Hepatitis B virus (HBV) infection is one of the main leading causes of hepatocellular carcinoma (HCC) worldwide. However, how reverse transcriptase (rt) gene contributes to HCC progression remains uncertain.
We enrolled a total of 307 chronic hepatitis B (CHB) and 237 HBV related HCC patients from 13 medical centers. Sequence features comprised multi-dimensional attributes of rt nucleic acid and rt/s amino acid sequences. Machine learning (ML) models were used to establish HCC predictive algorithms. Model performances were tested in the training and independent validation cohorts using receiver operating characteristic (ROC) and calibration plots.
Random forest (RF) model based on combined metrics (10 features) demonstrated the best predictive performances in both cross and independent validation (RFAUC=0.96, RFACC=0.90), irrespective of HBV genotypes and sequencing depth. Moreover, HCC risk score for individuals obtained from the RF model (AUC =0.966, 95% CI=0.922-0.989) outperformed α-fetal protein (AUC=0.713, 95% CI=0.632-0.784) in identifying HCC from CHB patients.
Our study provides evidence for the first time that HBV rt sequences contain vital HBV quasispecies features in predicting HCC. Integrating deep sequencing with feature extraction and ML models benefits the longitudinal surveillance of CHB and HCC risk assessment.

© The Author(s) 2020. Published by Oxford University Press for the Infectious Diseases Society of America. All rights reserved. For permissions, e-mail: journals.permissions@oup.com.