Journal of the Formosan Medical Association (Jul 2015)

Serum proteome predicts virological response in chronic hepatitis C genotype 1b patients treated with pegylated interferon plus ribavirin

  • Yi-Hao Yen,
  • Jyh-Chwan Wang,
  • Chao-Hung Hung,
  • Sheng-Nan Lu,
  • Jing-Houng Wang,
  • Tsung-Hui Hu,
  • Kwong-Ming Kee,
  • Chang-Chun Hsiao,
  • Chuan-Mo Lee

DOI
https://doi.org/10.1016/j.jfma.2013.04.013
Journal volume & issue
Vol. 114, no. 7
pp. 652 – 658

Abstract

Read online

Whether serum proteome changes can predict treatment response in chronic hepatitis C remains unclear. We investigated the association between serum proteome changes and virological responses in chronic hepatitis C virus genotype 1b (HCV-1b) patients treated with pegylated interferon (PegIFN) plus ribavirin (RBV). Methods: One hundred and thirty-six HCV-1b patients who had completed a course of PegIFN plus RBV for 24 weeks, had a 24-week follow-up, and had pretreatment serum available were enrolled. These patients were divided into training and validation groups. We used matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI–TOF/MS) for peptide profiling and ClinPro Tools version 2.0 bioinformatics software for data analysis. Results: Seventy-four patients (54%) had a sustained virological response (SVR), whereas 62 did not. We identified three protein peaks in pretreatment sera where the expression levels significantly differed between SVR and non-SVR (p < 0.05). Using the class prediction tool composed of the three protein peaks, we were able to correctly predict SVR in 95% of validation group patients with sensitivity = 95%, specificity = 56.3%, positive predictive value = 73.1%, and negative predictive value = 90%. We also identified a set of 20 protein peaks where the expression levels significantly differed in pretreatment sera between patients with nonresponse (NR) and virological response (SVR plus relapse; p < 0.05). Using the class prediction tool composed of these 20 protein peaks, we were able to correctly predict virological NR in 82% of validation group patients with sensitivity = 100%, specificity = 82%, positive predictive value = 92.6%, and negative predictive value = 100%. Conclusion: Pretreatment serum proteome allows prediction of SVR and NR to PegIFN plus RBV treatment in HCV-1b patients.

Keywords