Comparative Immunohistochemical Analysis of Clinicopathological Subgroups in Hepatocellular Carcinomas from Japan and Indonesia

Abstract

Background

Standardized pathological evaluation based on immunohistochemical (IHC) analysis could improve hepatocellular carcinoma (HCC) diagnoses worldwide. We evaluated differences in clinicopathological subgroups in HCCs from two academic institutions in Tokyo-Japan, and Jakarta-Indonesia.

Methods

Clinicopathological parameters and molecular expression patterns were evaluated in 35 HCCs from Indonesia and 41 HCCs from Japan. IHC analysis of biliary/stem cell (B/S) markers (cytokeratin 19, sal-like protein 4, epithelial cell adhesion molecule) and Wnt/β-catenin (W/B) signaling-related molecules (β-catenin, glutamine synthetase) could determine the IHC-based subgroups. For immuno-subtypes categorization, CD3/CD79α double immunohistochemistry was done to evaluate the infiltration of T and B cells. CD34 staining allowed identification of vessels that encapsulated tumor clusters (VETC).

Results

Indonesian HCC patients were mostly <60 years old (66%) with a hepatitis B virus (HBV) background (82%), in contrast to Japanese HCC patients (8% and 19%, respectively, both P < 0.001). In comparison with Japanese, Indonesian cases more frequently had >5 cm tumor size (74% vs 23%, P = 0.001), poor differentiation (40% vs 24%), portal vein invasion (80% vs 61%), and α-fetoprotein levels >500 ng/ml (45% vs 13%, P = 0.005). No significant differences were found in the proportions of B/S, W/B, and −/− subgroups from both countries. No immune-high tumors were observed among Indonesian cases, and immune-low tumors (66%) were more common than in Japanese cases (54%). VETC-positive tumors in Indonesia were significantly more common (29%), and most were in the HBV (90%) and −/− subgroups (90%), whereas Japanese VETC cases (10%, P = 0.030) were nonviral (100%) and W/B subgroups (75%).

Conclusion

IHC-based analysis more precisely reflected the clinicopathological differences of HCCs in Japan and Indonesia. These findings provide new insights into standardization attempts and HCC heterogeneity among countries.

According to the 2020 global cancer statistic estimates, liver cancer is the sixth most diagnosed cancer and the third leading cause of cancer-related deaths worldwide.¹ The overall burden from liver cancer is most pronounced in transitioning countries, and the rates are higher in Eastern Asia. The number of new liver cancer cases, which includes hepatocellular carcinoma (HCC), is predicted to increase by 55% between 2020 and 2040.²

In Japan, HCC is the fifth leading cause of death in men and women.³ HCC cases in Japan are predominantly caused by chronic hepatitis C virus (HCV) infections; however, in recent years, the proportion of HCC cases with no viral hepatitis background, known as non-B and non-C (NBNC) HCC, has dramatically increased.^3,4 Since the 1980s, Japan has operated a routine nationwide screening program for HCC, which has resulted in Japan having highly favorable outcomes for HCC treatment.⁵ The establishment of pathological diagnostic criteria for early HCC and recent advances in HCC subclassification have boosted the accuracy of HCC diagnosis and treatment strategies in Japan.⁶

As in Japan, Indonesia also has a high prevalence of liver cancer. According to the report from The Global Cancer Observatory 2020, liver cancer ranks as the fifth most common cancer and the fourth most common cause of cancer death among men and women in Indonesia.⁷ In a study of 282 HCC patients, 50% were caused by hepatitis B virus (HBV) infections and overall median survival rate was only 17 months between initial diagnosis and death. The tendency of patients to seek treatment only when clear symptoms have occurred and the lack of a surveillance program as a preventive measure are considered to be crucial factors contributing to the low survival rate of HCC patients in Indonesia.⁸

An important factor that contributes to the success of HCC treatment is the establishment of an accurate pathological diagnosis, the histological subtype, tumor grade, and vascular involvement. In addition to histopathological evaluation, recent advances in molecular subclassification by transcriptomic and genomic analyses have proposed HCC subclassifications associated with clinical features and distinct molecular pathways. Boyault et al. proposed the classification of HCC into six groups, termed G1–G6, based on 16 predictor genes,⁹ whereas Hoshida et al. classified HCC into subclasses S1–S3 based on meta-analysis of gene expression profiles.¹⁰ Although these approaches can identify more detailed molecular pathways, cost-effective and practically applicable methods that can be used in different countries with a range of clinical settings are needed.

Against this background, immunohistochemistry (IHC)-based evaluation to assess the expressions of various proteins resulting from genetic or epigenetic alterations during hepatocarcinogenesis may provide a realistic approach. The previous studies have shown that the expression of biliary/stem cell markers, which define the B/S subclass of HCC, is associated with more aggressive clinicopathological features roughly corresponding to Boyault’s G3 and Hoshida’s S1–S2 subclasses. The Wnt/β-catenin signaling activation-related marker-positive subclass (W/B group) corresponds with Boyault’s G5-G6 and Hoshida’s S3 subclasses. The remaining tumors, in which all these markers were negative (−/− group), are associated with better tumor differentiation.¹¹ Recently the role of immune cells in tumor development has received much attention, particularly since cancer immunotherapy has revolutionized conventional cancer treatment.¹² Analysis of the tumor microenvironment has revealed three distinct HCC immune subtypes, namely, immune-high, immune-mid, and immune-low. The immune-high HCC has a better prognosis compared with the immune-low subtype.¹³ HCCs with increased angiogenic factor expression showed a novel vascular pattern termed vessels encapsulating tumor clusters (VETC), which can be readily assessed by immunostaining for CD34; the presence of VETC has been shown to predict higher rates of metastasis and recurrence.¹⁴ The relationship between the HCC immune microenvironment and angiogenic factors may provide pathological insights into the optimum combination of immunotherapy and angiogenic therapy for advanced HCC treatment.¹⁵

Therefore, considering the different clinical settings and the practical ability to perform pathological analysis, we chose IHC-based subclassification to address the differences of HCC features in Japan and Indonesia. This is the first international collaborative study to examine HCC subclassification in Japan and Indonesia from the pathological perspective. Analysis of the different underlying HCC etiologies between the two countries could improve understanding of the nature of HCC heterogeneity. Moreover, the standardization of HCC subclassification systems across countries using a cost-effective and practically applicable method in daily practice will form a solid foundation for future progress.

MATERIALS AND METHODS

Ethics Declarations

The study in Japan was carried out in accordance with the principles of the 2013 Declaration of Helsinki and was approved by the Ethics Committee of Keio University School of Medicine with approval number 20040034. Informed consent was obtained from each patient.

The research in Indonesia received ethical approval from the Medical Ethics Committee of the Faculty of Medicine, Universitas Indonesia/Dr. Cipto Mangunkusumo Hospital, under the following number: KET-1193/UN2.F1/ETIK/PPM.00.02/2022. Informed consent was obtained from each patient.

Study Design and HCC Tissue Specimens

This study was performed under a memorandum of understanding between two academic institutions, Universitas Indonesia, Republic of Indonesia, and Keio University, Japan (number: 13/MOU/R/UI/2016). The design of the study was implemented according to the Agreement of Implementation between the Faculty of Medicine of Universitas Indonesia and Keio University (number: 125/AOI/FK/UI/2021). Study materials from Indonesia were transferred to Japan upon approval by the Material Transfer Agreement by the Agency of Health Development Policy in Indonesia.

From Japan, a total of 41 HCC tissue specimens were obtained from 37 patients who underwent surgical resection at Keio University Hospital in 2019 and 2020. Hematoxylin and eosin (H&E) staining was carried out on formalin-fixed paraffin-embedded sections (FFPE) of all resected specimens. Metastatic and recurrent tumors were excluded. From Universitas Indonesia, a total of 35 HCCs resected from 35 patients between 2013 and 2020 were prepared as unstained FFPE sections with corresponding H&E-stained slides. For each HCC specimen, a total of 10 unstained slides derived from FFPE blocks were prepared. All materials were sent to Japan and returned to Indonesia as described in the material transfer agreement.

Histopathological Analysis

Histopathological analysis of HCC specimens from Keio University was carried out by at least two liver pathologists based on the fifth edition of the World Health Organization classification of 2019.¹⁶ Histopathological analysis of HCC specimens from Indonesia University was carried out by pathologists from Indonesia. The clinical diagnoses and clinical parameter data were shared between the Indonesian and Japanese teams. Cholangiocarcinoma and combined HCC and cholangiocarcinoma cases were excluded from this study.

Immunohistochemical Analysis

All unstained slides from Japan and Indonesia were stained and analyzed following identical procedures in the department of Pathology, Keio University. Immunohistochemical staining was carried out using a Bond-Max automated staining machine (Leica Microsystems). Single immunohistochemical staining was carried out for the following markers: biliary/stem cell markers cytokeratin 19 (CK19), sal-like protein 4 (SALL4), and epithelial cell adhesion molecule (EpCAM); Wnt/β-catenin signaling-related markers β-catenin and glutamine synthetase (GS); cell proliferation marker Ki-67; and vascular endothelial marker CD34. Multiplex immunohistochemistry (dual staining on the Leica Bond-Max) was carried out for CD3 and CD79α antibodies to simultaneously evaluate the infiltration of T cells and B cells. This technique can also save time and minimize the use of tissue samples. The primary antibodies used in this study are listed in Supplementary Table 1.

The assessment criteria for positivity for each antibody are described as follows. For CK19 (membranous and/or cytoplasmic staining), SALL4 (nuclear staining), EpCAM (membranous staining), β-catenin (nuclear staining), strong staining in greater than or equal to 5% of tumor cells was considered positive. For GS, diffuse strong staining was considered positive, and for Ki-67, a labeling index of more than or equal to 30% was considered high.¹¹ CK19, SALL4, and/or EpCAM-positive cases were defined as B/S subgroup, β-catenin and GS positive cases were defined as W/B subgroup, all-negative cases were defined as −/− subgroup.¹¹ For CD34, the presence of CD34-positive vascular endothelium in a continuous line around a tumor cluster was defined as VETC-positive. The cut-off value for VETC-positive HCC was defined as VETC present in ≥1% of the tumor area.¹⁵ Immune cell infiltration assessment was performed as previously described.¹³ CD3-positive T cells (brown) and CD79α-positive B cells (red) were counted in at least five different high-power magnification fields (HPFs), and the mean values were recorded. Tumor with less than 50 cells/HPF was defined as belonging to the immune-low subtype, whereas tumor with 50 or more infiltrated T cells but not meeting the criteria for the immune-high pattern was defined as immune-mid. The presence of more than 100 T cells/HPF with 10 or more infiltrated B-plasma cells was designated as the immune-high subtype.

The evaluation of staining results was first analyzed independently by K. Effendi and Y. Kurebayashi. All staining results, including cases with subtleties of staining were confirmed together (finalized by Y. Kurebayashi). All cases were evaluated in the same manner and using the same criteria. The evaluation of Indonesian HCC cases was shared and discussed with N. Rahadiani and M. Stephanie.

Image Acquisition

Whole slide images were acquired from all immunostained slides using a NanoZoomer 2.0-HT slide scanner (Hamamatsu Photonics), and images were analyzed using NDP.view2 software (Hamamatsu Photonics). All images of slides from Indonesia were shared between Keio University and Universitas Indonesia.

Statistical Analysis

Associations between categorical variables were analyzed using the chi-square test. Statistical significance was considered to be established when the two-sided P value was less than 0.05 in all analyses. All statistical analyses were performed using IBM SPSS Statistics software version 25 (IBM, Armonk, NY, USA).

RESULTS

Clinicopathological Features of HCC Cases From Japan and Indonesia

The characteristics of HCC patients and tissue samples from Japan and Indonesia are summarized in Table 1 and Supplementary Table 2. Most HCC cases in Indonesia were found in the younger age group (<60 years old: 66%, 23/35), with a mean age 52.5 years, in contrast to the situation in Japan (<60 years old: 8%, 3/37: P < 0.001), with a mean age 71.7 years. The majority of HCC patients in Indonesia were infected with hepatitis B virus (82%, 28/35), compared with only 19% (7/37: P < 0.001) in Japan. Although we had a limited number of cases, a decrease in the trend of hepatitis C virus infections and an increase in NBNC HCC in Japan were already reflected in this study (62%, 23/27). Among NBNC group in Japan, 5 were nonalcoholic steatohepatitis (NASH) (22%, 5/23), 3 were alcohol-related HCC (13%, 3/23), and the etiology of the rest of them was unknown (Supplementary Table 2).

Table 1.

Clinicopathological Features of HCC Cases From Japan and Indonesia.

Clinicopathological features	Japan	Indonesia	P
Age	n = 37	n = 35	< 0.001∗∗
≥60	34	12
<60	3	23
Tumor size	n = 40	n = 35	0.001∗∗
≤5 cm	31	9
>5 cm	9	26
Etiology	n = 37	n = 35	< 0.001∗∗
HBV	7	28
HCV	7	6
NBNC	23	1
Tumor differentiation	n = 41	n = 35	0.341
Early—well	4	3
Moderately	27	18
Poorly	10	14
AFP	n = 31	n = 31	0.005∗
≤500 ng/ml	27	17
>500 ng/ml	4	14
vp/im	n = 41	n = 35	0.072
Positive	25	28
Negative	16	7
IHC-based subgroups	n = 41	n = 35	0.643
B/S	6	3
W/B	7	5
−/−	28	27
Immuno-subtypes	n = 41	n = 34	0.328
Immune-high	2	0
Immune-mid	17	12
Immune-low	22	22
VETC-positive pattern subgroups	n = 41	n = 34	0.030∗
VETC-positive	4	10
Non-VETC	37	24

Abbreviations: –/–, group in which all markers were negative; AFP, α-fetoprotein; B/S, biliary/stem cell markers positive group; HBV, hepatitis B virus; HCC, hepatocellular carcinomas; HCV, hepatitis C virus; IHC, immunohistochemical; NBNC, non-B, and non-C; VETC, vessels encapsulating tumor clusters; vp/im, portal vein invasion and/or intrahepatic metastasis; W/B, Wnt/ß-catenin signaling-related markers positive group.

Most HCC tumors from Indonesia were significantly larger (>5 cm: 74%, 26/35) than those seen in Japanese cases (23%, 9/40: P = 0.001). HCC cases from Indonesia also displayed higher frequencies of poor differentiation (40%, 14/35) and portal vein invasion (80%, 28/35) and higher α-fetoprotein (AFP) levels (45%, 14/31: P = 0.005) compared with Japanese cases. To get more information about tumor aggressiveness, we stained HCC samples from Indonesia with Ki-67 using a 30% cut-off value to determine the proliferation index.¹¹ Despite the higher tumor grades seen in Indonesian HCC cases, highly proliferative tumors were observed in only 7 of 35 (21%) cases, compared with 12 of 41 (29%) in Japanese cases (Supplementary Table 2). Immunohistochemistry of Ki-67 is known to be influenced by differences in the tissue handling of surgical specimens.¹⁷ Delayed or insufficient fixation in the initial tissue slide preparation may have affected the results, particularly for Indonesian cases.

Using a more standardized approach, we could observe that HCC cases from Japan and Indonesia have different dominant viral etiological causes and clinical backgrounds. The clinicopathological features seen in HCC cases from Indonesia are compatible with late-stage diagnosis, whereas many HCC cases from Japan showed early-stage diagnosis, underlining the differences in initial backgrounds.

Evaluation of IHC-based Subgroups of HCCs From Japan and Indonesia

Characterization of HCC into subgroups based on IHC staining which is easily conducted in most clinical settings is helpful. Representative images of the biliary/stem cell marker-positive group (B/S subgroup), defined as CK19, SALL4, and/or EpCAM-positive cases, are shown in Figure 1. In total, 6 of 41 (15%) cases from Japan belonged to the B/S subgroup, and in Indonesia 3 of 35 (9%) cases did (Table 1). All three cases in the Indonesian B/S subgroup had a background of HBV infection, portal vein invasion and/or intrahepatic metastasis, and high AFP levels. The Japanese B/S subgroup had a nonviral background (67%, 4/6), poor differentiation (83%, 5/6), and a high Ki-67 index (83%, 5/6) (Supplementary Table 3).

Figure 1.

Representative images of a CK19-positive, SALL4-positive, and EpCAM-positive HCC case from Indonesia. Positive expressions of CK19, SALL4, and EpCAM are associated with the biliary/stem cell markers group (B/S group). Scale bars = 100 μm.

The Wnt/β-catenin signaling-related marker-positive subclass (W/B subgroup) is defined as being β-catenin and GS positive (Figure 2) and is associated with intermediate aggressiveness; 5 of 35 (14%) Indonesian cases and 7 of 41 (17%) Japanese cases belonged to the W/B subgroup. Similar to the results of a previous study,¹¹ Japanese −/− subgroup cases were associated with lower AFP serum (61%, 17/28), better differentiation (82%, 23/28), and less proliferative activity (75%, 21/28). HCC cases from Indonesia also showed lower AFP serum (52%, 14/27), and less proliferative activity (74%, 20/27) (Supplementary Table 3). The −/− subgroup, which is associated with less aggressive clinicopathological features, was the most common subgroup in both Japanese (68%, 28/41) and Indonesian (77%, 27/35) cases.

Figure 2.

Representative images of a β-catenin-positive and GS-positive HCC case from Indonesia. Positive expressions of β-catenin and GS are associated with the Wnt/β-catenin signaling-related markers group (W/B group). Scale bars = 100 μm.

Although HCC cases from Indonesia exhibited more advanced features, the proportions of B/S, W/B and −/− subgroups in Japanese and Indonesian cases were not significantly different (Table 1). These data suggest that the contribution of biliary/stem cell features the poorer pathological characteristics seen in Indonesian cases is relatively small. Furthermore, the viral etiological factor did not appear to make a major contribution to the poor clinicopathological features associated with the B/S subgroup.

Evaluation of Immuno-subtypes in HCCs From Japan and Indonesia

Investigating the immuno-subtype profiles of HCCs from Japan and Indonesia will give additional information on how the tumor microenvironment is modified in HCCs with different clinical backgrounds.

According to the levels of infiltration of T cells and B-plasma cells, HCC was categorized into immune-high, immune-mid, and immune-low subtypes. Representative double IHC images of CD3-positive T cells (brown) and CD79α-positive B cells (red) are shown in Figure 3. We found two cases of immune-high tumors among Japanese HCC samples (5%, 2/41), but no immune-high tumors were found among Indonesian cases. The immune-low subtype was the most frequently seen immuno-subtype for both Indonesian and Japanese cases; however, the frequency of immune-low subtype tumors among Indonesian cases (65%, 22/34) was slightly higher than that among Japanese cases (54%, 22/41) (Table 1, Supplementary Table 3). Most of the immune-low subtypes from Indonesia had a background of HBV infection (18 of 22, 82%), whereas Japanese cases predominantly had an NBNC background (16 of 22, 73%).

Figure 3.

Representative images of double immunohistochemistry for CD3 (brown) and CD79α (red) from HCC cases from Japan. According to the levels of infiltration of T cells and B-plasma cells, HCC was categorized into immune-high, immune-mid, and immune-low subtypes. Scale bars = 100 μm.

Interestingly, both Japanese and Indonesian immune-low HCCs showed an increased tendency toward portal vein invasion and/or intrahepatic metastasis, i.e., 18 of 22 (82%) Indonesian cases, and 13 of 22 (59%) Japanese cases (Supplementary Table 3). These data suggest that decreased levels of immune infiltration could contribute to the invasive potential of tumor cells, regardless of the underlying etiology or level of tumor differentiation.

Evaluation of the VETC Pattern in HCCs From Japan and Indonesia

VETC is a unique vascular pattern that is distinguished by the presence of CD34-positive vascular endothelium structures that surround and separate tumor clusters as shown in Figure 4.

Figure 4.

Representative images of vessels encapsulating tumor clusters (VETC)-positive HCC cases from Indonesia. The cut-off value for VETC-positive HCC was defined as VETC present in ≥1% of the tumor area. Scale bars = 100 μm.

The frequency of VETC-positive tumors was significantly higher in Indonesian cases (29%, 10/34) than in Japanese cases (10%, 4/41, P = 0.030) (Table 1) as the macrotrabecular pattern was more often seen in Indonesian cases (5/10) than in Japanese cases (1/4). Further observation showed that all VETC-positive HCCs from Japan had portal vein invasion and/or intrahepatic metastasis and belonged to the immune-low subtype. Similar findings were observed in VETC-positive HCCs from Indonesia (90%, 9/10 and 67%, 6/9, respectively) (Table 2).

Table 2.

Clinicopathological Features in Regard With VETC-Positive Pattern of HCC Cases From Japan and Indonesia.

Clinicopathological features	VETC-positive pattern
	Japan	Indonesia
Etiology	n = 4	n = 10
HBV	0	9
HCV	0	1
NBNC	4	0
Tumor size	n = 4	n = 10
≤5 cm	2	3
>5 cm	2	7
AFP	n = 4	n = 10
≤500 ng/ml	3	4
>500 ng/ml	1	6
vp/im	n = 4	n = 10
Positive	4	9
Negative	0	1
IHC-based subgroups	n = 4	n = 10
B/S	0	0
W/B	3	1
−/−	1	9
Immuno-subtypes	n = 4	n = 9
Immune-high	0	0
Immune-mid	0	3
Immune-low	4	6

Abbreviations: -/-, group in which all markers were negative; AFP, α-fetoprotein; B/S, biliary/stem cell markers positive group; HBV, hepatitis B virus; HCC, hepatocellular carcinomas; HCV, hepatitis C virus; IHC, immunohistochemical, NBNC, non-B, and non-C; VETC, Vessels encapsulating tumor clusters; vp/im, portal vein invasion and/or intrahepatic metastasis; W/B, Wnt/ß-catenin signaling-related markers positive group.

Moreover, for both Indonesian and Japanese cases, we observed that none of the VETC-positive HCCs were in the B/S subgroup, suggesting that VETC formation is not directly related to the biliary/stem cell phenotype. Most of the VETC-positive HCCs from Japan had a nonviral background and were mainly found in the W/B subgroup (3 of 4), in accordance with the previous reports indicating that VETC enrichment is seen in HCC groups related with deregulation of the Wnt/β-catenin signaling pathway.¹⁸ In contrast, VETC-positive HCCs from Indonesia had an HBV background and were in the −/− subgroup (9 of 10), with a trend toward higher AFP levels (6/10) and larger tumor sizes (7/10).

Although Japanese and Indonesian HCCs had different underlying backgrounds, the VETC-positive pattern in both countries showed features indicating poorer prognoses, thereby supporting previous reports showing an association between VETC and lower levels of inflammatory infiltrates and frequent microvascular invasion, both of which lead to more unfavorable outcomes.^18,19

DISCUSSION

HCC cases in Japan and Indonesia evidently have different underlying background factors. Recent trends have indicated that the incidence of nonviral etiology (NBNC) HCC is greatly increasing in many countries, including Japan.^3,4 This changing situation in Japan was reflected in our current findings that more than half of the cases were in the NBNC group. In contrast, Indonesia has a high prevalence of HBV infections, and HCC patients are younger than they are in Japan. The average age at onset of HBV-related HCC is reportedly nearly 10 years less than that of HCV-related HCC.²⁰ Our results showed that 66% of Indonesian HCC patients were younger than 60 years old, and this also reflected the current age profile of HCC in Indonesia where the mean age of HCC patients in 2020 was reported as 55 years.⁸ The clinicopathological features of HCC constitute an important criterion that may determine the success of HCC treatment. An expanded criterion called the 5-5-500 rule (nodule size ≤5 cm in diameter, nodule number ≤5, and serum AFP ≤500 ng/ml) has been proposed in Japan for candidates for living-donor liver transplantation (LDLT).²¹ The eligibility requirements for undergoing LDLT are more easily fulfilled in the setting of Japanese HCC patients because they often have early-stage HCC with a smaller tumor size. These findings highlight the importance of conducting screening and surveillance for preventive measures and early detection of HCC as is done in Japan.⁵ With much effort expended toward early detection and regular screening programs, particularly for those with HBV infection, we hope to reach a similar situation in Indonesia in the coming years.

Previously, we showed that HBV-related HCC was more often seen in HCCs expressing biliary/stem cell markers (B/S subgroup), which is associated with poorer prognosis compared with the other subgroups.^11,22 In the current investigation, we expected to see a higher number of B/S subgroup HCCs among Indonesian cases compared with Japanese cases, because most of the Indonesian cases were HBV positive. However, there were no significant differences seen in the proportions of the B/S, W/B, and −/− HCC cases from Japan and Indonesia. Most cases from both countries belonged to the −/− group, which previously had been associated with lower AFP levels, better differentiation, and a decreased frequency of portal vein invasion and/or intrahepatic metastasis.¹¹ However, although 77% of Indonesian HCC cases were in the −/− subgroup, they also had poorer clinical features than Japanese HCC cases. Therefore, we suggest that the features indicating poorer prognosis seen in Indonesian cases mainly resulted from the delay in diagnosis rather than the aggressive phenotype resulting from the acquisition of biliary/stem cell features. Although different underlying etiological factors play roles in the progression of HCC, other factors such as surveillance, comorbidities, and stage at diagnosis are also crucial.²³

The clinical success of immune checkpoint inhibitors against many cancers, including advanced HCC, has created renewed interest in the tumor immune microenvironment.¹² In this study, we found that no Indonesian cases were of the immune-high subtype characterized by increased B-plasma cell infiltration; indeed, there was an increased prevalence of immune-low subtype tumors. Previous studies using a mouse model of liver cancer indicated that mice lacking B cells develop more tumors and larger tumors than wild-type control animals. Accordingly, patients with high B cell densities tend to have smaller tumor sizes with the absence of vascular invasion.²⁴ Consequently, the low frequency of B cell-positive HCC cases from Indonesia may partly explain their larger tumor sizes and the presence of vascular invasion; besides the presence of external factors is also important, e.g., late-stage diagnosis. Furthermore, immune-low subtype cases from both Japan and Indonesia showed an increased frequency of portal vein invasion and/or intrahepatic metastasis, supporting the association of immune-low tumors with poorer prognosis as described in a previous report.¹³

Previous reports have shown an association between VETC-pattern HCC and worse prognosis.^14,19 VETC formation may promote microvascular invasion and create an immunosuppressive tumor microenvironment, which could also explain the increased tendency toward portal vein invasion and/or hepatic metastasis in immune-low HCCs.^14,18 We also found that most VETC-positive HCCs from Japan and Indonesia were Immune-low HCCs, suggesting that VETC positivity could be a prominent pathological finding affecting prognosis, both in early-stage settings as seen in Japanese cases, and in more advanced settings as seen in Indonesian cases.²⁵ Interestingly, VETC-positive HCC cases from Japan mainly belonged to the W/B subgroup, whereas VETC-positive HCCs from Indonesia were mainly in the −/− subgroup and had higher AFP levels and a larger tumor size. VETC-positive HCCs in the W/B subgroup has been associated with increased expression of angiogenic factors such as fibroblast growth factor 2 which might provide an environment suitable for the formation of VETC.¹⁸ On the other hand, VETC-positive HCCs in the non-W/B subgroup as commonly seen in Indonesian cases, were mostly associated with HBV infection and may reflect a different mechanism in VETC enrichment, such as the p53 phenotype and the macrotrabecular massive subtype.¹⁹ Further investigations to assess the mechanisms underlying VETC formation in different subgroups would help to predict patient prognosis and facilitate the optimum treatment, such as combination of immunotherapy and angiogenic therapy.

The current study had some limitations. First, because of the mostly advanced-stage presentation in Indonesia, we could only obtain few samples from resectable HCC cases, and we could not completely deny the factor of late-stage presentation in the features of our subtypes’ differences. Second, due to differences in backgrounds, timing of diagnosis, and availability of patient follow-up, it was difficult to collect overall survival data. Third, the initial preparation of FFPE tissue slides was done separately in each country, and the variable methods of collection, processing, including shipping the slides may have affected the results of IHC staining for some markers.²⁶ Future studies using larger sample sizes and the inclusion of overall survival data analysis should be done to obtain a clearer overview and significant results. Furthermore, standardization and proper tissue handling across laboratories or institutions will help to improve the quality of IHC staining.

In conclusion, the similarities and differences in findings between HCC cases from Japan and Indonesia underline the heterogeneous features of HCC; these features also present a challenge to the establishment of a unified HCC subclassification method across countries. Nevertheless, our study showed that immunohistochemistry-based HCC subclassification, which is relatively easy to conduct in daily clinical settings, could be used to reflect HCC features more precisely despite the HCCs having different clinical backgrounds. The consistent assessment of HCC subtypes with various backgrounds from the pathological perspective would be helpful in the global management of HCC.

Credit authorship contribution statement

Kathryn Effendi conducted the study and wrote the manuscript. Nur Rahadiani and Marini Stephanie provided and interpreted HCC slides from Indonesia. Olivia Maurine Jasirwan and Ridho Ardhi Syaiful provided clinical data information for HCC samples from Indonesia. Yutaka Kurebayashi and Hanako Tsujikawa interpreted the findings. Michiie Sakamoto reviewed the manuscript. All authors approved the final manuscript.

Conflicts of interest

The authors have none to declare.

Acknowledgements

The authors would like to thank Dr. Ken Yamazaki and Dr. Kosuke Matsuda for their help in providing statistical and staining analysis. Thanks also go to all members in Yonken in the Department of Pathology, Keio University School of Medicine, Japan, for their help in technical assistance and to Mrs. Kayoko Takashima for assisting in grant funding documentation.

Funding

This study was partially supported by the Keio University Medical Science Fund – The Mitsunada Sakaguchi Grant for International Collaborative Research.

Appendix A. Supplementary data

The following is the Supplementary data to this article: