Non-invasive models for predicting histology in patients with chronic hepatitis B

Chun-Tao Wai1, Chee Leong Cheng, Aileen Wee

  • 1Department of Medicine, National University Hospital, Singapore. mdcwct@nus.edu.sg

Insights

Non-invasive models for predicting liver fibrosis in chronic hepatitis B (CHB) patients are unreliable. Existing models from chronic hepatitis C (CHC) studies showed poor accuracy in CHB patients, indicating a need for new diagnostic approaches.

Area of Science:

  • Hepatology
  • Gastroenterology
  • Medical Diagnostics

Background:

  • Chronic hepatitis B (CHB) requires accurate non-invasive methods to assess liver fibrosis and cirrhosis.
  • Limited research exists on non-invasive fibrosis prediction models specifically for CHB patients.
  • Existing models developed for chronic hepatitis C (CHC) may not be applicable to CHB.

Purpose of the Study:

  • To evaluate non-invasive markers for diagnosing significant fibrosis and cirrhosis in CHB patients.
  • To assess the accuracy of established non-invasive fibrosis models from CHC in a CHB cohort.

Main Methods:

  • Histological evaluation of liver biopsies using the Ishak score in treatment-naïve CHB patients.
  • Development and validation of non-invasive fibrosis markers using training and validation sets.
  • Assessment of Area Under the Receiver Operating Characteristic Curve (AUROC) for various markers and models, including AST, AST/ALT ratio, and APRI.

Main Results:

  • Significant fibrosis was present in 47% and cirrhosis in 19% of the 218 evaluated CHB patients.
  • Platelet count was the only significant predictor of fibrosis/cirrhosis in multivariate analysis, with modest AUROC values (0.63 and 0.73).
  • Non-invasive models derived from CHC studies demonstrated poor accuracy in the CHB cohort.

Conclusions:

  • Non-invasive fibrosis prediction models validated in CHC patients are unsuitable for diagnosing significant fibrosis or cirrhosis in CHB patients.
  • Simple, readily available non-invasive markers lack sufficient accuracy for predicting cirrhosis in CHB.
  • Further research is needed to develop reliable non-invasive diagnostic tools for CHB.
Abstract