Page 781 - Read Online

Yoneoka et al. Hepatoma Res 2020;6:67 I http://dx.doi.org/10.20517/2394-5079.2020.57 Page 9 of 11

who previously used a similar grouping system based on mRECIST to demonstrate that an elevated NLR
is associated with tumor progression after intra-arterial therapy. Although other studies have shown that
an elevated NLR was associated with poor overall survival following TARE, none of these studies reported
specifically on tumor progression based on imaging [22,23] . These results taken together suggest that NLR may
be a valuable prognostic marker in TARE.

Notably, we found that an elevated pre-treatment PLR predicted non-response to TARE in univariate
analysis but was not a significant variable in our multivariate model. This suggests that the pre-treatment
NLR may be superior to pre-treatment PLR in predicting non-response to TARE. Nonetheless, this
result may be limited by our small sample size. To our knowledge, this was the first study to examine the
[24]
prognostic capabilities of PLR in TARE based specifically on tumor response to therapy. D’emic et al.
previously suggested in their study of 116 patients who received selective internal radiation therapy that
pre-treatment PLR > 78 was the most predictive serum marker associated with improved overall survival.
However, it is difficult to make definitive conclusions about NLR and PLR in HCC as their study also
included other cancer types and only 37 patients had HCC. Future studies are therefore needed to compare
the prognostic capabilities of PLR compared to NLR in TARE.

On TTP analysis, we found that pre-treatment NLR ≥ 2.83 and pre-treatment PLR ≥ 83 were both
associated with a higher proportion of tumor progression at 6 months post-TARE. The median TTP was
[3]
not yet reached in all groups. This is consistent with previous results published by Salem et al. , who found
that the median TTP for radioembolization was greater than 26 months. On the basis of these results,
both the pre-treatment NLR and pre-treatment PLR may have utility in predicting tumor progression at
6 months following TARE. Nonetheless, the prognostic value of NLR could have a distinct advantage over
PLR because pre-treatment NLR < 2.83 was also associated with response to treatment in our multivariate
logistic regression analysis. NLR may therefore have greater clinical utility than PLR as pre-treatment NLR
was predictive of both tumor progression and potential response to therapy in our cohort. In comparison,
pre-treatment PLR was only predictive of tumor progression in our TTP analysis.

[25]
The ALBI grade was a newer model proposed by Johnson et al. that offered better discriminatory
capabilities compared to the Child-Pugh class. While other studies reported that the ALBI grade was
predictive of survival following TARE, our multivariate model did not find the ALBI grade helpful in
predicting response to TARE [26,27] . Since the ALBI grade likely reflects underlying liver function, it may be
more suitable for determining longer-term overall survival following TARE, rather than predicting specific
tumor response to treatment.

The underlying mechanism behind NLR and PLR is not well understood. However, it is generally
recognized that inflammation plays a key role in the development of cancer [15,17] . Neutrophils can favor
a pro-mutagenic state with the abundant release of reactive oxygen species and proteases . In addition,
[28]
platelets may support a pro-tumor microenvironment with the release of angiogenic factors such as
[29]
vascular endothelial growth factor and basic fibroblastic growth factor . These observations, coupled with
the fact that lymphopenia has been associated with advanced disease in various tumors, may be reflected
[30]
in systemic inflammation-based markers such as NLR and PLR . Nonetheless, more research is needed to
better understand the basis of these two markers.

This study was limited in that this was a single center study with a small sample size. This study was also
retrospective, and the exact timings of imaging and laboratory studies were not collected consistently as
part of a study protocol. Missing data in some patients may have also contributed to our small sample size.
In addition, several patients may have had laboratory data collected for other medical issues unrelated to
TARE, which may have influenced NLR and PLR.