A novel risk stratification model based on the Children's Hepatic Tumours International Collaboration-Hepatoblastoma Stratification and deoxyribonucleic acid methylation analysis for hepatoblastoma.

ABSTRACT

INTRODUCTION: Hepatoblastoma (HB) is the most common paediatric liver tumour, and epigenetic aberrations may be important in HB development. Recently, the Children's Hepatic Tumors International Collaboration-Hepatoblastoma Stratification (CHIC-HS) developed risk stratification based on clinicopathological factors. This study aimed to construct a more accurate model by integrating CHIC-HS with molecular factors based on DNA methylation. METHODS: HB tumour specimens (N = 132) from patients treated with the Japanese Pediatric Liver Tumors Group-2 protocol were collected and subjected to methylation analysis by bisulfite pyrosequencing. Associations between methylation status and clinicopathological factors, overall survival (OS), and event-free survival (EFS) were retrospectively analysed. We investigated the effectiveness of the evaluation of methylation status in each CHIC-HS risk group and generated a new risk stratification model. RESULTS: Most specimens (82%) were from post-chemotherapy tissue. Hypermethylation in ≥2 of the four genes (RASSF1A, PARP6, OCIAD2, and MST1R) was significantly associated with poorer OS and EFS. Multivariate analysis indicated that ≥2 methylated genes was an independent prognostic factor (hazard ratios of 6.014 and 3.684 for OS and EFS, respectively). Two or more methylated genes was also associated with poorer OS in the CHIC-very low (VL)-/low (L)-risk and CHIC-intermediate (I) risk groups (3-year OS rates were 83% vs. 98% and 50% vs. 95%, respectively). The 3-year OS rates of the VL/L, I, and high-risk groups in the new stratification model were 98%, 90%, and 62% (vs. CHIC-HS [96%, 82%, and 65%, respectively]), optimising CHIC-HS. CONCLUSIONS: Our proposed stratification system considers individual risk in HB and may improve patient clinical management.