This retrospective study was conducted at the Kurume University Hospital and Omuta City Hospital. The protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the ethics committees of the Kurume University School of Medicine (approval no. 19203). An opt-out approach was employed to obtain informed consent from the patients, and personal information was protected during data collection.

The patient inclusion criteria for this study were as follows: i) diagnosis of unresectable HCC according to the Barcelona Clinic Liver Cancer (BCLC) staging system (15); ii) age ≥18 years; iii) Eastern Cooperative Oncology Group performance status 0; iv) history of pretreatment with MTAs; and v) completion of follow-up until death or study cessation (August 30, 2020). The patient exclusion criteria were as follows: i) history of a malignant tumor other than HCC in the 5 years preceding the study; ii) participation in any clinical trial; iii) Child-Pugh class C; iv) creatinine >1.5 mg/dl; v) chronic heart failure; vi) infiltrative HCC; vii) presence of ascites; viii) esophageal varices with a high risk of rupture; and ix) history of liver transplantation.

A total of 16 consecutive HCC patients who received RAM treatment between September 19, 2019 and July 31, 2020 were registered. The data cutoff date for this analysis was August 31, 2020. Patients meeting any of the exclusion criteria were excluded from the analysis (n=2). Thus, a total of 14 patients were enrolled in the study.

CONUT score, used to assess nutritional status, was calculated from serum albumin level, total cholesterol level, and lymphocyte count, as previously described (9) Albumin concentrations of ≥3.5, 3.0-3.49, 2.5-2.99, and <2.5 g/dl were scored as 0, 2, 4 and 6 points, respectively. Total lymphocyte counts of ≥1,600, 1,200-1,599, 800-1,199, and <800/µl were scored as 0, 1, 2 and 3 points, respectively. Total cholesterol concentrations of ≥180, 140-179, 100-139 and <100 mg/dl were scored as 0, 1, 2 and 3 points, respectively. Liver function was evaluated using the albumin-bilirubin (ALBI) score, as previously described (16) It based on serum albumin and total bilirubin levels; ALBI-score = [log10 bilirubin (µmol/l) x 0.66] + [albumin (g/l) x -0.085], and was graded as following: ≤-2.60 = ALBI grade 1, >-2.60 to ≤-1.39 = ALBI grade 2, >-1.39 = ALBI grade 3).

The skeletal muscle index (SMI) was evaluated at the third lumbar vertebra level on computed tomography (CT) scans, which were obtained as part of the HCC assessment (17) SMI were calculated by normalizing the L3 skeletal muscle areas by the square of the height (m²) (18). The targets of measurement were psoas, erector spinae, quadratus lumborum, transversus abdominis, external and internal obliques, and rectus abdominis. The measurement was performed by two government-certified physical therapists (S.K. and K.H.) who were blinded to the patients' information. This analysis was performed using diagnostic software ImageJ Version 1.50 software (National Institutes of Health) (19).

We measured VFA using diagnostic CT scans at the umbilical level, as previously described (20) CT scanning was performed for HCC evaluation. The measurement was performed by two government-certified physical therapists (S.K. and K.H.) who were blinded to the patients' information. The VFA was also measured by the diagnostic software ImageJ Version 1.50 software (National Institutes of Health) (19).

Muscle quality was evaluated by measuring the IMAT content. The IMAT was calculated as the psoas muscle-to-subcutaneous fat attenuation ratio using diagnostic CT scans at the umbilical level as previously described (21) A higher IMAT indicates a greater amount of adipose tissue within the skeletal muscle (22) The SMI, VFA, and IMAT were measured by two government-certified physical therapists (S.K. and K.H.) who were blinded to patient information.

HCC was diagnosed using a combination of tests for serum tumor markers, such as AFP and des-γ-carboxy prothrombin (DCP), and imaging procedures, such as ultrasonography, computed tomography (CT), and magnetic resonance imaging (MRI). HCC was classified using the BCLC staging system (15).

RAM was intravenously injected at a dose of 8 mg/kg once every 2 weeks. Therapeutic response was evaluated according to the modified Response Evaluation Criteria in Solid Tumors (mRECIST) (23), using dynamic CT or magnetic resonance imaging. The evaluation was conducted 4-6 weeks after initiation of treatment with RAM, and thereafter, at intervals of 2-3 months until death or study cessation.

AEs and ascites were assessed every month after the initiation of RAM treatment. AEs were assessed according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.0.

Ascites were assessed by ultrasonography and CT scan images. Severe ascites were defined as diuretic-resistant ascites.

All data are expressed as frequency or median (range). All statistical analyses were performed using JMP Pro, version 14 (SAS Institute Inc.). PFS was calculated using the Kaplan-Meier method and analyzed using the log-rank test. We also performed decision tree analysis to identify factors associated with severe ascites, as previously described (24).

Patient profiles are summarized in Table I. The median patient age was 72.5 years, and 28.5% of the patients were female. The median body mass index (BMI) was 20.7 kg/m², and 78.6% of patients (11/14) showed Child-Pugh class A. ALBI grade 2 was observed in 92.3% of the patients, and ALBI 2a was observed in 42.9%. The frequencies of CONUT scores of 0-1 (normal nutrition), 2-4 (mild malnutrition), 5-8 (moderate malnutrition), and ≥9 (severe malnutrition) were 21.4, 42.9, 35.7 and 0.0%, respectively (Table I). BCLC stage B HCC was seen in 42.8% of the patients. The median IMAT and VFA were -0.51 and 79.92 cm², respectively. Prior treatment with sorafenib (SORA), SORA+LEN, and LEN was seen in 28.6 (4/14), 35.7 (5/14), and 35.7% (5/14) of the patients, respectively. The median observation period was 4.5 months (1.3-11.5 months) (Table I).

The therapeutic responses to RAM are shown in Table II. Complete response, partial response, stable disease, and progressive disease were observed in 0, 14, 36 and 50% of the patients, respectively (Table II). The overall objective response rate and disease control rate were 14 and 50%, respectively (Table II). The median PFS time was 2.1 months (Fig. 1).

AEs determined by the attending physician are shown in Table III. Ascites was the most frequent AE and was seen in 57.1% (8/14). Appetite loss and fatigue were seen in 50 (7/14) and 50% (7/14), respectively. The infusion reaction was seen in 7.2% (1/14).

RAM-related severe ascites developed in 57.1% (8/14) of the patients, and the median onset was 37.5 days (14-61 days) after RAM treatment initiation. Although ascites was treated with diuretics, RAM was discontinued in 87.5% (7/8) of patients who developed severe ascites. A representative case of the development of severe ascites treated with RAM is shown in Fig. 2A and B. Visceral inversion was seen in this case. No ascites was seen when multiple recurrent hepatic nodules developed after LEN treatment (Fig. 2A). Severe ascites developed 27 days after the initiation of treatment with RAM (Fig. 2B).

In this study, severe ascites developed in 57% of all the subjects by the time of study cessation. To determine the profiles associated with RAM-related severe ascites, decision tree analysis was performed. We revealed that the CONUT score was the first splitting variable for the development of RMA-related severe ascites. In patients with a CONUT score <5, the second splitting variable was the IMAT (Fig. 2). Although severe ascites developed in all patients with an IMAT <-0.54 and a CONUT score <5, severe ascites developed in only 16% of patients with an IMAT ≥-0.54 and a CONUT score <5 (Fig. 3).