*Corresponding author: Feng Xia, Institute of Hepatobiliary Surgery and Southwest Cancer Center, Army Medical University, China
How to cite this article: Xuesong L, Wan Yee L, Deyu G, Shuguo Z, Geng C, et al. Hepatic Regeneration index and Completeness of Regeneration in Remnant Liver after Partial Hepatectomy for Hepatocellular Carcinoma and Cirrhosis. Open Access J Surg. 2021; 13(4): 555868 DOI: 10.19080/OAJS.2021.13.555868.
This study aimed to evaluate whether the regeneration index (RIx) and completeness of regeneration of remnant liver volume (RLV) are related to liver fibrotic stages, resected liver parenchymal volumes (RLPV) and long-term survival outcomes in patients with hepatocellular carcinoma (HCC) and cirrhosis.
Sixty-two HCC patients were included in regular CT volumetry assessments for up to 17 months after partial hepatectomy. The RIx of remnant liver was calculated by the formula at the different time points as: [regenerated RLV- original RLV] / original RLV ×100. The patients were divided into subgroups according to the fibrotic stages and the RLPV as a percentage of total liver volume (TLV). The mean RIx and completeness of regeneration at 17 months of RLV were compared between each of the subgroups.
The mean RIx declined as the fibrotic stages increased. It increased with increasing percentages of RLPV to TLV. RIx was high as measured by CT volumetry around 1 month after hepatectomy (t1) and then reached a plateau, especially in patients with severe liver fibrosis or with a minor hepatic volume resection. There were no significant differences in long-term disease-free survival and overall survival outcomes between the high RIx (≥50 percent) and low RIx (<50 percent) subgroups.
Liver regeneration had basically been completed (94.1%±14.1) within 4 months in cirrhotic livers after hepatectomy. The RIx showed no correlation with long-term disease-free and overall survival outcomes for HCC patients with cirrhosis.
According to the annual projections, the World Health Organization estimated that more than 1 million patients would die from hepatocellular carcinoma (HCC) in 2030. The main etiologies of HCC are hepatitis B or C virus infection and alcohol abuse. Partial hepatectomy (PH) is still the first-line treatment for HCC [1-3]. However, the rule governing regeneration of the remnant liver after hepatectomy for these patients remains largely unclear . The future Remnant Liver Volume (future RLV) and the restoration of the Remnant Liver Volume (RLV) by liver regeneration have been accurately measured and monitored by
CT volumetry before and after surgery [5-7]. Many researchers
reported that the remnant liver regenerates immediately after
PH in non-cirrhotic livers. The early Regeneration Index (RIx)
on day 7 ranges from 20% to 60%, and the RLV reaches 85%
of its initial size at 6 months after right hepatectomy. The most
relevant element related to RIx is the size of the original RLV [8-
10]. Few studies have been conducted on RIx and restoration
of RLV by regeneration after PH in cirrhotic liver. These studies
just revealed that the regenerative capacity of cirrhotic liver is
lower and is more complicated than non-cirrhotic livers. Little is
known on the impact of fibrotic stages of the liver and resected
liver parenchymal volumes on liver regeneration of the remnant
liver after hepatectomy. The study aimed to clarify the RIx and the
completeness of regeneration of liver volume in the remnant liver
in patients with cirrhotic liver after hepatectomy. The patients
were divided into subgroups based on the fibrotic stages and
percentages of Resected Liver Parenchymal Volume (RLPV). The
regeneration of the remnant liver after surgery was assessed by
CT volumetry for more than one year. The relationship between
liver regeneration and long-term survival outcomes were also
Consecutive patients who underwent partial hepatectomy for
Hepatocellular Carcinoma (HCC) from October 2010 to April 2014
at the Department of Hepatobiliary Surgery, Southwest Hospital
were retrospectively studied. The demographic characteristics
and operative findings of these patients are summarized in (Table
1). No patients received TACE or chemotherapy, and no portal vein
invasion was detected before surgery. The resected specimens
were all verified to be HCC on histopathological examination. The
study was approved by the Ethics Committee of the Southwest
Hospital of the Army Medical University, and it was censored on
April 30, 2016.
* Value expressed in median (interquartile range) and variance with range in parentheses.
All patients underwent a pre-operatively and four postoperative
CT volumetry assessments. The mean interval between
the preoperative CT volumetry and operation was 7 days. The
mean time of the first postoperative CT volumetry was 58.3±39.2
days (range 5 to 132 days), the second CT volumetry 188.9±52.7
days (range 77 to 275 days), the third CT volumetry 318.7±61.8
days (range 203 to 436 days), and the fourth CT volumetry
506.1±139.4 days (range 326 to 894 days). These time points
were defined as t1, t2, t3, t4. The CT scanning included the entire
upper abdomen, and all acquisitions started from the top of the
liver in a craniocaudal direction to cover the whole liver during a
single breath holding in maximal inspiration. The pre-contrast and
then the hepatic arterial, portal venous, and delayed phases were
obtained. Intravenous contrast enhanced imaging was obtained
using a nonionic, iodinated contrast material (370 mg I/mL
iopromide, Ultravist, Bayer Shering Pharma, Berlin, Germany) at a
dose of 1.8 mL/kg of body weight. This contrast was administered
with a power injector at a flow rate of 5 mL/second through an
18-gauge intravenous catheter placed in an antecubital vein. The
imaging slice thickness was 5 mm.
Anatomical liver resections were based on the Couinaud’s liver
segments. Non-anatomical hepatectomy followed the resection
planes as planned in a preoperative simulator. The patients were
divided into 3 subgroups according to the percentages of the
total non-tumorous liver volume resected. The R1 group (n=22),
0-19%. The R2 group (n=32), 20-39%. The R3 group (n=8), ≥40%.
All the surgically resected specimens were studied
histopathologically. Using the metavir scoring system to assess
the liver fibrosis stage by an experienced pathologist based
on predefined criteria . The patients were classified into 4
subgroups: the F1 subgroup with stellate enlargement of portal
tracts but without septa formation (n=2); the F2 subgroup with
enlargement of portal tracts with rare septa formation (n=9);
the F3 subgroup with numerous septa formation (n=22); the F4
subgroup with cirrhosis (n=29). No patients in this study were in
the F0 stage of the metavir score.
The liver parenchymal and tumorous contours of each of the
histopathological slices were delineated manually. The volumes were calculated by summation of the areas on each axial section. Major vessels (e.g., inferior vena cava, extrahepatic portal vein)
and fissures (e.g., fissure for ligamentum teres) were excluded
from the volume summation. Volumetric data was calculated
manually by the software of the manufacturer workstation (Amira
4.1.2) to include the Total Liver Volume (TLV), Tumor Volume
(TV), resected liver parenchymal volume (RLPV) before surgery,
and Remnant Liver Volume (RLV) at each CT volumetric study
time points. The original remnant liver volume (original RLV)
was calculated by [TLV (without TV)-RLPV (without TV)] before
surgery. RLV was calculated at each CT volumetric study time
points. The remnant liver RIx was calculated as: Remnant liver RIx
(%) = [(regenerated RLV at each follow-up time point – original
RLV)/original RLV] ×100.
All statistical analyses were performed by the SAS9.3 software.
Quantitative data with normal distribution were expressed as
median (interquartile range). Comparisons between groups of
quantitative data were performed with the Kruskal-Walli’s test
or the Wilcoxon test. A P value of less than 0.05 was considered
as statistically significant. The association between patients’
parameters and regeneration of original RLV was analyzed by the
linear regression analysis. The relationship between the liver RIx
and tumor recurrence were assessed by the logrank test.
For the 62 patients in this study, the types of liver resection
included right hepatectomy (n=4), central hepatectomy (n=3), left
hepatectomy (n=6), extended left hepatectomy (n=4), and minor
liver resection (including segmentectomy and non-anatomical
resection (n=45). There were no patients who developed liver
failure. The 90-days postoperative mortality rate was 0%. The
median of TLV (volume tumor excluded) was 1064.7 ml, the
median TV was 167.5 ml, and median original RLV was 786.5
ml. The patient’s clinicopathologic factors, conditions during
operation and operation methods are summarized in (Table 1). The
median of RIx at CT volumetric assessments at the postoperative
time points t1, t2, t3 and t4 were 25.8(0.5-91.2), 27.0(1.0-107.0),
26.5(2.2-139.5), and 33.9(0.3-127.2) respectively.
The liver volumes and RIx at different fibrotic stages are
summarized in (Table 2). Among the F1 to F4 subgroups, the
pre-operative TLV and original RLV showed no significant
differences. At t1, the median RLV for the F1 to F4 subgroups increased to 1086.2(1053.8-1118.6) ml, 1325.0(867.3-1732.3) ml, 1006.7(652.4-1534.4) ml and 1083.4(600.0-1606.2) ml
respectively. The corresponding RIx reached to 38.4(0.8-76.0),
28.6(2.5-72.9), 30.0(5.0-91.2) and 24.2(0.9-84.3) percent
respectively. Then the RIx increased further only in the subgroup
F1 from t2 to t4, but it slowed down in the subgroups of F2 to
F4. The median RIx of each subgroup increased fastest from liver
resection to t1. Using the metavir system, the median RIx were
compared among the subgroups at a single follow-up time point.
The median RIx was 80.1(20.6-139.5) in the F1 subgroup, and
23.7(2.2-116.7) in the F4 subgroup at t3, although there were no
significant differences among the subgroups, but it seemed that
there is a downward trend for RIx accompanied by fibrotic stage
increased. The mean RIx in relation to the different metavir scores
at each follow-up time points are shown as a box plot in (Figure
1A). The restoration of volume in the remnant livers after surgery
in each subgroup is shown in (Figure 2A). The median RLV were
restored to 93.7±5.7, 97.8±17.2, 97.7±14.1 and 90.2±13.0 percent
of the TLV at t1 in the subgroups F1 to F4 respectively, then the
increase of median RLV all reached a plateau from t2 to t4.
Note: F1 subgroup (n=2), Stellate enlargement of portal tract but without septa formation; F2 subgroup (n=9), Enlargement of portal tract with rare
septa formation; F3 subgroup (n=22), Numerous septa formation; F4 subgroup (n=29), Cirrhosis. The data are presented as median (interquartile range).
# Mark located in the right side of RIx in F1 subgroup means no significant difference between F1 and F4 subgroupsat t1 (P=0.936), and t2 (P= 0.629), and at t3 (P=0.376), and t4 (P= 0.469).
# Mark located in the right side of RIx in F2 subgroup means no significant difference between F2 and F4 subgroups at t1 (P= 0.810), and t2 (P=0.877), and t3 (P= 0.668), and t4 (P= 0.295).
Note: all patients were classified into 3 subgroups according to percentage of removal of total non-tumorous liver volume, R1 group (n=22), 0-19
percent; R2 group (n=332), 20-39 percent (n=7); R3 group(n=8), more than and equal to 40 percent. Comparisons between groups of quantitative data were performed with Kruskal-Walli’s test and Wilcoxon test, a P value of less than .05 was considered significant.
*Mark located in the right side of RIx in R1 subgroup means significant difference between R1 and R3 subgroups at t1 (P= 0.00035), and t2 (P= 0.0006035), and t3 (P= 0.00003), and t4 (P= 0.00032).
*Mark located in the right side of RIx in R2 subgroup means significant difference between R2 and R3 subgroups at t2 (P= 0.01070.006), and t3 (P= 0.00470.010), and t4 (P= 0.0042008).
#Mark located in the right side of RIx in R2 subgroup means no significant difference between R2 and R3 subgroups at t1 (P= 0.2858054).
The patients were classified into three subgroups of R1, R2 and
R3 according to the percentages of resected liver volume. There
were no significant differences in age or sex distribution among
them. The increase in RLV and RIx in relation to the percentages
of liver resection are summarized in (Table 3). The mean preoperative
TLV were 1069.5(600.0-1624.6) ml, 1058.1(652.4-
1624.6) ml, 1068.6(863.0-1606.2) ml, respectively. The mean TV
were 117.8(11.0-1349.1) ml, 190.6(10.1-1989.3) ml, 186.9(16.7- 990.3) ml, respectively with no significant differences among the groups. The original RLV were 889.6(537.2-1442.8) ml,
767.4(488.5-1248.4) ml, and 535.6(428.4-888.5) ml, respectively.
For the subgroups R1, R2 and R3 at t1, the RIx reached 13.0(0.5-
54.7), 27.2(0.9-91.2), 43.9(21.6-84.3) percent, respectively.
The RIx only increased in R3 from t2 to t4, but it was less in R1
compared to R2. The median RIx of each subgroup all increased
fastest within t1. The median RIx were compared among the
subgroups at a single follow-up time point (Table 3). Shows that
the RIx reached to 43.9(21.6-84.3), 67.8(22.9-107.0), 67.3(13.8-
139.5), 85.5(22.0-127.2) percent in the R3 subgroup from t1 to t4.
However, the RIx only reached to 13.0(0.5-54.7), 22.6(1.0-50.8),
21.5(2.2-69.9), 22.1(0.3-96.6) percent in the R1 subgroup from
t1 to t4. There were significant differences between the 2 groups
(p<0.05). The same results were also found between the R2 and
R3 subgroups. Thus, the more the liver volume was resected, the
higher was the RIx acquired by the remnant liver from t1 to t4
. The median RIx in relation to the different RLPV at each
follow-up time point are shown as a box plot in (Figure 1B). When
completeness of regeneration of RLV was studied, the median RLV
of the R1 subgroup reached to its initial volume before surgery.
However, the median RLV of the R3 subgroup was only restored
to 91 percent of its original liver volume at t4. (Figure 2B) shows
the median RLV as a percentage of TLV at different follow-up time
The analysis included the following factors which might
impact on liver regeneration for the patients: BMI, original RLV,
pre-surgical platelet count, prothrombin time (PT), tumor volume,
serum total bilirubin, and follow-up time. Partial regression
coefficients evidenced with stepwise regression revealed a
significant association between the mean RIx and original RLV (p
The recurrence rate for the patients was 75.8 percent (47/62),
and the death rate was 38.7 percent (24/62). The patients were
then divided into two subgroups by the percentages of RIx into the
two subgroups: subgroup 1 (<50 percent); and subgroup 2, (≥50
percent). (Figure 3) shows that there was no significant difference
between the two groups in median DFS and OS outcomes.
Previous research reported that the RIx of non-cirrhotic liver
after surgery rapidly reached to 28-64 percent in the 1st month,
indicating that the early postoperative period was crucial for liver
regeneration, and the RLV increased to nearly its original size at 6
month (13,14). However, the RIx in post-hepatectomy for cirrhotic
liver is still unclear. Most patients in our study had a metavir score
of F3 (n=22) or F4 (n=29). As our results showed that the mean
RIx for all the patients was 25.8(0.5-91.2) percent at t1, then the
increase in the RIx became slower from t2 to t4, liver regeneration
occurs even in cirrhotic livers in the early postoperative period. In
our study the RLV almost reached to its original volume within four
months after hepatectomy, the major liver regeneration occurred
within t1. The fibrotic stages and the resected liver parenchymal
volume (RLPV) had remarkable effects on the completeness of
regeneration of the liver remnant and the RIx in cirrhotic patients.
These observations can be explained by the defective adaptation
to surgical stress, the decreased regenerative capacity of the
remnant liver parenchymal cells , the decreased portal blood
flow to the remnant liver than in healthy people, and the excessive
expression of IL-6 and CPR which are harmful to patient’s
recovery and liver regeneration . What are the mechanisms
for cirrhosis to affect liver regeneration? In our study, the RIx was
obviously higher in F1 than in F4 at all the different follow-up time
points (Table 2), especially at t3 (80.1 vs 23.7 percent). Although
the RIx did not show a significant difference between them, but
the median RIx in F4 was obviously lower than RIx in the healthy
liver previously reported . Thus, the regeneration capacity was
inhibited with increasing fibrotic stages. In addition, the RIx only
reached to 69.6(12.0-127.2) percent at t4 in the F1 subgroup,
the decrease in the RIx and RLV may be explained by tumor
recurrence in a limited number of patients in the F1 subgroup.
Kele and his associates reported that the RIx was linearly related
to the resected liver parenchymal volume (RLPV) at 6 months
after operation in non-cirrhotic livers8. In our study, the RIx was
also strongly correlated with the RLPV, showing an increasing
tendency at each follow-up time point when the liver remnant
was regenerating. Besides, the significant difference among the
subgroups suggested that the RIx had a positive correlation with
A previous study showed that some pre- and post-operative
factors had the potential to predict liver regeneration . To
detect the factors which might play a role on the RIx in patients
with cirrhosis, our multiple linear regression suggested that only
BMI and original RLV were significant factors. The follow-up time
did not influence RIx. Our results revealed a negative relationship
between original RLV and RIx (data not shown), which indicated
that the smaller the RLV the more the release of proinflammatory
cytokines which promote liver growth. Thus, the rate of TLV
is more important to liver regeneration than the type of liver
resection. A high tumor recurrence rate exists in patients after
partial hepatectomy for HCC. The recurrence and death rates in
our study were 73.9% and 52.2%, respectively. Tumor inoculation
in animal models showed that tumor grows faster after PH and
tumor cells can be stimulated by factors associated with liver
regeneration [19,20]. Based on the RIx in our study, all patients
were categorized into two subgroups by fifty percentage of RIx.
The long-term DFS and OS outcomes showed no significant
difference between the two subgroups. Our results indicated
that liver regeneration might not directly promote liver tumor
recurrence in patients with cirrhosis. There are limitations of
this study. First, the number of patients in our study was small.
Second, there is a potential of patient selection bias. Third,
increase in liver volume by CT volumetry may not entirely reflect
the increase in volume were all due to functional regeneration of
liver parenchymal cells, but rather the increase could also be due
to oedema or congestion of the remnant liver.
Liver regeneration in cirrhotic patients after partial
hepatectomy was almost complete within 2 months after partial
hepatectomy. The RIx and RLV increased very little after 4
months post-operation. The RIx declined with increasing fibrotic
stages. The RIx was positively correlated with the resected liver
parenchymal volume (RLPV). The completeness of regeneration
of liver remnant volume was impaired with increasing fibrotic
stages and RLPV, especially in the F4 and R3 subgroups when the
remnant liver volume failed often to reach their original volumes
before surgery. The RIx had no direct impact on long-term survival
outcomes in our study.
Authors’ contributions: Xuesong Li contributed to the
conception/design of the research. All authors contributed to
acquisition, analysis, or interpretation of data as well as drafting
the article. All authors agree to be fully accountable for ensuring
the integrity and accuracy of the work and approved the final