Association of Interleukin-28B Genotype and Hepatocellular Carcinoma Recurrence in Patients with Chronic Hepatitis C

Hepatocellular carcinoma (HCC) is the seventh most common cancer worldwide and the third most common cause of cancer mortality (1). HCC usually develops in patients suffering from chronic hepatitis B or chronic hepatitis C (CHC). Although hepatic resection has been considered the most efficient therapy for HCC, it is only suitable for 20% to 35% of patients because of poor hepatic reserve (2). Radiofrequency ablation (RFA) has therefore been introduced as a minimally invasive therapy for such cirrhotic patients and is widely applicable with little effect on hepatic reserve. Moreover, randomized (3, 4) and nonrandomized (5, 6) controlled studies revealed no statistical difference in patient survival between resection and RFA.

Despite these curative treatments of HCC, its recurrence remains common. Several studies have identified potential risk factors for HCC recurrence, including the presence of cirrhosis, high α-fetoprotein (AFP) levels, large tumor foci, and tumor multiplicity (7, 8).

The interleukin-28B (IL-28B) gene, also known as IFN-λ3, is a newly described member of the family of IFN-related cytokines (9) and shares the same biologic properties as type I IFNs (10). Recently, several single-nucleotide polymorphisms (SNP) in the IL-28B locus have been associated with the effectiveness of pegylated-IFN and ribavirin combination therapy for CHC (11, 12). We previously confirmed this relationship and revealed that the IL-28B genotype is associated with the expression of hepatic IFN-stimulated genes (ISG) in patients with CHC (13). Others have also described an association between the IL-28B genotype and the outcome of CHC therapy, biochemical factors, and histologic findings (14, 15); however, the relationship between the IL-28B genotype and the biology and clinical course of HCC remains unknown. In this study, therefore, we investigated the association between the IL-28B genotype and clinical outcome after initial curative treatment of HCC and clarified the molecular features in relation to the IL-28B genotype.

A total of 852 patients were admitted to the Department of Gastroenterology, Kanazawa University Hospital, Kanazawa, Japan between January 2000 and March 2012 for the treatment of developed HCC. The major background liver disease was hepatitis C virus (HCV; n = 502), followed by hepatitis B virus (n = 148). Treatment of HCC included surgical resection in 175 patients and RFA in 390 patients. The choice of treatment procedure was determined according to the extent of the tumor and the hepatic functional reserve as assessed by Child's classification that forms the Japanese HCC Guidelines (16, 17). In some cases indicated for surgical resection, we conducted RFA on patients who refused surgical resection, and we consequently excluded these patients on the basis of Japanese HCC guidelines.

Study inclusion criteria were: (i) Child–Pugh class A or B; (ii) the presence of up to 3 tumors, each 3 cm or less; (iii) HCV infection (positive for HCV RNA, patients with sustained viral response were excluded); (iv) radical treatment by either surgical resection or RFA; and (v) availability of blood samples for genetic analyses (Supplementary Fig. S1). Consequently, 183 patients were studied and their baseline characteristics are reported in Table 1. Informed consent was obtained from all patients before therapy. The experimental protocol was approved by the Human Genome, Gene Analysis Research Ethics Committee of Kanazawa University (Approval No. 260), and the study was conducted in accordance with the Declaration of Helsinki.

HCC diagnosis was based predominantly on image analysis. Patients underwent dynamic computed tomography (CT) and/or dynamic MRI and abdominal angiography with CT imaging in the arterial and portal flow phase. HCC was diagnosed if a liver nodule showed hyperattenuation in the arterial phase and washout in the portal or delayed phase or showed typical hypervascular staining on digital subtraction angiography (18).

Hepatic resection was carried out under intraoperative ultrasonographic monitoring and guidance. Anatomic resection was conducted in 9 patients and nonanatomic resection was conducted in 20 patients. Curative resection was defined as removal of all recognizable tumors with a clear margin (19). RFA was conducted using either the radiofrequency interstitial tumor ablation system (RITA; RITA Medical Systems Inc.) or the cool-tip system (Tyco Healthcare Group LP). All procedures were conducted according to the manufacturer's protocol. In the case of RFA, dynamic CT was conducted 1 to 3 days after therapy and the ablated area was evaluated. Complete ablation was defined as no enhancement in the ablated area on the dynamic CT. When complete ablation was not achieved, additional ablation was considered.

All patients were followed up by ultrasound and contrast enhancement 3-phase CT or MRI every 3 months. Local tumor progression was defined as the reappearance of tumor progression adjacent to the treated site and distant recurrence as the emergence of one or several tumor(s) not adjacent to the treated site. Patients with confirmed recurrence received further treatment such as resection, RFA, transarterial chemoembolization, and chemotherapy depending on the condition. Time to recurrence (TTR) was defined as the period from the date of therapy until the detection of tumor recurrence, death, or the last follow-up assessment. For TTR analysis, the data were censored for patients without signs of recurrence.

Genomic DNA was extracted from peripheral blood samples using the QIAamp DNA Blood Mini Kit (Qiagen) according to the manufacturer's instructions. An IL-28B SNP (rs8099917) was determined using TaqMan Pre-Designed SNP Genotyping Assays as described previously (12). A custom assay was created by Applied Biosystems for rs12979860. We determined IL-28B genetic variations in all patients included in this study.

Resected cancer and noncancerous liver tissue specimens were immediately frozen in liquid nitrogen and kept at −80°C until required for RNA preparation. Liver tissue RNA was isolated using the RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. Isolated RNA was stored at −70°C until required. The quality of isolated RNA was estimated after electrophoresis using an Agilent 2001 Bioanalyzer. Microarray analysis using an Affymetrix Human 133 Plus 2.0 microarray chip was conducted as described previously (13). The microarray data have been submitted to the Gene Expression Omnibus (GEO) public database at National Center for Biotechnology Information (NCBI, Bethesda, MD; accession number GSE41804).

Affymetrix GeneChip array data were normalized, preprocessed, and analyzed using R (20) and Bioconductor (21) software. Raw CEL file data were normalized using the MAS 5.0 algorithm as implemented in the affy package. Normalized data were log₂ transformed and assessed using gene set enrichment analysis (GSEA), which is a bioinformatics method to assess whether genes with known biological/molecular function are concomitantly upregulated or downregulated in a certain gene expression dataset (22). GSEA was conducted using a parametric analysis of gene set enrichment (PAGE; ref. 23). The Gene Ontology gene set collection C5 of the Molecular Signatures Database (22) was downloaded from the Broad Institute and loaded into the R environment.

We also investigated the gene set differentially expressed HCC-infiltrating mononuclear inflammatory cells studied previously (24). Z scores and P values of all gene sets were calculated using the PGSEA package and an estimate was made as to whether certain gene sets, and therefore functional gene categories, were differentially regulated in HCC tissue from patients with the IL-28B TT genotype and the IL-28B TG/GG genotype.

Hierarchical clustering was conducted with Cluster software using Pearson's correlation distance metric and average linkage followed by visualization in Treeview software.

Noncancerous liver tissue that had been surgically resected from patients with HCC and liver specimens obtained by needle biopsy from the background liver of patients with HCC were fixed in 10% buffered formalin and embedded in paraffin. Each paraffin-embedded specimen was sliced into 3 to 4 μm sections and stained with hematoxylin and eosin. Each specimen was semiquantitatively analyzed by assigning a score according to each of the following features: (i) severity of inflammatory cell infiltration (0 for none, 1 for minimal, 2 for mild, 3 for moderate, and 4 for severe) in the periportal, intralobular, and portal areas; (ii) the severity of the F stage of fibrosis (0 for F0, 1 for F1, 2 for F2, 3 for F3, and 4 for F4; ref. 25); the degree of lymphoid aggregates in the portal area (0 for none, 1 for mild, 2 for scattered, 3 for clustered, 4 for lymph follicle without germinal center, and 5 for lymph follicle with germinal center); the severity of portal sclerotic change, perivenular fibrosis, and pericellular fibrosis (on a scale of 0–4 with 0 for none to 4 for severe); the severity of damage to the bile duct (on a scale of 0–4 with 0 for none to 4 for disappearance); the existence of bridging necrosis (0 for none and 1 for existence); the severity of irregular regeneration of hepatocytes as described previously (on a scale of 0–4 with 0 for none to 4 for severe; ref. 26); the grade of steatosis (on a scale of 0–4 with 0 for none to 4 for severe). The irregular regeneration score was based on the findings of a map-like distribution, anisocytosis, and pleomorphism of the hepatocytes, bulging of the regenerated hepatocytes and proliferation of atypical hepatocytes and oncocytes.

Paraffin-embedded specimens were sliced into 3 to 4 μm sections, deparaffinized, and subjected to heat-induced epitope retrieval at 98°C for 40 minutes. After blocking endogenous peroxidase activity using 3% hydrogen peroxide, the slide was incubated with appropriately diluted primary antibodies. Antihuman CD4, antihuman CD8 and antihuman CD14 mouse monoclonal antibodies were used to evaluate the immunoreactivity of HCC using a DAKO EnVision+TM kit, as described in the manufacturer's instructions.

We semiquantitatively analyzed tumor tissues by assigning a score to the severity of CD4-positive and CD8-positive lymphocyte infiltration in the tumor tissue (0 for none, 1 for mild, 2 for moderate, and 3 for severe).

Fisher exact probability test was used to compare categorical variables and the Mann–Whitney U test was used to compare continuous variables; a P value of less than 0.05 was considered statistically significant. The TTR survival curve was analyzed using the Kaplan–Meier curve and compared by the log-rank test. Univariate Cox regression analysis was conducted to identify TTR predictors out of clinical and biologic parameters [sex, age, IL-28B genotype, therapy, platelet count, alanine aminotransferase (ALT), γ-GTP, albumin, prothrombin activity, bilirubin, Child–Pugh class, history of IFN therapy, AFP, and des-γ-carboxy prothrombin (DCP)] and tumor factors (size and number).

Multivariate analysis was conducted using the Cox regression model with backward elimination (27). The significance level for removing a factor from the model was set to 0.05. A bootstrap technique was applied to confirm the choice of variables (27). One thousand bootstrap samples were generated using resampling with replacement and Cox regression analysis with backward elimination was applied to each sample. The percentage of samples (from the total of 1,000) for which each variable was included in the model was calculated. In multivariate analysis, we evaluated two models that contained either Child–Pugh class or its components to avoid multicollinearity. Data analysis was conducted with R software. We used functions from the Regression Modeling Strategies library for validation with the bootstrap technique (28).

We genotyped 183 patients with HCC for the IL-28B rs8099917 TT, TG, and GG genotypes and observed respective proportions of 67.8% (124 of 183), 30.6% (56 of 183), and 1.6% (3 of 183), which is a similar distribution to that observed in several Japanese studies of patients with CHC (13, 14, 29, 30). Although the prevalence of the TG/GG genotype was higher than that of the general population (12%–16%; refs. 12, 31, 32), there was no significant difference between our result and that of HCV-infected patients in a previous study. There was also no significant difference in clinical variables between the TT and TG/GG genotypes (Table 1).

We next genotyped 160 of 183 cases for rs12979860 and our findings were largely in concordance with those of rs8099917, with the exception of 1 case (0.6%). The haplotype of the case showed that rs8099917 was TT and rs12979860 was CT suggesting that these 2 loci are in a haplotype block with a high level of linkage disequilibrium, as previously reported (13, 30). Genotype distribution analysis showed that rs8099917 was in Hardy–Weinberg equilibrium, so we selected it for further investigation.

During the median follow-up period of 2.5 years (range, 0.3–7.2 years), 118 of 183 patients developed HCC recurrence. Local tumor progression was seen in 13 patients treated by RFA and in only 1 patient treated by resection. The local tumor progression rate and distant recurrence rate were 2.6% and 21.2% in the first year and 8.3% and 54.2% within 2.5 years, respectively. These results are comparable with previous reports by others (33, 34). The type of recurrence was also comparable between IL-28B genotype groups.

HCC TTR was also analyzed using multivariate Cox regression analysis using 15 clinical parameters and the IL-28B genotype. With a significance level of 0.05 for removing a variable in a Cox regression with backward elimination, the IL-28B genotype was selected as the final model (Table 2). To confirm this decision, a bootstrapping technique was applied. The percentages of inclusion among the 1,000 samples created by the bootstrapping technique for variables are shown in Table 2. The percentage of inclusion for the IL-28B genotype was 80.4%. Frequencies of another variable were lower than 40%. The bootstrap procedure result confirmed the variables chosen for the final model.

In univariate Cox regression analyses, the IL-28B genotype was associated with HCC recurrence (Table 2). The TTR survival curve was analyzed using the Kaplan–Meier curve and log-rank test (Fig. 1), and patients with the IL-28B TT genotype showed a significantly shorter median TTR (1.61 years) than those with the IL-28B TG/GG genotype (2.58 years; P = 0.007).

To clarify the molecular mechanism influencing HCC recurrence, we histologically analyzed 141 noncancerous liver tissues according to previously published criteria (Table 3; ref. 26). The mean score of the degree of inflammatory cell infiltration in the periportal area was significantly higher in TT genotype patients (2.804) than TG/GG genotype patients (2.513; P = 0.025); the degree of inflammatory cell infiltration in the intralobular area was also higher in the TT genotype (2.522) than the TG/GG genotype (2.308), although this did not reach statistical significance (P = 0.08). Furthermore, the mean score of the degree of hepatocyte anisocytosis was significantly higher in the TT genotype (1.891) than the TG/GG genotype (1.385; P = 0.024). Anisocytosis is characterized by viability of cell size with focal dysplastic change and indicates irregular regeneration of hepatocytes. The irregular regeneration score was higher in the TT genotype (2.207) than the TG/GG genotype (1.795), albeit not significantly (P = 0.105).

We next compared the gene expression profile of HCC tissues and noncancerous liver tissues of both the IL-28B TT and IL-28B TG/GG genotype. Ten patients with HCC were selected from each IL-28B genotype and their gene expression was determined using Affymetrix genechip analysis (Supplementary Table S1). We recently reported that expression of hepatic ISGs is downregulated in individuals with the IL-28B TT genotype, whereas the expression of other immune response-related genes was shown to be upregulated (13). Therefore, to validate our expression data, we compared the expression of ISGs and other immune response-related genes in the present study with that of the previous study. We analyzed the expression data of 20 patients from the current study in addition to another series of 91 patients with CHC from our previous study.

One-way hierarchical clustering using 28 representative ISGs showed that patients with the IL-28B TG/GG genotype had higher expression of hepatic ISGs, whereas patients with the TT genotype showed lower expression of hepatic ISGs in CHC tissues and noncancerous background liver tissue, confirming our previous data (Fig. 2A and Supplementary Table S2). Expression of hepatic ISGs in HCC tissues was lower than in background liver tissues, with no relationship to the IL-28B genotype. Hierarchical clustering of 51 representative immune response-related genes from the Gene Ontology gene set of the Molecular Signatures Database indicated that their expression was upregulated in TT genotype compared with TG/GG genotype tissues, with the exception of HCC tissues (Fig. 2B and Supplementary Table S2). Upregulation of immune response-related genes suggests that hepatic inflammation is more severe in TT genotype patients, which is consistent with our histologic findings and recent studies that reported an association between high serum ALT levels and the IL-28B TT genotype (14, 29).

We applied PAGE to identify gene sets differentially regulated between the different IL-28B genotypes from the whole gene expression profiles derived from HCC tissues. Analysis of groups of genes involved in a specific function enables significant differences to represent a biologically meaningful result (23). Many gene sets associated with the immune system (e.g., the immune system process, T-cell activation, regulation of T-cell activation, and T-cell proliferation) showed a significant increase in their expression in patients with HCC with the IL-28B TG/GG genotype (Supplementary Table S3). This PAGE profile was consistent with the hierarchical clustering of 51 immune response-related genes (Fig. 2B) and suggests that the immune response to tumors might be more intensive in IL-28B TG/GG genotype HCC than IL-28B TT genotype HCC.

To verify our PAGE profile, we histologically compared HCC tissue of 20 cases of the IL-28B TT genotype and 12 cases of the TG/GG genotype using immunohistochemical staining with antibodies against helper T cells (CD4) and cytotoxic T cells (CD8). The mean score of the degree of CD8⁺ lymphocyte infiltration in the tumor tissue was significantly higher in the TG/GG genotype (1.75) than the TT genotype (1.175; P = 0.047; Supplementary Table S4). A representative case is shown in Fig. 3. There was no morphologic alteration associated with the IL-28B genotype. Immunohistochemical analysis showed intratumoral infiltration of CD4⁺ and CD8⁺ lymphoid cells and slight infiltration of monocytes/macrophages in HCC of the IL-28B TG/GG genotype, compared with little infiltration of lymphocytes or monocytes/macrophages in HCC of the IL-28B TT genotype.

Furthermore, the gene set differentially expressed in HCC-infiltrating mononuclear inflammatory cells from our previous study (24) was upregulated in HCC of the IL-28B TG/GG genotype (Z score, −9.879; P < 0.001). One-way hierarchical clustering was carried out of 122 genes involved in the gene set differentially expressed in HCC-infiltrating mononuclear inflammatory cells. Most of the 122 genes were expressed at high levels in many HCC tissues of IL-28B TG/GG genotype patients (Supplementary Fig. S2).

IL-28B is a recently identified, novel IFN-λ family member that shares the same biologic properties as type I IFNs (9). Recent reports have shown a significant association between IL-28B allelic variants and treatment outcome in CHC (11, 12). IL-28B genotyping is therefore considered to be a suitable pretreatment predictor of treatment response for individual patients, and also an indicator of biochemical and histologic findings in patients infected with HCV (14). In this study, we determined that the IL-28B genotype is associated with HCC recurrence in patients with CHC as patients with the IL-28B TT genotype showed a significantly higher incidence of recurrence than those with the IL-28B TG/GG genotype after curative therapy. To our knowledge, this is the first study to reveal an association between the IL-28B genotype and HCC recurrence and molecular features in patients with CHC.

To date, there are several contradicting results about the association of the IL-28B genotype and progression of liver disease including the development of HCC. Fabris and colleagues and Eurich and colleagues reported that patients with a T allele in rs12979860 (G allele in rs8099917) were at a high risk of progressing to liver cirrhosis and HCC (35, 36), however, these reports have not yet been confirmed by others. A large-scale European genome-wide association study (GWAS) recently identified a weak protective role for the rs12979860 T allele in the progression of fibrosis during HCV infection (37), whereas a Japanese GWAS identifying a susceptibility locus for HCV-induced HCC found no association of rs12979860 and rs8099917 SNPs with HCC (38). In support of these findings, Joshita and colleagues reported no association between the IL-28B genotype and the incidence of primary HCC (39). These results show a good concordance with those of the present study, which revealed that the IL-28B genotype was not associated with HCC incidence before treatment (Table 1). Furthermore, closer histologic assessment showed a high score of periportal inflammation and pericellular fibrosis in the rs8099917 TT genotype (CC in rs12979860). This suggests that our patient selection process was not biased, and that our results are in agreement with the Japanese study and are comparable with the European study.

To date, the reasons for contradicting results about the association of the IL-28B genotype and progression of liver disease have not been clear, however, clinical bias such as patient number, history of treatment, virus genotype, and titer and racial differences may affect the results. It should be noted that significant differences in genotype frequencies with respect to ethnic/racial groups have previously been reported for IL-28B SNPs (11). To overcome these limitations, a future cross-sectional prospective study should be conducted.

Several risk factors for HCC recurrence have previously been reported, including the presence of cirrhosis, high AFP levels, and multiplicity of tumors 7, 8). However, multivariate analysis and the bootstrap procedure of the present study revealed that the IL-28B genotype was independent indicators for recurrence, suggesting that it is stronger predictors of HCC recurrence than other factors.

The expression of hepatic ISGs was higher in IL-28B TG/GG genotype patients than IL-28B TT genotype patients with CHC in this study. This confirms our previous findings in a different cohort and those of another research group (13, 40). Several ISGs have been reported to have antiproliferative and proapoptotic functions (41, 42), and IFN-α (type I IFN) has also been found to inhibit metastasis and human HCC recurrence after curative resection mediated by angiogenesis (43). Indeed, IL-28B rs8099917 is associated with early HCC recurrence (<1 year), possibly because of the intrahepatic metastasis of HCC in this study (Fig. 1 and Supplementary Table S5). These reports and our findings suggest that high expression of hepatic ISGs might cause the low HCC recurrence in the IL-28B TG/GG genotype, although the mechanism of this association remains unknown.

Microarray, histologic, and immunohistochemical analysis in the present study showed that the immune response was more severe in chronic hepatitis and noncancerous tissue of IL-28B TT genotype compared with TG/GG genotype patients. Serum ALT levels were also higher in the IL-28B TT genotype, albeit not significantly. These results support previous findings that showed higher serum ALT levels and more severe liver inflammation in TT genotype compared with TG/GG genotype patients with HCC (14, 29). Irregular regeneration of hepatocytes develops as a result of repeated cycles of necrosis and regeneration of hepatocytes and was previously reported to be an important predictive factor for the development of HCC (26). We histologically showed that the degree of hepatocyte anisocytosis was more severe in noncancerous livers of TT genotype than TG/GG genotype patients, perhaps because of IL-28B genotype-dependent hepatic inflammation. This might also affect the late recurrence of HCC (>1 year) as a result of the multicentric occurrence of HCC in background liver disease. In the late recurrence group, IL-28B TT genotype patients showed a shorter TTR than IL-28B TG/GG genotype patients although this did not reach statistical significance (P = 0.086; Supplementary Fig. S3; Supplementary Table S6).

Previous studies showed that the gene expression profile of noncancerous liver tissue was associated with late recurrence HCC and the multicentric occurrence of HCC (44). However, the gene set expression of these studies did not differ between the IL-28B TT and TG/GG genotypes in the present study. Although the reason for this discrepancy is unclear, the IL-28B genotype may affect early recurrence more than late recurrence, and the limited number of patients and the short follow-up period may affect statistical comparisons. Therefore, further investigations with a large series of patients are necessary to clarify whether IL-28B genotype-dependent inflammation influences HCC recurrence.

On the other hand, the gene expression profile and histologic analyses showed that more lymphocytes infiltrate into the tumor tissue of the IL-28B TG/GG genotype than the TT genotype. Chew and colleagues previously showed that 14 intratumoral immune gene signatures were able to identify molecular cues driving the tumor infiltration of lymphocytes and predict the survival of patients with HCC, particularly during the early stages of disease (45). We can confirm that the expression of some of these 14 genes was higher in TG/GG genotype than TT genotype patients (Supplementary Fig. S4), supporting the association of the IL-28B genotype, HCC recurrence, and histologic findings. The presence of lymphocyte infiltration in HCC was also reported as a negative predictor of HCC recurrence after liver transplantation (46), and this phenomenon may contribute to a lower incidence of HCC recurrence in the TG/GG genotype.

It may seem contradictory that the immune response in noncancerous liver was more severe in TT genotype than TG/GG genotype patients despite the fact that the expression of immune genes was higher in tumor tissue and more lymphocytes infiltrated the tumor in the TG/GG genotype compared with the TT genotype. Although we are unable to explain this contradiction, it is conceivable that the host immune reaction has a differential role between tumor and nontumor tissue. Moreover, HCV-specific T-cell immune responses, which are essential for disease control, are attenuated in patients with CHC, and T-cell exhaustion has recently been implicated in the deficient control of chronic viral infections. On the other hand, little is known on self- and tumor-specific T-cell responses in patients with HCC. While several reports have shown the existence of exhausted T cells in a tumor environment, impaired T-cell responses to tumors are unlikely to be simply explained by T-cell exhaustion (47).

Anergy or other functional statuses such as suppressive immunity by tumor cells should be considered in tumor immunity. Therefore, differences in immunity to viral antigens and self- and tumor-antigens could explain our findings, although further work should be carried out to confirm these conclusions. We have preliminarily confirmed that the ratio of regulatory T cells is higher in the peripheral blood of IL-28B TT genotype HCC patients than IL-28B TG/GG genotype patients, although there is no significant difference between non-HCC IL-28B TT genotype and IL-28B TG/GG genotype patients (data not shown). Although the cause of this phenomenon is unclear, our gene expression profile of noncancerous liver and tumor tissues suggests paradoxical roles for the immune response in CHC and HCC depending on IL-28B genotype; it will be necessary to clarify these mechanisms in future investigations.

Recently, a sustained virologic response (SVR) to CHC antiviral treatment was shown to be associated with a lower risk of HCC recurrence (48). Although we did not include patients with SVR in the current study, we nevertheless observed that they showed a longer recurrence-free survival than patients infected with HCV, independent of IL-28B genotype (data not shown). This result together with the association between the IL-28B genotype and response to antiviral treatment promotes recommendations for aggressive CHC antiviral treatment, especially in cases with the IL-28B TT genotype.

RFA is a recently developed technique and its efficacy has been reported equal to that of surgical resection, especially in early-stage HCC (3–6). In the European Association for the Study of the Liver–European Organisation for Research and Treatment of Cancer (EASL-EORTC) guidelines, RFA is considered the standard care for patients with Barcelona Clinic Liver Cancer stage 0-A tumors not suitable for surgery and whether or not RFA can be considered a competitive alternative to resection is uncertain (49). In our study, the local tumor progression rate was not statistically different between RFA and resection cases. However, further studies with an appropriate sample population are necessary to clarify the comparison of RFA and resection. The present study has some limitations. It was a retrospective cohort and a single-center study, so it was difficult to completely eliminate bias. Further prospective studies of a larger series of patients should be conducted to validate our results. As a consequence of the small sample size and even smaller number of patients undergoing surgical resection, we could not show an association between IL-28B genotype and HCC recurrence in the surgical resection group (data not shown). However, we did find no significant difference in TTR between RFA and surgical resection, confirming previous findings.

In conclusion, we found that the IL-28B rs8099917 TT genotype is associated with shorter TTR in patients with HCC with CHC. Microarray analysis showed a high expression of ISGs in background liver and high expression of immune system-related genes in tumor tissues of the IL-28B TG/GG genotype. Histologic findings also showed that more lymphocytes infiltrated into tumor tissues in the TG/GG genotype. The IL-28B genotype therefore is a candidate useful genetic marker to predict HCC recurrence as well as the response to pegylated-IFN and ribavirin combination therapy for CHC.

No potential conflicts of interest were disclosed.

Conception and design: M. Honda, Y. Sakai, M. Moriyama, S. Kaneko

Development of methodology: M. Moriyama

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Hodo, A. Tanaka, K. Arai, T. Yamashita, T. Yamashita, A. Sakai, Y. Nakanuma, M. Moriyama, S. Kaneko

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Hodo, M. Honda, M. Sasaki, M. Moriyama

Writing, review, and/or revision of the manuscript: Y. Hodo, M. Honda, M. Moriyama

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Hodo, A. Tanaka, Y. Nomura, K. Arai, E. Mizukoshi, M. Sasaki, M. Moriyama, S. Kaneko

Study supervision: Y. Sakai, E. Mizukoshi, M. Moriyama

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.