Progenitor cell markers predict outcome of patients with hepatocellular carcinoma beyond Milan criteria undergoing liver transplantation

Abstract

Background & aims: In patients with hepatocellular carcinoma (HCC), liver transplantation (LT) is an excellent therapy if tumor characteristics are within the Milan criteria. We aimed to define genomic features enabling to identify HCC patients beyond Milan criteria who have acceptable transplant outcomes.

Methods: Among 770 consecutive HCC patients transplanted between 1990 and 2013, 132 had tumors exceeding Milan criteria on pathology and were enrolled in the study; 44% of the patients satisfied the 'up-to-7 rule' [7=sum of the size of the largest tumor and the number of tumors]. Explant tumors were assessed for genomic signatures and immunohistochemical markers associated with poor outcome.

Results: At a median follow-up of 88months, 64 patients had died and 45 recurred; the 5-year overall survival (OS) and recurrence rates were 57% and 35%, respectively. Cytokeratin 19 (CK19) gene signature was independently associated with recurrence [Hazard ratio (HR)=2.95, p<0.001], along with tumor size (HR=3.37, p=0.023) and presence of satellites (HR=2.98, p=0.001). S2 subclass signature was independently associated with poor OS (HR=3.18, p=0.001), along with tumor size (HR=5.06, p<0.001) and up-to-7 rule (HR=2.50, p=0.002). Using the presence of progenitor cell markers (either CK19 or S2 signatures) patients were classified into poor prognosis (n=58; 5-year recurrence 53%, survival 45%) and good prognosis (n=74; 5-year recurrence 19%, survival 67%) (HR=3.16, p<0.001 for recurrence, and HR=1.72, p=0.04 for OS).

Conclusions: HCC patients transplanted beyond Milan criteria without gene signatures of progenitor markers (CK19 and S2) achieved survival rates similar as those within Milan criteria. Once prospectively validated, these markers may support a limited expansion of LT indications.

Keywords: Gene expression; Gene signature; Prognosis; Stem cell; Survival.

Fig. 1.. Flow chart of the study.

The initial cohort included 770 patients that were consecutively transplanted for HCC at Mount Sinai Hospital, New York (n = 590) and Mayo Clinic, Scottdale Arizona (n = 180). A total of 206 patients were beyond Milan criteria based on the explant pathology. Tumors with HCC-ICC features, diffused pattern, necrotic tissue, macro-vascular invasion or metastasis were excluded. From the remaining 141 cases, six tumors with poor quality RNA were discarded. Total RNA from 135 tumors was subjected to transcriptome profiling. Three samples had poor quality profile and were excluded and eventually 132 tumors were tested for the presence of previously reported outcome-associated gene signatures. OLT, orthotopic liver transplantation; ICC, intrahepatic cholangiocarcinoma; FFPE, Formalin Fixed Paraffin Embedded; DASL HT, cDNA-mediated Annealing, Selection, Extension, and Ligation (DASL) High Throughput (HT) Assay.

Fig. 2.. Prediction of gene expression signatures.

Only signatures that were able to assign patients into good and poor prognosis groups with FDR >0.05 were included. (A) Each column represents different sample and each row different signature. Positivity of each signature is represented by red bars. Events (recurrence or 5-year death) are shown with black bars. (B) Visualization of cramer’s V coefficient for the pair-wise comparison of gene signatures. The scale from blue to red represents the strength of correlation (red represents the highest correlation). The signatures are clustered according to their correlation.

Fig. 3.. Kaplan-Meier curves and estimates of overall survival and recurrence in patients beyond Milan undergoing liver transplantation (n = 132).

(A and B) Overall survival and recurrence in this cohort. Patients positive for either CK19 or S2 signatures (progenitor cell signatures) were assigned into a poor outcome group. (C and D) Overall survival and recurrence in patients beyond Milan with or without CK19/S2 signatures. p values are calculated based on the log-rank test.