The prognostic impact of perioperative dynamic changes in cachexia index in patients with hepatocellular carcinoma

Abstract Background The cachexia index (CXI), which consists of skeletal muscle, inflammation, and nutritional status, has been associated with prognosis in patients with hepatocellular carcinoma (HCC). We hypothesized that dynamic changes in CXI might be associated with long‐term outcomes in HCC. Methods This study comprised 131 patients who had undergone primary hepatic resection for HCC between 2008 and 2019. Preoperative CXI (pre‐CXI) and postoperative CXI (post‐CXI) were calculated by the following formula: skeletal muscle index x serum albumin level / neutrophil‐to‐lymphocyte ratio. Pre‐ and post‐CXI were classified into two groups (high vs. low). We retrospectively investigated the association of perioperative dynamic changes in CXI with disease‐free and overall survival. Results In multivariate analyses, negative HBs‐antigen (p = 0.02), high serum PIVKA‐II level (p < 0.01), poor tumor differentiation (p = 0.02), multiple tumors (p < 0.01), microvascular invasion (p < 0.01), partial resection (p < 0.01), postoperative complications (p < 0.01), and low‐pre‐CXI (p < 0.01) were significant predictors of disease‐free survival, while high ICGR15 (p = 0.01), poor tumor differentiation (p < 0.01), multiple tumors (p = 0.01), postoperative complications (p < 0.01), low‐pre‐CXI (p < 0.01), and low‐post‐CXI (p < 0.01) were significant predictors of overall survival. Low‐post‐CXI was associated with older age (p = 0.045), larger tumor (p < 0.01), longer operation time (p = 0.047), greater intraoperative bleeding (p < 0.01), and intraoperative blood transfusion (p < 0.01). Moreover, dynamic changes in CXI were associated with overall survival in each subgroup of patients with low‐pre‐CXI (p = 0.02) or high‐pre‐CXI (p = 0.03). Conclusions Not only post‐CXI but also dynamic changes in CXI from pre‐ to post‐hepatectomy can be a prognostic indicator of HCC, providing a compelling rationale for aggressive perioperative nutritional and physical interventions to improve long‐term outcomes.


| INTRODUC TI ON
Hepatocellular carcinoma (HCC) is the seventh most common cancer and the second leading cause of cancer mortality worldwide, accounting for 75%-85% of primary liver cancers. 1 The general curative treatment for HCC is surgical resection, but the recurrence rate after hepatic resection remains high, 2 suggesting the need to improve the therapeutic strategy for HCC.The risk stratification is important to develop more personalized therapy, which may improve the prognosis of HCC after resection.It has been reported that pre-and postoperative nutritional and inflammatory indexes reflect the prognosis after surgery for HCC. 3 Cachexia is a complex wasting syndrome seen in patients with cancer and many chronic diseases. 47][8] Although each component that contributes to forming cancer cachexia has been associated with prognosis, CXI, a comprehensive biomarker of cancer cachexia, has been more strongly associated with poor oncological outcomes in patients with HCC. 9 Given major surgeries including hepatectomy can develop postoperative malnutrition, inflammation, and loss of skeletal muscle mass due to their invasiveness, 10,11 the degree of cachexia might change from before to after surgery.However, the prognostic association of postoperative CXI or dynamic changes in CXI with survival in patients with cancer including HCC has not been investigated.Therefore, in this study, we retrospectively investigated whether postoperative CXI and dynamic changes in CXI from pre-to post-hepatectomy reflect the prognosis of patients after primary hepatic resection for HCC.

| Patient selection
This retrospective study analyzed 131 patients with HCC who underwent primary hepatic resection at the Department of Surgery, Jikei University Hospital, Tokyo, Japan, between January 2008 and December 2019.We excluded patients with other malignancies, perioperative mortality, and unavailable data on preoperative and postoperative CXI.Clinical information, surgical and pathological findings, and postoperative clinical courses were collected from medical records and a prospectively maintained database of patients at our institution.This study was approved by the Ethics Committee of the Jikei University School of Medicine (#27-177).

| Treatment and follow-up
Generally, the extent of hepatic resection was determined based on the retention rate of indocyanine green (ICG) at 15 min (ICG R15 ) and the hepatic reserve, as described by Miyagawa et al. 12 The nomenclature of the liver segments and types of operation followed the Tokyo 2020 terminology. 13The type of resection was either anatomical resection (hemihepatectomy, sectionectomy, or segmentectomy) or nonanatomical limited partial resection. 14The Tumor-Nodes-Metastasis (TNM) classification was based on the tumor pathology and the General Rules for the Clinical and Pathological Study of Primary Liver Cancer by the Liver Cancer Study Group of Japan. 15e recurrence of HCC was defined as newly detected hypervascular hepatic or extrahepatic tumors by ultrasonography, computed tomography, magnetic resonance imaging, or angiography with or without an increase in the serum alpha-fetoprotein (AFP) levels or levels of protein induced by vitamin K absence or antagonist-II (PIVKA-II).For recurrent HCC in the liver, repeated hepatic resection, local ablation therapy, or transcatheter arterial chemoembolization (TACE) was performed based on the hepatic functional reserve, judged mainly by ICG R15 .Extrahepatic recurrences were mainly treated with molecular-targeted drugs.Postoperative complications were classified according to the Clavien-Dindo classification. 16

| Assessment of the cachexia index (CXI)
The skeletal muscle index (SMI) was calculated using the right psoas muscle based on a previously described method. 9We simply measured the lengths of the major and minor axes of the right psoas muscle at the caudal end of the third lumbar vertebra, and calculated the psoas muscle mass area (PMA).SMI was calculated as PMA / the height squared (cm 2 /m 2 ).We assessed preoperative PMA by computed tomography (CT) taken last before surgery and postoperative PMA by CT taken closest to 90 days between 45 and 135 days after surgery as the first postoperative surveillance is performed at 3 months after surgery.The height was measured preoperatively.As previously described, 17 the neutrophil-to-lymphocyte ratio (NLR) was calculated by dividing the absolute neutrophil count by the absolute lymphocyte count.We assessed preoperative NLR by the complete blood count within 7 days before surgery and postoperative NLR by the complete blood count performed on the same day as the CT scan date used to calculate the postoperative psoas muscle.
[8][9] Given skeletal muscle mass varies by sex, we classified preoperative CXI into two groups (high vs. low) based on the receiver operating characteristic (ROC) curves in strata of sex.Preoperative CXI cut-off level for each sex was determined by maximizing the Youden index (the sum of sensitivity and specificity) for predicting 5-year survival on each ROC curve.The cut-off value for postoperative CXI was set to the same value as for preoperative CXI.

| Definition of sarcopenia
We assessed sarcopenia using preoperative PMA.As previously described, 18 preoperative PMA below the median for each sex was defined as sarcopenia.The sex-specific cut-off values for preoperative PMA were 35.1 cm 2 for men and 19.9 cm 2 for women.

| Statistical analyses
All statistical analyses were conducted using EZR ver.1.54, 19 which is for R, and all p-values were two-sided.We used the two-sided α level of 0.05.Our primary analyses were assessment of the survival association of postoperative CXI and dynamic change in CXI with disease-free and overall survival.All other tests, including assessment of risk estimates, represented secondary analyses.
Data are expressed as median and interquartile range or number and ratio.Univariate analysis was performed using the Mann-Whitney U-test, chi-square test, or Fisher's exact test, as appropriate.Preoperative CXI cut-off level for each sex was determined by maximizing the Youden index (the sum of sensitivity and specificity) for predicting 5-year survival on each ROC curve among the patients who were followed up for over 5 years or who died within 5 years of follow-up.
The Kaplan-Meier method was used to estimate cumulative survival probabilities, and the differences between groups were compared using the log-rank test.
In the sensitivity analysis of clinicopathologic variables for survival after hepatic resection for HCC, the cut-off value of postoperative CXI for each sex was determined based on the ROC analysis of survival status at 5-year follow-up in the same way as that of preoperative CXI.As an exploratory analysis, to evaluate the changes in the numerical value of CXI from before to after surgery, the ratio of postoperative CXI to preoperative CXI was calculated as an indicator of the changes in the numerical value of CXI from before to after surgery.We then investigated the prognostic value of this ratio using univariate and multivariate analyses.

| Univariate and multivariate analyses of clinicopathological variables in relation to disease-free survival after hepatic resection for HCC
Table 2 shows the association between clinicopathological variables and disease-free survival after hepatic resection for HCC.In the univariate analysis, disease-free survival was significantly associ-

| Univariate and multivariate analyses of clinicopathological variables in relation to overall survival after hepatic resection for HCC
Table 3 shows the association between clinicopathological variables and overall survival after hepatic resection for HCC.In the univariate analysis, overall survival was significantly associated

| Survival according to dynamic changes in CXI from before to after hepatic resection for HCC, and clinicopathological variables affecting the dynamic changes in CXI
We assessed the association of dynamic changes in CXI from before to after hepatic resection for HCC with disease-free survival and overall survival, and investigated clinicopathological variables The chi-square test was performed to compare categorical variables, while Mann-Whitney U-test was performed to compare continuous variables.

TA B L E 1 Clinicopathological characteristics of enrolled patients and univariate analysis of clinicopathological variables in relation to postoperative cachexia index (CXI).
affecting the dynamic changes in CXI.Among patients with preoperative high CXI, although there was no statistical difference in disease-free survival (p = 0.71, Figure 3A), patients with a change from high preoperative CXI to low postoperative CXI had significantly worse overall survival than those with a maintained high CXI from before to after surgery (p = 0.03, Figure 3B).A change from high preoperative CXI to low postoperative CXI was significantly associated with longer operation time (p = 0.04), greater intraoperative bleeding (p < 0.01), and intraoperative blood transfusion (p < 0.01) (Table 4).Among patients with low preoperative CXI, patients with a change from low preoperative CXI to high postoperative CXI had a tendency toward better disease-free survival (p = 0.19, Figure 3C) and had significantly better overall survival compared to those with a maintained low CXI from before to after surgery (p = 0.03, Figure 3D).Older age was significantly associated with a change from low preoperative CXI to high postoperative CXI (p < 0.01) (Table 4).Among the patients with low postoperative CXI, patients with high preoperative CXI showed better diseasefree and overall survival compared with those in patients with low preoperative CXI.Also, among the patients with high postoperative CXI, patients with high preoperative CXI showed better disease-free and overall survival compared with that in patients with low preoperative CXI (Figure S2).Regardless of preoperative CXI status, patients with high postoperative CXI tended to receive repeated hepatic resection as a treatment for recurrence of HCC, while patients who maintained low CXI from before to after surgery tended to receive best supportive care (Table 4).Due to the limited number of cases, there were no significant differences between the four subgroups classified by the status of preoperative and postoperative CXI.However, differences in overall survival were observed according to treatment after recurrence in both high postoperative CXI group (p = 0.01) and low postoperative CXI group (p = 0.02) (Figure S3).In both the high and low postoperative CXI groups, patients who received repeated hepatic resection or radiofrequency ablation (RFA) had significantly better overall survival compared to those who received chemotherapy, TACE, or best supportive care (p < 0.01, Figure S3).

| Preoperative CXI, postoperative CXI, and ratio of perioperative change in CXI and survival after hepatic resection for HCC
As a sensitivity analysis, we investigated the prognostic value of postoperative CXI using the different cut-off values of preoperative CXI.Similar to the cut-off values of preoperative CXI, based on the ROC analysis of survival status at 5-year follow-up, the optimal cut-off values of postoperative CXI for each sex (Figure S4) were determined to be 11.4 for men and 19.0 for women, with areas under the curve of 0.603 (95% CI: 0.458-0.748)and 0.600 (95% CI: 0.154-1.00),respectively.When postoperative CXI was classified into the two groups using these cut-off values, postoperative low CXI had a trend toward worse disease-free survival (p = 0.09), and was significantly associated with worse overall survival (p < 0.01) (Figure S5).
As an exploratory analysis, the ratio of postoperative CXI to preoperative CXI was calculated and we investigated the prognostic value of this ratio.The median (IQR) value of the ratio of perioperative change in CXI, the ratio of postoperative CXI to preoperative CXI, was 1.19 (0.85-1.60).The cut-off value of the ratio of perioperative change in CXI was determined to be 0.50 based on the median absolute distance from 1.00 of the ratios of postoperative CXI to preoperative CXI.Low ratio of perioperative change in CXI was significantly associated with worse disease-free survival (p < 0.01) and overall survival (p = 0.03) (Figure S6).In the multivariate analysis, low ratio of perioperative change in CXI was significantly associated with worse disease-free (p < 0.01, Table S1) and overall survival (p < 0.01, Table S2).

| DISCUSS ION
In the present study, we found that postoperative CXI and dynamic changes in CXI from before to after hepatic resection were significantly associated with overall survival in patients with HCC.
F I G U R E 2 Kaplan-Meier curves of the disease-free survival (A) and overall survival (B) after hepatic resection for hepatocellular carcinoma according to the status of postoperative cachexia index.Postoperative low cachexia index was significantly associated with worse disease-free survival (p < 0.01) and overall survival (p < 0.01).Underlying the pathogenesis of cachexia are many inflammatory cytokines such as tumor necrosis factor-alpha (TNFα), interleukin-1beta (IL-1β), and interleukin-6 (IL-6), which cause malnutrition and loss of skeletal muscle mass. 20,21Given these inflammatory cytokines are associated with increased NLR, 22 CXI consisting of skeletal muscle index, NLR, and serum albumin level, may be associated with these inflammatory cytokines.Surgery itself also induces these inflammatory cytokines through the activation of macrophages and neutrophils. 23The majority of postoperative patients have inadequate protein and energy intake. 24stoperative complications cause a prolonged systemic inflammatory response, decreased nutritional intake, and delayed ambulation, 3,10 and sustained inflammation causes skeletal muscle wasting 20 and persistent anorexia. 21Therefore, surgical stress may worsen the degree of preoperative cachexia or contribute to developing postoperative cachexia.Since postoperative cachexia, represented by low postoperative CXI, was associated with poor overall survival, perioperative exacerbation of cachexia should be avoided.Perioperative risk stratification based on cachexia may allow for an improved treatment strategy for HCC.

(A) ( B)
In this study, patients with low CXI preoperatively but who improved their CXI postoperatively were younger than those who did not improve their low CXI until postoperatively.Surgical resection of tumors can potentially improve cancer cachexia because it removes the tumors producing inflammatory cytokines that cause the status of cachexia.Given aging leads to delayed wound healing, increased susceptibility to infection, loss of muscle quality, and chronic inflammation called inflammaging, 25,26 it may be difficult for older people to improve their cachexia even after tumor removal.Therefore, older people with preoperative cachexia may require more intensive preoperative nutritional and physical therapy to improve their cachexia.In patients with severe nutritional risk, nutritional support for 10-14 days prior to major surgery and the addition of oral nutritional supplements to the regular diet both during hospitalization and after discharge have been recommended. 27Long-chain N-3 fatty acids have reduced inflammatory cytokines such as IL-6 or C-reactive protein and resting energy expenditure in cancer patients. 27Both preoperative and postoperative exercise interventions have been shown to increase a patient's muscle mass. 27,28CXI can potentially reflect the effect of these interventions on patient's condition and might be a useful biomarker that contributes to comprehensive evaluation.On the other hand, many of the patients with high CXI preoperatively but worsened CXI postoperatively had a greater degree of surgical stress, including longer operation time, greater intraoperative blood loss, and blood transfusion.Severe surgical stress results in the overproduction of inflammatory cytokines, 23 which may become the trigger for cachexia.In fact, excessive intraoperative blood loss is undesirable because it leads to prolonged operation time and blood transfusion, and is associated with poor prognosis of HCC. 29 For patients without preoperative cachexia, massive blood loss should be avoided as much as possible to prevent the formation of postoperative cachexia.A previous study has shown that laparoscopic hepatic resection was superior to open hepatic resection regarding intraoperative blood loss and blood transfusion rate. 30

3 | RE SULTS 3 . 1 |
Patient characteristics and univariate analysis of clinicopathological variables in relation to postoperative CXI Clinicopathological characteristics of the enrolled patients are presented in Table 1 as median and interquartile range (IQR), or as number and ratio.The median (IQR) follow-up period of 131 patients after hepatectomy was 5.2 (3.0-7.1)years.During this follow-up, 83 of 131 patients experienced tumor recurrence (63.4%), and the median (IQR) time to recurrence following hepatic resection was 1.6 (0.7-3.0) years.At the time of postoperative CXI calculation, 10 of 131 patients had developed a recurrence.In the current study, the 5-year disease-free survival and overall survival rates after hepatic resection for HCC were 55.7% and 76.3%, respectively.The median (IQR) values of preoperative and postoperative CXI (Figure 1) were 23.5 (14.2-33.6)and 25.8 (16.0-39.2),respectively.Eighty of 131 patients (61%) increased their postoperative CXI compared to their preoperative CXI, while 51 patients (39%) decreased their postoperative CXI.Based on the ROC analysis of the survival status at the 5-year follow-up, the optimal cutoff values of preoperative CXI for each sex (Figure S1) were set as 11.5 for men and 7.8 for women with areas under the curve of 0.599 (95% confidence interval (CI): 0.455-0.742)and 0.800 (95% CI: 0.468-1.00),respectively.Table 1 also shows the univariate analysis of clinicopathological variables in relation to postoperative CXI, of which the cut-off values for each sex were the same as those of preoperative CXI.In the univariate analysis, low postoperative CXI was associated with older age (p = 0.045), greater tumor size (p < 0.01), longer operation time (p = 0.047), greater intraoperative bleeding (p < 0.01), intraoperative blood transfusion (p < 0.01), and low preoperative CXI (p < 0.01).
Years after surgery Overall survival rate Postoperative high CXI (n=111) Postoperative low CXI (n=20) Postoperative high CXI (n=111) Postoperative low CXI (n=20) change in CXI was associated with both disease-free and overall survival.Our findings indicated the potential of dynamic change in CXI from before to after surgery to enable the fine risk stratification of patients, leading to the development of a personalized treatment strategy for HCC.
Minimally invasive surgery can potentially contribute to maintaining the CXI.Furthermore, for patients with low postoperative CXI, nutritional and physical intervention is required to improve their prognosis.This study had several limitations.First, there was a potential risk of selection bias because this study was a retrospective review.Second, given the number of subjects in the subgroup analysis of preoperative CXI, the statistical power to investigate the effects of dynamic changes in CXI on survival was limited.More comprehensive large-scale studies are expected to validate our findings.Third, postoperative CXI in this study was evaluated at approximately 3 months after liver resection for HCC as the first surveillance is usually performed at 3 months.However, skeletal muscle can change in the early operative period and the serial change of the skeletal muscle in the perioperative period is unknown.Therefore, further studies are needed to evaluate CXI at serial time points including the perioperative period and later than 3 months after surgery.Fourth, the effect of preoperative interventions to improve patient conditions could not be evaluated in this retrospective study.As several studies indicated that nutritional and physical interventions can improve cancer cachexia,27,28 further prospective studies are necessary to confirm whether the improvement of CXI by interventions can contribute to a better prognosis.In conclusion, not only postoperative CXI but also dynamic changes in CXI from pre-to post-hepatectomy can be a prognostic indicator in patients with HCC after hepatic resection.Furthermore, surgical stress in patients with high preoperative CXI and age in patients with low preoperative CXI may be associated with lower postoperative CXI, leading to worse overall survival.AUTH O R CO NTR I B UTI O N SMA and KH developed the main concept and designed the study.YY, KO, KF, MT, YS, SO, and MM were responsible for acquisition of F I G U R E 3 Kaplan-Meier curves of the disease-free and overall survival after hepatic resection for hepatocellular carcinoma according to the status of postoperative cachexia index in preoperative high (A and B) or low (C and D) cachexia index group.Postoperative low cachexia index was significantly associated with worse overall survival both in preoperative high cachexia index group (p = 0.03) and in preoperative low cachexia index group (p = 0.02).

Variables Total (n = 131) Postoperative CXI p-value a High (n = 111) Low (n = 20)
Abbreviations: AFP, alpha-fetoprotein; BTF, blood transfusion; CXI, cachexia index; HBsAg, hepatitis B surface antigen; HCV-Ab, hepatitis C virus antibody; ICG R15 , retention rate of indocyanine green at 15 min; PIVKA-II, protein induced by vitamin K absence or antagonist-II.a Univariate and multivariate analyses of prognostic factors for disease-free survival in patients with hepatocellular carcinoma after hepatic resection.
scatter plots representing changes from preoperative to postoperative CXI (median with interquartile range).The median (interquartile range) values of preoperative and postoperative CXI were 23.5 (14.2-33.6)and25.8(16.0-39.2),respectively.TA B L E 2Abbreviations: AFP, alpha-fetoprotein; BTF, blood transfusion; CI, confidence interval; CXI, cachexia index; HBsAg, hepatitis B surface antigen; HCV-Ab, hepatitis C virus antibody; HR, hazard ratio; ICG R15 , retention rate of indocyanine green at 15 min; PIVKA-II, protein induced by vitamin K absence or antagonist-II; NS, not significant.a Univariate and multivariate analyses of prognostic factors for overall survival in patients with hepatocellular carcinoma after hepatic resection.