PREDICTING HEPATOCELLULAR CARCINOMA (HCC) OUTCOMES IN HCV PATIENTS USING MOLECULAR MARKERS.Room: 210
Valeria Mas, Daniel Maluf, Kellie Archer, Kenneth Yanek, Bridgette Williams, Robert Fisher. Surgery and Biostatistics, Virginia Commonwealth University, Richmond, VA.
Background. Liver transplantation (LT) is a curative treatment for small HCC. Potentially curable higher HCC stage patients (Pts) are denied LT because the lack of progression (PRG) and recurrence (REC) markers. Molecular markers are being studied to better predict post-LT outcomes.
Patients and Methods. 38 HCV-HCC Pts candidates for LT were prospectively studied (26.421.7mo). Gene expression analysis (GExp) of tumor samples was performed using high-density oligonucleotide arrays. The robust-multiarray average method was used to estimate probe set (Psets) expression summaries. In assessing Pset significantly associated with overall survival, the time-dependent covariate transplant was included in the Cox proportional hazards model to adjust for differences in survival between LT vs. not-LT Pts. A second time-dependent survival analysis considering time to PRG or REC as a covariate in the Cox proportional hazards model to adjust for differences in survival among Pts due to disease PRG was used.
Results. Twenty Pts (52.6%) underwent LT, 13 (34.2%) progressed while waiting for LT, 4 (10.5%) still alive waiting LT, and 1 (2.6%) died without PRG while waiting for LT. Differences in GExp among Pts who underwent or did not progress while waiting LT (n=25) vs. those whose disease progressed while waiting for LT (n=13) were assessed. Fifty-four Psets were significantly differentially expressed (=0.001).
Among LT Pts, 46 Psets were differentially expressed comparing LT Pts with an explanted stage of T3/T4 who also recurred (n=5) vs. LT Pts with the same explanted stage who have not recurred (n=5). Genes involved in the host viral response (i.e., MX1, OAS2), immunoregulatory processes (i.e., STAT1, CCL27), and anti-apoptosis (i.e., FAIM3) were over-expressed in LT Pts with REC. Age, race, TNM stage, AFP levels, and MELD score were not significant predictors of survival in this group. However, 22 Psets were found to significantly predict survival. Survival genes were categorized as associated with inhibition of cell growth and inductors of apoptosis (i.e., TNFAIP3, FOXO3A). There were 36 Psets that were significantly associated with survival after adjusting for the time-dependent covariate PRG or REC (i.e., BTG2, PTPRD).
Conclusions. A set of genes was significantly associated with survival indicating that a limited set of markers may be sufficient to predict the prognosis of HCV-HCC Pts at the transplant evaluation time.
Keywords: Gene expression; Hepatocellular carcinoma; Liver transplantation
Wednesday, July 26, 2006 11:50 AM
Concurrent Session 65: Transplant Associated Malignancy: Clinical Investigations (10:30 AM-12:30 PM)