The value of HIFI model based on cell-free DNA genomic features in early diagnosis and prediction of hepatocellular carcinoma

Abstract

Background The detection of cell-free DNA (cfDNA) is a promising early diagnosis method for hepatocellular carcinoma (HCC). We validated the cfDNA-based comprehensive diagnostic model (HIFI) and explored its value in the early diagnosis and prediction of HCC. Methods cfDNA detection was performed in 126 patients, the results of HIFI were obtainedand compared with HCC markers. Moreover, we conducted a comparison between the predictive ability of aMAP score and the HIFI model for HCC at 10 months of follow-up. Results Chi-square test and Kappa analysis showed that the diagnostic consistency between HIFI model and traditional methods was significantly higher than that of AFP and PIVKA-II (Kappa=0.461 vs 0.251 vs 0.287, P≤0.05). Moreover, the consistency of HIFI model combined with PIVKA-II was higher than that of AFP combined with PIVKA-II and HIFI model combined with AFP (Kappa=0.564 vs 0.267 vs 0.317, P≤0.05). We compared HIFI model with aMAP score at 10 months, the sensitivity, specificity, PPV and NPV of HIFI model in predicting HCC was 70.59%, 90.32%, 66.67% and 91.80%. When 50/60 is set as the cut-off value of aMAP score, the sensitivity, specificity, PPV and NPV was 94.12 / 76.47%, 29.03 / 72.58%, 26.67 / 43.33% and 94.74 / 91.84%. Conclusions The study demonstrated the superiority of the cfDNA-based HIFI model as a method of early warning and diagnosis of HCC. Moreover, the predictive efficacy of the HIFI model remained high at 10 months, in terms of specificity and PPV, the HIFI model was better than aMAP score.