A prognostic fingerprint in liver transplantation for hepatocellular carcinoma based on plasma metabolomics profiling.

INTRODUCTION

Tumor recurrence is a major cause of post-transplant mortality in liver transplantation for hepatocellular carcinoma (HCC). This study aimed to explore an effective noninvasive approach to accurately predict post-transplant tumor recurrence.

MATERIALS AND METHODS

Metabolomics profiling was performed on pre-operative plasma from 122 HCC patients undergoing liver transplantation, 52 healthy controls (HC) and 25 liver cirrhosis (LC) patients.

RESULTS

Five prognostic metabolites were identified by univariate analysis (P < 0.01), including phosphatidylcholine (PC) (16:0/P-18:1), PC(18:2/OH-16:0), PC(o-16:0/20:4), nutriacholic acid and 2-oxo-4-methylthiobutanoic acid. In the HCC group, PC(o-16:0/20:4), nutriacholic acid and 2-oxo-4-methylthiobutanoic acid were decreased, while PC(18:2/OH-16:0) was elevated compared with the LC group (e < 0.05). PC(16:0/P-18:1) was associated with tumor size, vascular invasion, and neutrophil-lymphocyte ratio (NLR; P < 0.05). Moreover, PC(18:2/OH-16:0) was also related to tumor number and NLR (P < 0.05). Multivariate cox regression showed that PC(16:0/P-18:1), PC(18:2/OH-16:0), nutriacholic acid and alpha-fetoprotein (AFP) were independent risk factors for tumor recurrence (P < 0.01). A prognostic fingerprint was established as a nomogram, which divided the patients into low risk (n = 45), moderate risk (n = 48) and highrisk groups (n = 29) with discriminated prognosis (P < 0.001). In patients fulfilling the Hangzhou criteria, the fingerprint/nomogram could also successfully stratify the patients into two groups with different recurrence risk (P < 0.05).

CONCLUSIONS

The established pre-operative plasma fingerprint/nomogram is efficient in the prediction of recurrence risk, which could facilitate candidate selection in liver transplantation for HCC.