elevated. Therefore, the possibility to predict patients response to the above treatment is of paramount importance. The progress in the field of informatics and its large use for decision making has led to the development of novel techniques related to the so-called Artificial Intelligence, even including artificial neural networks (ANNs). In chronic viral hepatitis data are lacking. By means of an artificial neural network (ANN), 300 patients treated with IFN plus RIB were retrospectively analyzed with the aim to predict the response to the treatment. One hundred patients resulted responders and 200 non-responders at the end of treatment and during the follow up. For evaluating the prediction of treatment response, six ANNs with 16 neurons of input, an hidden layer with 7 neurons and an output layer with one neuron were utilized. The ANN model generated a positive predictive value (i.e. posterior probability of treatment response) ranging from 57% to 75% while the negative one (i.e. posterior probability of no response to treatment) was comprised between 52% and 71%. The highest level of diagnostic accuracy was 70%. In conclusion, ANNs appear to be a promising tool in the prediction of treatment response in patients with chronic hepatitis C. However, additional prospective studies are necessary to ultimately validate this predictive method.

Keywords: hepatitis c, prediction, interferon, ribavirin, artificial neural network

Rights & PermissionsPrintExport Cite as

Article Details

Year: 2004
Page: [2101 - 2109]
Pages: 9
DOI: 10.2174/1381612043384240
Price: $65

Article Metrics

PDF: 5