A total of 105 patients with primary HCC were treated using hepatic partial resection at the Department of Digestive Surgery, Kagoshima University Graduate School of Medicine (Kagoshima, Japan) between January 1996 and December 2002. The patients comprised 83 men and 22 women, with a median age of 64.7 years. Of these patients, 19 patients tested positive for the hepatitis B surface antigen, 73 were positive for the antibody to HCV and 13 were negative for the two viruses. The mean tumor diameter was 46.4 mm (range 10–150 mm). Macroscopically, 67 cases (63.8%) had simple nodular tumors, whereas 38 cases (36.2%) comprised other types. Microscopically, 20 tumors (19.0%) were well-differentiated HCC, 74 tumors (70.5%) were moderately differentiated and 8 tumors (7.6%) were poorly differentiated. A total of 36 tumors had infiltrated blood vessels and 27 cases presented with intrahepatic metastasis. For comorbidities, 41 cases presented with hypertension, 34 cases with diabetes mellitus and 12 cases with hyperlipidemia (Table I). Patients with a history of treatment for HCC and synchronous or metachronous multiple cancers in other organs were excluded from the present study. Follow-up data were obtained from all patients post-surgery, with a median follow-up time of 53 months.

The present study was approved by the Ethics Committees of the Graduate School of Medical and Dental Sciences, Kagoshima University (Kagoshima, Japan; registration number 25–39) and was conducted according to the Declaration of Helsinki. Written informed consent was obtained from each patient, including consent to publish and to report individual data from the participant.

Tissues were immersed into 10% neutral buffered formalin for 24 h at room temperature immediately after resection and then embedded with paraffin. Consecutive 4-µm sections were cut from each paraffin-embedded block. Sections were immunostained with anti-CD68 (cat. no. M0876; Dako; Agilent Technologies, Inc., Santa Clara, CA, USA), anti-CD163 (cat. no. Mob460; Diagnostic BioSystems, Inc., Pleasanton, CA, USA) and FRβ antibody, which was kindly provided by Professor Matsuyama, Department of Immunology, Graduate School of Medical and Dental Sciences, Kagoshima University (23). Briefly, following blocking with 0.3% H₂O₂/methanol for 30 min at room temperature, specimens were blocked with PBS containing 5% normal horse serum (Vector Laboratories Inc., Burlingame, CA, USA) for 30 min at room temperature. Anti-CD68, anti-CD163 and FRβ antibodies were used at a dilution of 1:100. Following overnight incubation at 4°C with the primary antibodies, specimens were briefly washed in PBS and incubated at room temperature for 30 min with polymer detection system used at 1:1 dilution (cat. no. 424132; Histofine simple stain MAX PO; Nichirei Biosciences, Tokyo, Japan). The specimens were washed with PBS and developed for 2 min at room temperature using diaminobenzidine solution (Dako; Agilent Technologies, Inc.). Specimens were subsequently washed with water and counterstained with Meyer's hematoxylin for 30 sec at room temperature (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany).

To evaluate the results of immunohistochemical staining, the immunostained sections were scanned using a BX50-32 light microscope and DP71-SET-A digital camera (both Olympus Corporation, Tokyo, Japan) at a magnification of ×40. The sections were analyzed with cellSens Standard software (v1.11; Olympus Corporation). The six fields of intra-tumor and peri-tumor lesions with the greatest staining intensity in each specimen were selected, the number positive cells in each field was counted manually using high power (×200 magnification) light microscopy and the mean number of positive cells for each specimen was calculated. The peri-tumoral lesion was defined as the area within 2 mm from the external edge of the tumor. The intra-tumoral lesion was defined as the remaining area of the tumor. Two investigators assessed the slides without knowing the clinicopathological features and were blinded to each other's evaluation. They agreed on all slides examined. The mean values for the positive cells in the two locations were evaluated.

Clinicopathological factors selected for evaluation included preoperative laboratory values [including indocyanine green retention rate at 15 min (ICGR15) and tumor markers α-fetoprotein (AFP) and des-gamma-carboxy prothrombin (DCP)]. Histopathological diagnosis was based on evaluation of tumor size, the number of tumor nodules, lymph node metastasis and infiltration of blood vessels (portal vein and hepatic artery and/or vein). The tumor stage and pathological parameters were determined according to the General Rules for the Clinical and Pathological Study of Primary Liver Cancer (24). The overall survival (OS) was calculated from the date of resection to the date of mortality regardless of the cause. Recurrence-free survival (RFS) was calculated from the date of resection to the date that the tumor recurrence was diagnosed or from the date of resection to the last visit if recurrence was not diagnosed.

An unpaired t-test was used to evaluate continuous variables. Analysis of variance (ANOVA) was used to determine whether there was a significant difference between groups of tumor stage and histological differentiation. When ANOVA showed a significant result, the Tukey honest significance difference test was used to define between which groups there was a significant difference. The cumulative OS and RFS rates were calculated using the Kaplan-Meier method and tested using the Generalized Wilcoxon test. Data are presented as the mean ± standard deviation. Correlations were analyzed using Pearson's correlation coefficient. Statistical analyses were performed using the SPSS statistical software package (version 24; IBM Corp., Armonk, NY, USA). P<0.05 was considered to indicate a statistically significant difference.

Positive staining for CD68, CD163 and FRβ was assessed in intra-tumor and peri-tumor lesions, and was mainly observed in the cytoplasm of stroma cells. The majority of tumor cells and hepatic cells were negatively stained (Fig. 1).

To elucidate the biological significance of CD68 and CD163 expression in HCC, the number of CD68 and CD163 positive cells was compared with the clinicopathological features of 105 patients. The number of CD68 positive cells was significantly correlated with that of CD163 positive cells (R=0.823; P<0.001; Fig. 2). The number of CD68 positive cells was significantly higher in patients with stage IV cancer compared with stage III cancer (P=0.035; Table II) and lower in patients with simple nodular type tumors, which had good prognosis, compared with patients who were not diagnosed with simple nodular type tumors (P=0.021; Table II). CD163 positive cells were significantly more abundant in patients with median tumor size ≥3.5 cm (P=0.049; Table III) and in patients with poorly differentiated HCC (P=0.038; Table III). The CD163/68 ratio was also compared with clinicopathological features. The ratio was significantly lower in patients with stage IV cancer (P=0.048 vs. stage I; P=0.017 vs. stage III; Table IV), DCP abnormalities (P=0.047; Table IV), infiltration to blood vessels (P=0.016; Table IV) and intrahepatic metastasis (P=0.050; Table IV). A lower ratio appeared to be associated with the clinical malignancy.

To clarify the role of the M2 macrophage in HCC, further examination was performed to evaluate the expression of FRβ as another M2 macrophage marker, as well as the number of FRβ positive cells compared with the clinicopathological features and CD68 and CD163 expression. No significant correlation was observed between FRβ positive cells and clinicopathological features (Table V); however the number of FRβ positive cells was significantly correlated with CD68 expression and with CD163 expression (R=0.694, P<0.001; R=0.471, P<0.001, respectively; data not shown).

Patients were divided into subgroups depending on the number of CD68 and CD163 cells observed and the CD163/68 ratio, as well as into FRβ positive or negative groups. The OS and RFS for the groups were then compared. No significant differences in OS and RFS were identified between groups (Figs. 3–6).

When a cut-off value of 1.0 was set for the CD163/68 ratio, OS (for the lower ratio group was significantly shorter compared with the higher ratio group median survival 84.2 vs. 72.3 months; P=0.046; Fig. 7). However, no significant differences in RFS were identified between the two groups (Fig. 7).