The importance of body composition assessment for patients with advanced hepatocellular carcinoma by bioelectrical impedance analysis in lenvatinib treatment

Background and aims The aim of this study was to investigate the relationship between body composition before lenvatinib treatment and prognosis in patients with hepatocellular carcinoma (HCC). We also assessed the relationship between the rate of change in body composition after lenvatinib treatment and prognosis. Methods Eighty-one patients with advanced HCC who were treated with lenvatinib were enrolled. We assessed prognosis, various clinical data, body composition parameters obtained by bioelectrical impedance analysis (BIA), and handgrip strength. Results Multivariate analysis showed that an extracellular water to total body water ratio (ECW/TBW) ≤ 0.400 at treatment initiation was associated with longer overall survival (OS), progression-free survival (PFS), and post-progression survival (PPS) (OS: hazard ratio [H0R], 4.72; 95% CI, 12.03–11.00; P < 0.001; PFS: HR, 2.66; 95% CI, 1.33–5.34; P = 0.0057; PPS: HR, 3.08; 95% CI, 1.32–7.18; P = 0.0093). Multivariate analysis also showed that the skeletal muscle mass index (SMI) of the arm at treatment initiation was associated with a longer PFS (HR, 2.12; 95% CI, 1.23–3.64; P = 0.0069). In the group with an ECW/TBW ≤ 0.400 before lenvatinib treatment, univariate analysis showed that the rate of change in only the arm SMI was associated with a longer OS and PFS. Conclusion Body composition assessment by BIA before and after lenvatinib treatment is useful in predicting prognosis in lenvatinib-treated patients with HCC.

Introduction Hepatocellular carcinoma (HCC) is one of the leading causes of cancer-related death worldwide [1][2][3]. HCC commonly occurs in patients with chronic hepatitis or liver cirrhosis due to hepatitis B virus, hepatitis C virus, alcohol use, nonalcoholic steatohepatitis, or diabetes [4]. In recent years, the prognosis of HCC has improved due to progress in imaging technology and therapeutic strategies. However, advanced HCC still has a poor prognosis.
Lenvatinib was approved for insurance coverage in Japan in March 2018 for advanced HCC as a first-line therapy. The REFLECT study demonstrated noninferiority of lenvatinib compared with sorafenib [5]. In that study, lenvatinib had a significantly better response as assessed by the modified Response Evaluation Criteria in Solid Tumors and a longer progression-free survival (PFS) than sorafenib, although overall survival (OS) was similar between the two drugs [6].
A diagnosis of sarcopenia requires both a decrease in handgrip strength and skeletal muscle mass loss according to the diagnostic criteria of the Japan Society of Hepatology [7]. The prognosis of cirrhotic patients with sarcopenia is generally poor [8]. Previous studies have reported that in patients with HCC, sarcopenia is associated with a worse prognosis and an increased rate of cancer recurrence [9,10]. In addition to the collective measure of sarcopenia, skeletal muscle mass loss and a decrease in handgrip strength are also independently associated with a poor prognosis [11][12][13][14][15][16]. However, the studies that led to these conclusions used single-slice computed tomography (CT), not bioelectrical impedance analysis (BIA), to measure skeletal muscle mass. In this study, we investigated the relationship between body composition before lenvatinib treatment and prognosis in patients with HCC, as well as the relationship between the rate of change in body composition during lenvatinib treatment and prognosis by measuring skeletal muscle mass with BIA.

Patients treated with lenvatinib
The study was conducted in accordance with the Declaration of Helsinki, and the study protocol was approved by the Hiroshima University Hospital Institutional Ethics Committee (E-882). Written informed consent was obtained from each participating patient. This retrospective cohort study included 81 patients with advanced HCC who had been treated with lenvatinib at our hospital between April 2018 and April 2021. Of the 81 patients, 26 were included in our previous study [17]. These patients were considered unfit for surgery, liver transplantation, repeat locoregional therapy, repeat transcatheter arterial chemoembolization, or repeat hepatic arterial infusion chemotherapy. The inclusion criteria for treatment with lenvatinib were an Eastern Cooperative Oncology Group performance status score of � 1 and Child-Pugh class A. In this study, 61 patients received lenvatinib as their first-line therapy and 20 patients had been previously treated with other molecularly targeted therapies or immune checkpoint inhibitors. The exclusion criteria for this study were (1) Child-Pugh class B or ascites, (2) a short observation period (< 1 month), (3) absence of proper image analysis, and (4) absence of body composition measurements by BIA before and after the start of lenvatinib treatment.
Tumour staging was based on the Tumour-Node-Metastasis staging system of the Liver Cancer Study Group of Japan [18]. Tumours in stage I patients had three features: (1) they were solitary, (2) they measured < 2 cm in diameter, and (3) they had no vessel invasion (n = 0). Those in stage II fulfilled two of the above three features (n = 13, 16%); those in stage III fulfilled one of the above three features (n = 27, 33%); those in stage IVA fulfilled none of the above three features, but had lymph node metastases and no distant or intrahepatic metastases (n = 15, 19%); and those in stage IVB fulfilled none of the above features, but had distant metastases (n = 26, 32%).

Body composition analysis of patients receiving lenvatinib
Body composition was measured by BIA using Inbody 720 1 (BioSpace Co. Ltd., Seoul, Korea). The BIA used in this study was a 4-pole, 8-point direct segment multi-frequency BIA that can evaluate not only total muscle mass and fat mass, but also muscle mass and fat mass specifically of the right arm, left arm, trunk, right leg, and left leg. The first body composition measurement was taken within 1 month of starting lenvatinib treatment, and another measurement was taken around 1 month after the end of lenvatinib treatment. We had initially planned to take body composition assessments 1 month after the start of lenvatinib treatment and every 1 to 3 months thereafter, but some patients did not cooperate, and thus, we could not be performed this analysis as planned. The skeletal muscle mass index (SMI) (kg/m 2 ) was calculated by dividing the limb skeletal muscle mass (kg) by the square of the height (m 2 ). Arm SMI was calculated by dividing the arm's skeletal muscle mass (kg) by the square of the height (m 2 ). Leg SMI was calculated by dividing the leg's skeletal muscle mass (kg) by the square of the height (m 2 ). According to the Japan Society of Hepatology criteria, 7 patients with SMI values < 7.0 kg/m 2 for males and < 5.7 kg/m 2 for females were defined as having skeletal muscle mass loss. The BIA also measured the total fat mass and the extracellular water to total body water ratio (ECW/TBW), an oedema index. Because excessive ECW results in an oedematous state, ECW/TBW is an index that reflects the degree of oedema in a person. An ECW/ TBW > 0.400 was defined as an overhydrated state [17,19]. The rate of change in body composition was calculated by subtracting the data before (or within the first month of) lenvatinib treatment from the data after lenvatinib treatment and dividing by the data before lenvatinib treatment.

Statistical analysis
In this study, we set the cutoff values for SMI and handgrip strength according to the sarcopenia diagnostic criteria of the Japanese Society of Hepatology, and those for arm SMI and leg SMI at the median value. Continuous variables were expressed as median (range), while categorical variables were expressed as absolute and relative frequencies. The Mann-Whitney U test was used to compare continuous data. Either Pearson's chi-square test or Fisher's exact test was used to compare significant differences in the distribution of categorical variables. The percent change in body composition was correlated using the nonparametric Spearman's rank correlation coefficient. OS, PFS, and post-progression survival (PPS) were evaluated using Kaplan-Meier survival curves and a log-rank test. Logistic regression analysis and Cox regression analysis were carried out for multivariate analysis. Only those factors with a P < 0.05 in the univariate analysis were subsequently assessed in the multivariate analysis. A P value < 0.05 denoted a statistically significant difference. All statistical analyses were carried out with the Predictive Analytics Software R version 3.5.2.

Amount of change in body composition during lenvatinib treatment
The median duration from the start of lenvatinib treatment to the 2nd body composition assessment after starting lenvatinib treatment was 5.9 months (range: 0.8-24.4 months). The correlation between the amount of change in body composition and the duration of lenvatinib treatment is shown in Fig 4. Total SMI and leg SMI measured by BIA showed an increase after treatment in skeletal muscle mass in 36 and 40 cases, respectively. Therefore, we examined the correlation between the rate of change in ECW/TBW and the rate of change in SMI and leg SMI and found that it was significantly correlated (r = 0.492, P < 0.001 and r = 0.586, P < 0.001; Fig 5).

Association with rate of change in body composition and prognosis
The median rates of change in arm SMI, leg SMI, ECW/TBW, and body fat percentage were −7.22%, −0.12%, −2.09%, and −9.57%, respectively. We set the median as the cutoff value and examined the relationship between the rate of change in body composition and prognosis, but none of the results were significantly different. Next, we examined the association between the rate of change in body composition and prognosis in a subgroup of patients with an ECW/ TBW < 0.400 before lenvatinib treatment, as this group had a better prognosis. The arm SMI change > −7.22% group had a longer OS and PFS compared to the change < −7.22% group (Fig 6). There were no significant differences between prognosis and the rates of change in leg SMI, ECW/TBW, and body fat percentage.

Discussion
In HCC treatment, only our previous report has examined body composition measurement in BIA [17], and there are no reports on body composition changes in BIA. ECW/TBW is an index that reflects oedema, and a high ECW/TBW indicates increased fluid retention [19,20].  Previously, we reported that patients with an ECW/TBW > 0.400 have a shorter duration until reduction or withdrawal of lenvatinib and have difficulty maintaining the relative dose intensity, though PFS did not differ based on ECW/TBW [17]. In this study, the number of cases included was higher than in the previous study, offering a greater statistical power. In addition, OS and PPS were also evaluated in addition to PFS due to the longer follow-up duration. In this study, we found that an ECW/TBW > 0.400 before lenvatinib treatment was associated with shorter OS, PFS, and PPS. Notably, ECW/TBW increases as cirrhosis progresses, with the median value of ECW/TBW being 0.393 in Child-Pugh class A patients, 0.402 in Child-Pugh class B patients, and 0.405 in Child-Pugh class C patients [21]. Based on this, patients with an ECW/TBW > 0.400 were speculated to have a poor liver functional reserve, elevated blood   levels of lenvatinib, and earlier onset of adverse events. In this study, the median duration to first dose reduction or drug withdrawal was 1.6 months. The median duration to first dose reduction or drug withdrawal in the group with ECW/TBW > 0.400 was 0.6 months, which was significantly shorter than that in the group with ECW/TBW < 0.400, which was 2.0 months. The short duration of the first dose reduction or drug withdrawal may suggest a decrease in relative dose intensity (Fig 7). Therefore, an ECW/TBW > 0.400 may have contributed to the shortened OS and PFS because these patients were less likely to continue lenvatinib treatment. With this result, we reconfirmed that ECW/TBW is useful in screening Child-Pugh class A patients for tolerability of lenvatinib treatment. Previous reports have shown that loss of skeletal muscle mass is associated with a worse prognosis in HCC [11][12][13][14][15][16]. Other reports indicate that skeletal muscle mass loss is a prognostic factor in lenvatinib treatment [15,16]. There are some methods to measure SMI: single-slice CT, BIA, and dual energy X-ray absorptiometry (DXA). Previous studies have used single-slice CT to measure skeletal muscle mass [11][12][13][14][15][16]. Single-slice CT is a common test, but is difficult to repeat due to radiation exposure. DXA also involves radiation exposure, but the low dose allows for repeat examinations and regional body composition assessment. BIA is simple, noninvasive and repeatable as DXA. In this study, we used BIA to measure body composition because BIA has the greatest advantage over single-slice CT or DXA in that there is no radiation exposure no matter how many times it is measured. When we measured SMI using BIA, however, we found that SMI was not associated with prognosis in this study, as previous reports using CT have found. This may be due to some limitations of BIA. Namely, the chemical composition of fat-free masses (i.e., water, proteins, etc.) have considerable inter-and intraindividual variability because of changes in fat-free mass that occur during disease states [22]. In addition, one report showed a decrease in skeletal muscle mass measured by BIA after haemodialysis [23]. Therefore, SMI may be overestimated due to oedema when it is measured by BIA. In fact, in our study, there was a positive correlation between the rate of change in ECW/TBW and the rate of change in SMI after lenvatinib treatment, and thus, SMI before treatment may have been affected by the oedema rate. Moreover, another study reported that the arm SMI value was better associated with prognosis compared to the leg SMI value in patients with cirrhosis [24]. When SMI was examined separately for the upper and lower limbs in this study, the group with a low SMI of the arm had a shorter PFS than the group with a high SMI of the arm. These data suggest that SMI as measured by BIA may be higher than it really is when there is fluid overload, with leg SMI being more susceptible to oedema. In this study, we examined the correlation between the rate of change of ECW/TBW for arm SMI and leg SMI. ECW/TBW showed a positive correlation with the leg SMI, but not with the arm SMI. Therefore, to accurately assess SMI during lenvatinib treatment in HCC patients, it should be assessed with arm SMI.
We previously reported that tyrosine kinase inhibitor treatment for HCC caused skeletal muscle loss [25]. In this study, we found that the rate of change in arm SMI had an impact on prognosis in the group with an ECW/TBW < 0.400. Several factors are associated with skeletal muscle mass loss. A previous study showed that reasons for skeletal muscle depletion in HCC include decreased physical activity and poor nutrition due to disease progression and the adverse effects of treatment, as well as increased expression of inflammatory cytokines in patients with cancer [9]. Furthermore, sarcopenia is associated with alterations in the phosphoinositide PI3k/Akt/mTOR signalling pathway [26][27][28][29], which is associated with muscle protein synthesis. Both sorafenib and lenvatinib, used for HCC treatment, inhibit VEGFRmediated signalling and carnitine transporters, which may suppress signalling in the downstream PI3k/Akt/mTOR pathway, leading to subsequent skeletal muscle mass loss [30]. Skeletal muscle mass loss caused by molecularly targeted drugs is associated with poor prognosis in non-small cell lung cancer [31,32], but this has not been reported in HCC until now. This is the first report that skeletal muscle mass loss during lenvatinib treatment is associated with a worse prognosis in HCC. When checking the rate of change of skeletal muscle mass in BIA, it is important to check it using the rate of change of arm SMI. This study has several limitations, such as the small sample size, retrospective design, and short-term nature. Also, the timing for the second assessment of body composition varied between patients. In future studies, we expect the number of cases available for study to increase, the follow-up duration to be lengthened, and the timing of measuring body composition after lenvatinib treatment to be matched.
In conclusion, body composition measurement using BIA is simple, non-invasive, repeatable, and does not involve radiation exposure. Additionally, BIA can measure not only skeletal muscle mass, but also body fat percentage and ECW/TBW. Since arm SMI is associated with PFS, and ECW/TBW is associated with OS, PFS, and PPS, assessing body composition with BIA before lenvatinib treatment is extremely useful as a prognostic factor. We also found that a great decrease in arm SMI was associated with prognosis in patients with an ECW/ TBW � 0.400 before treatment. Our data indicate that it is important to assess body composition before as well as during lenvatinib treatment.