Page 21 - Read Online

Page 10 of 12 Nishioka et al. Hepatoma Res 2018;4:1 I http://dx.doi.org/10.20517/2394-5079.2017.46

Several limitations of the present study should be recognized. First, the tumor samples analyzed in this study
were collected from patients recruited from a single medical center in the state of Hawaii. This raises the
possibility of selection bias. However, the likelihood of such bias is reduced given the prospective nature
of this study and the fact that our center is responsible for treating most of the HCC patients in Hawaii.
Because Hawaii is a small territory, the number of patients presenting annually with HCC is also relatively
small despite the high incidence, and thus the statistical power of this study is limited. However, unlike
unsupervised methods, GSEA has been found to be fairly robust with sample size in the range of the present
[35]
study . Another potential limitation relates to the fact that gene expression analysis was performed by
sampling only a small peripheral portion of the tumor. Because of this, our results cannot account for the
possibility of intra-tumoral heterogeneity. Notwithstanding this methodologic limitation, the results of this
study did produce a biologically-coherent depiction of HCC tumors associated with high serum AFP levels.

Our study provides additional data supporting the clinical relevance of gene signatures for HCC derived
from many different cohorts, including those from Asia, Europe, and North America. Because there are
[36]
studies suggesting that ancestry and genetics may influence HCC genomes , it is prudent to validate
predictive gene signatures for HCC in a broad spectrum of patients before accepting them into mainstream
[10]
application. While some gene signatures for HCC have already been subject to further validation , none
have been thoroughly validated to a global extent. Our study, conducted in a racially and ethnically diverse
HCC cohort, provides further evidence to support the generalizability of gene signatures for clinical
molecular classification of HCC. Specifically, we confirmed that several externally derived molecular sub-
classes of HCC associated with distinct molecular features and survival outcomes could be detected with
statistical confidence in a cohort of patients from Hawaii. The generalization of these signatures will support
their use in multi-center efforts aimed at developing targeted therapies for HCC.

In conclusion, herein we provide supporting evidence that a molecular classification system for HCC
developed using cohorts from North America, Europe, and Asia is applicable to patients in Hawaii. Similar
to other cohorts, the findings in the present study also indicate that elevated AFP is significantly associated
with more aggressive tumor characteristics and poor clinical outcome, as well as gene expression related
to cell cycle progression, DNA damage response, and MYC oncogene pathways. Confirming the ability
to apply the same molecular classification system to tumors from different populations is a crucial step to
broadening the use of genomic enrichment strategies in global multi-center clinical trials. Establishing that
similar distributions of tumor sub-classes exist in different populations will also increase confidence that
molecularly-targeted therapies found to be beneficial in one cohort can be similarly effective in cohorts from
other populations.

DECLARATIONS
Authors’ contributions
Conception: Wong LL, Kwee SA
Study design: Nishioka ST, Wong LL, Kwee SA
Data collection: Sato MM, Wong LL, Tiirikainen M, Kwee SA
Data analysis: Nishioka ST, Sato MM, Tiirikainen M, Kwee SA
Interpretation of results and manuscript writing: Nishioka ST, Sato MM, Wong LL, Tiirikainen M, Kwee SA

Data source and availability
Corresponding author may be contacted for any data inquiries.

Financial support and sponsorship
This work was supported by U.S. National Institutes of Health grants R01CA161209-06 and U54MD007584-07.