Page 190 - Read Online

Page 8 of 11 Costa et al. Hepatoma Res 2018;4:35 I http://dx.doi.org/10.20517/2394-5079.2018.06

were more homogeneous. In addition, sorafenib only acted on the DEN model, decreasing tumor growth and
perfusion . Similar findings were demonstrated in another study that observed a low objective response
[12]
(3%) by RECIST criteria . This cytotoxic response pattern was also observed in another clinical study, in
[37]
which approximately 42% of patients were stable or had a minor/partial response .
[38]
In the current study, we found the presence of advanced fibrosis or cirrhosis in all animals, as well as
advanced HCC with vascular invasion in 75% of control animals (n = 4) and 43% of animals in the treated
group (n = 8). These findings highlight the clinical relevance of our model, as most preclinical studies used
younger animals with less advanced disease, thus not reflecting the hepatic microenvironment observed
in humans .
[36]
Despite treatment with sorafenib, the mortality rate (60%) was similar in both groups, resulting from the
severity of liver cirrhosis and advanced HCC; sometimes with decompensation in ascites (about 10% in both
groups) and pulmonary metastasis. In humans, the prognosis of advanced HCC is bleak, with a median
survival of 6 months or 25% in 1 year . Park et al. used PET/CT with [ F]FDG to evaluate rats exposed
[15]
[24]
18
only to intraperitoneal DEN, administered once a week for 16 weeks. They reported a mortality rate similar
to our study, or about 65% in week 19. A more precise evaluation of the effect of sorafenib on survival was
not possible because all animals were euthanized 3 days after the last PET scan to minimize risk of losing
more animals prior to the endpoint of the study.

The development of a biochemical marker or diagnostic tool to identify the most undifferentiated and
aggressive tumors has been studied in an attempt to better select patients for curative treatment [17,24] .
[ F]FDG PET appears to be a potential tool, because it has been shown that higher values of [ F]FDG uptake
18
18
correlates with lower overall survival, advanced HCC, undifferentiated histology (loss of p53 expression),
and increased liver transplantation recurrence [17,19,22,39] . Lee et al. retrospectively evaluated patients who
[23]
underwent liver transplantation for HCC and had [ F]FDG PET prior to surgery. They found that HCC
18
18
patients with lower [ F]FDG uptake had more highly differentiated HCC (Grades I/II Edmondson-Steiner
classification), with lower rates of microvascular invasion and no recurrence of HCC after a 3-year follow-
up : results that are similar to the results found in our work.
[23]
Kim et al. showed that the [ F]FDG uptake calculated as ratio between tumor and liver adjacent tissue was
[17]
18
more accurate than tumor uptake in predicting HCC post-transplant recurrence (SUVmax Tumor/SUVmax
18
Liver 0.869 vs. tumor SUVmax 0.762). In our study, the best correlation of [ F]FDG uptake and HCC Grades
at III/IV was found with SUVmax Tumor/SUVmax Muscle (R = 0.54, P = 0.006), followed by SUVmax
2
tumor (R = 0.44, P = 0.01) and Tumor SUVmax/Liver SUVmax ratio (R = 0.42, P = 0.02): somewhat lower
2
2
than what was reported by Kim et al. .
[17]
The absence of a pretreatment (16th week) [ F]FDG PET scan is one limitation of our study, as we could not
18
follow the evolution of the same node with one diagnostic tool throughout the experiment. This decision was
based on two reasons: (1) high probability of increasing mortality rate on additional anesthesia during image
acquisition (animals were already weak at that point) and (2) at pretreatment time point it was necessary
to ensure the presence of lesions that could be detected by PET (> 1 mm) at a later point in time; therefore,
a method with higher spatial resolution was chosen (US) to check homogeneity of lesion count and size
between groups. US was able to show that the groups were homogeneous and that comparison between
groups with PET at the end of the experiment was feasible. Although US is more sensitive in the detection of
small lesions and offers good localization, it is not able to provide detailed information about tumor grade.
Therefore we chose to use PET as a correlate for tumor aggression.

We showed that sorafenib could be responsible for reduced number of nodules and aggressiveness of HCC
(more Grade I/II lesions than III/IV), as well as reduced tumor [ F]FDG uptake. [ F]FDG PET could be used
18
18