MicroRNA based Prediction of Posthepatectomy Liver Failure and Mortality Outperforms Established Markers of Preoperative Risk Assessment

Abstract

Background: Posthepatectomy liver failure (PHLF) continues to be the most significant factor-determining outcome after hepatic resection, accounting for nearly half of postoperative mortality. In this study, we evaluated whether a newly developed commercially available test measuring circulating microRNAs (miRs) could predict PHLF and compared it with other established liver function tests.

Patients and methods: A total of 329 patients undergoing liver resection were included and postoperative outcome was assessed. Our previously described P-score, calculated on the basis of three circulating microRNAs (miR-122-5p, miR-192-5p, miR-151a-5p) using the hepatomiR^® CE-IVD test, was evaluated and compared with other predictors of PHLF, namely indocyanine green (ICG)-clearance as well as the combined aspartate aminotransferase (AST)-to-platelet ratio index (APRI) and albumin-bilirubin grade (ALBI) score.

Results: Compared with both other liver function tests, P-scores were superior in predicting PHLF and PHLF grades B and C (PHLF B + C) (PHLF B + C: hepatomiR^® AUC = 0.835, APRI + ALBI AUC = 0.807; retention rate at 15 min (R15) AUC = 0.690; plasma disappearance rate (PDR) AUC = 0.691). We also documented a superior positive (77%) and negative predictive value (> 90%) for PHLF, along with a close association with postoperative overall survival. A health-economic analysis demonstrated the cost-effectiveness of hepatomiR^® in terms of life-years gained due to improved patient risk stratification.

Conclusions: The hepatomiR^® P-score outperforms established liver function tests utilized in daily clinical practice for predicting PHLF and identifies patients possibly better served with alternative treatments. A health-economic assessment allowed us to demonstrate that optimized preoperative risk-assessment leads to a cost-effective improvement in patient outcomes.

Fig. 1

HepatomiR^® in the whole cohort and in comparison with other liver function tests: differences of P-scores among patients without and with PHLF, ROC curves comparing the predictive potential of hepatomiR^® for PHLF grades A–C and PHLF grade B + C (A); differences of APRI + ALBI and ICG-clearance (PDR and R15) in patients without and with PHLF (B); ROC curves comparing the predictive potential of hepatomiR^®, APRI + ALBI, and ICG-clearance (C); APRI+ALBI aspartate aminotransferase-to-platelet ratio index and albumin-bilirubin grade, AUC area under the curve, ICG indocyanine green, PHLF posthepatectomy liver failure, PHLF B+C PHLF grades B and C, ROC receiver operating characteristics; * p < 0.05, ** p < 0.01, *** p < 0.0001

Fig. 2

Risk groups of hepatomiR^® and other liver function tests compared: crosstabs illustrating the event of PHLF across different risk groups of hepatomiR^® and ICG-clearance (A)* and APRI + ALBI (B) cut-offs; * total number of patients who received ICG-clearance testing is included; N number, APRI+ALBI aspartate aminotransferase-to-platelet ratio index and albumin-bilirubin grade, PHLF posthepatectomy liver failure

Fig. 3

HepatomiR^® in tumor subgroups and overall survival: differences in hepatomiR^® without and with PHLF in the respective tumor subgroups, ROC curve analyses of different tumor subgroups (A), other* benign liver disease, and other malignancies; bar graphs depicting differences between low- and high-risk groups for the outcome parameters PHLF and overall survival (OS) up to 5 years across the low-risk cut-off P > 0.59 and the high-risk cut-off P > 0.68, the incidence of events in % is given on the y-axis (B); AUC area under the curve, CCa cholangiocellular carcinoma, HCC hepatocellular carcinoma, mCRC metastasized colorectal cancer, OS overall survival, PHLF posthepatectomy liver failure, 90 DM 90-day mortality; * p < 0.05, ** p < 0.01, *** p < 0.0001

Fig. 4

Overall survival differences between hepatomiR^® risk groups: Kaplan–Meier survival plots illustrating the difference in overall survival (OS) among both cut-offs, P > 0.59 and P > 0.68 (A), in the tumor subgroup HCC (B) and mCRC (C); the cumulative survival is plotted on the y-axis, the survival in months is plotted on the x-axis; a log-rank test was utilized to assess statistical differences; HCC hepatocellular carcinoma, mCRC metastasized colorectal cancer, No. number