Expression of eukaryotic translation initiation factor 3 subunit B in liver cancer and its prognostic significance.

Liver cancer is one of the major malignancies with the worst prognosis among all solid tumor types. It is therefore ponderable to explore prognostic biomarkers and therapeutic targets for liver cancer. Eukaryotic translation initiation factor 3 subunit B (EIF3B) is closely linked to the transcription initiation of cancer-associated genes. In the present study, EIF3B was indicated to be a potential prognostic biomarker of liver cancer. The mRNA expression level of EIF3B in liver cancer was assessed by analyzing the Cancer Genome Atlas dataset. χ2 and Fisher's exact tests were used to assess the association of EIF3B expression with clinical parameters. Receiver-operating characteristic curve analysis was used for evaluating the diagnostic value of EIF3B. Overall and relapse-free survival were assessed using Kaplan-Meier curves to determine the association between EIF3B expression and survival. Univariate and multivariate Cox regression analysis were performed to identify the factors affecting overall/relapse-free survival. Gene set enrichment analysis (GSEA) was used to identify signaling pathways associated with EIF3B in liver cancer. It was revealed that EIF3B was highly expressed in liver cancer tissues and it had a promising diagnostic ability. Furthermore, the survival analysis indicated that patients with high EIF3B expression generally had shorter overall as well as relapse-free survival. Univariate and multivariate Cox analysis suggested that high EIF3B mRNA expression may serve as an independent biomarker for the prognostication of patients with liver cancer. GSEA suggested that MYC-V1 (HALLMARK_MYC_TARGETS_V1 geneset; P=0.009), MYC-V2 (HALLMARK_MYC_TARGETS_V2 geneset; P=0.004) and DNA repair pathways (HALLMARK_DNA_REPAIR geneset; P<0.001) were differentially enriched in high EIF3B expression and low EIF3B expression groups. In conclusion, high EIF3B expression was indicated to be an independent prognostic biomarker for patients with liver cancer. Copyright: © Yue et al.

Introduction

Liver cancer is a common malignant tumor type with high morbidity and mortality (1). Although various treatments have been improved, the mortality rate of liver cancer is still high and the prognosis remains poor (2,3). Therefore, prognostic biomarkers of liver cancer have become one of the hotspots of current research (4). The discovery of accurate prognostic biomarkers may contribute to clinical guidance in order to improve the evaluation system of liver cancer.

The family of eukaryotic translation initiation factors (EIFs) participates in eukaryotic translation by regulating the interaction between ribosomes and RNA. It is the rate-limiting step of protein synthesis and participates in numerous processes that are deregulated in cancer cells, including DNA repair and proliferation, cell cycle and apoptosis (5). EIF 3 subunit B (EIF3B) is an important member of the family of EIFs and has been observed to be overexpressed in numerous cancer types, including clear cell renal cell carcinoma (6), esophageal squamous cell carcinoma (7), glioblastoma (8), ovarian cancer (9), osteosarcoma (10) and lung cancer (11), and has an important role in the progression and prognosis of several cancer types (12-14). In addition, Golob-Schwarzl et al (15) reported that EIF3B was upregulated in hepatitis C virus (HCV)-associated hepatocellular carcinoma (HCC). However, the precise role of EIF3B in liver cancer has remained elusive.

To further evaluate the roles of EIF3B in patients with liver cancer, the expression of EIF3B was examined in a dataset from The Cancer Genome Atlas (TCGA) database. The χ2 and Fisher's exact tests were used to assess the association of EIF3B with clinicopathological parameters and demographic features. Receiver operating characteristics (ROC) curve analysis was used for evaluating the diagnostic value of EIF3B. Kaplan-Meier overall survival and relapse-free survival analysis were performed to determine the association between EIF3B expression and survival. Univariate and multivariate Cox regression analysis were performed to identify the factors affecting overall survival and relapse-free survival. Furthermore, gene set enrichment analysis (GSEA) was used to explore EIF3B-associated signaling pathways.

Materials and methods

Data source

The clinical information of patients and their RNAseq data were obtained from TCGA (https://cancergenome.nih.gov/). All patients from the liver hepatocellular carcinoma (LIHC) cohort were screened based on TCGA inclusion and exclusion pre-selection criteria.

Statistical analysis

R (version 3.6.1; The R Foundation) (16) was used for statistical analysis (t-test, Kruskal-Wallis with Dunn's post-hoc test, Wilcoxon sum-rank test) and generation of images. The ggplot2 package (17) was used to draw boxplots of the EIF3B expression in subgroups by clinical characteristics. χ2 and Fisher's exact tests were applied to estimate the significance of the association between EIF3B expression and clinicopathological or demographic characteristics. The pROC package (18) was used to plot the ROC curves and assess the diagnostic ability of EIF3B, and patients were divided into a high expression group and low expression group according to the best operating system cut-off value determined by the Youden index. The survival package (19) was used to draw survival curves. A univariate Cox linear regression model was utilized to select correlative variables affecting survival time. Multivariate Cox regression analysis was employed to evaluate the independent influencing factors of survival time.

GSEA

GSEA may be used to determine whether a predefined set of genes is able to indicate significant, consistent differences between two biological states (20). In the present study, GSEA was performed in the ‘h.all.v6.2.symbols.gmt’ and ‘c2. cp.biocarta.v6.2.symbols.gmt’ gene sets using GSEA3.0 software. The standardized enrichment fraction was obtained by 1,000 permutation analyses.

Results

Patient characteristics

In Table I, the clinical features of the 373 patient cohort, including sex (female 121, male 252), age (16-90 years old, median 61, mean 59.47), use of radiation therapy, residual tumor, relapse-free survival, histological type, stage, vital status, survival data, T/N/M classification and EIF3B expression are provided.

Table I

Demographic and clinical characteristics of the cohort from The Cancer Genome Atlas-liver hepatocellular carcinoma dataset.

Characteristics	N (373)
Age (years)
<55	117 (31.45)
≥55	255 (68.55)
NA	1 (0.00)
Sex
Female	121 (32.44)
Male	252 (67.56)
Histological type
Fibrolamellar carcinoma	3 (0.80)
Hepatocellular carcinoma	363 (97.32)
Hepatocholangiocarcinoma (mixed)	7 (1.88)
Histologic grade
NA	5 (1.34)
G1	55 (14.75)
G2	178 (47.72)
G3	123 (32.98)
G4	12 (3.22)
Stage
NA	24 (6.43)
I	172 (46.11)
II	87 (23.32)
III	85 (22.79)
IV	5 (1.34)
T classification
NA	2 (0.54)
T1	182 (48.79)
T2	95 (25.47)
T3	80 (21.45)
T4	13 (3.49)
TX	1 (0.27)
N classification
NA	1 (0.27)
N0	253 (67.83)
N1	4 (1.07)
NX	115 (30.83)
M classification
M0	267 (71.58)
M1	4 (1.07)
MX	102 (27.35)
Radiation therapy
NA	25 (6.70)
No	340 (91.15)
Yes	8 (2.14)
Residual tumor
NA	7 (1.88)
R0	326 (87.40)
R1	17 (4.56)
R2	1 (0.27)
RX	22 (5.90)
Vital status
Deceased	130 (34.85)
Alive	243 (65.15)
Sample type
Primary tumor	371 (99.46)
Recurrent tumor	2 (0.54)
Overall survival (ten years)
No	237 (64.58)
Yes	130 (35.42)
Relapse-free survival (ten years)
No	179 (55.94)
Yes	141 (44.06)
EIF3B
High	109 (29.22)
Low	264 (70.78)

NA, not available; EIF3B, eukaryotic translation initiation factor 3 subunit B; G1-4, grade relating to degree of differentiation; T1-4, size and or extension of the primary tumor; TX, tumor could not be assessed; N0, no regional lymph node metastasis; N1, regional lymph node metastasis present; NX, lymph nodes could not be assessed; M0, no distant metastasis; M1, metastasis to distant organs; MX, metastasis could not be assessed; R0, no residual tumor visible under the microscope; R1, residual tumor visible under the microscope; R2, residual tumor visible to the naked eye; RX, residual tumor could not be assessed.

Expression of EIF3B in liver tissues

Boxplots revealed that EIF3B was significantly upregulated in liver cancer compared with that in normal liver tissues (Fig. 1A; P<0.001). Furthermore, EIF3B was also differentially expressed between subgroups by vital status (P=0.005) and histologic grade (P<0.001; Fig. 1).

Figure 1

Differences in EIF3B expression among subgroups of patients. Boxplots showing differences in EIF3B expression according to (A) tissue type (P<0.001 vs. normal tissue controls), (B) age, (C) sex, (D) histologic grade (P<0.001), (E) histological type, (F) T classification, (G) N classification, (H) M classification, (I) radiation therapy, (J) residual tumor classification, (K) sample type, (L) clinical stage and (M) vital status (P=0.005). EIF3B, eukaryotic translation initiation factor 3 subunit B. G1-4, grade relating to degree of differentiation; T1-4, size and or extension of the primary tumor; TX, tumor could not be assessed; N0, no regional lymph node metastasis; N1, regional lymph node metastasis present; NX, lymph nodes could not be assessed; M0, no distant metastasis; M1, metastasis to distant organs; MX, metastasis could not be assessed; R0, no residual tumor visible under the microscope; R1, residual tumor visible under the microscope; R2, residual tumor visible to the naked eye; RX, residual tumor could not be assessed.

Diagnostic capability of EIF3B

To assess the diagnostic performance of EIF3B in liver cancer, ROC curve analysis was used. The area under the curve (AUC) was 0.821, indicating that EIF3B has moderate diagnostic ability. In addition, similar AUCs were obtained for distinguishing normal liver tissues from liver cancer at specific stages (stage I, 0.785; stage II, 0.850; stage III, 0.853; stage IV, 0.916; Fig. 2).

Figure 2

Receiver-operating characteristic curves for EIF3B in The Cancer Genome Atlas-liver hepatocellular carcinoma dataset. The ability of EIF3B to distinguish between the following was assessed: (A) Non-tumor vs. tumor sample; non-tumor sample vs. tumor sample of (B) stage I, (C) stage II, (D) stage III and (E) stage IV. AUC, area under the curve; EIF3B, eukaryotic translation initiation factor 3 subunit B.

Association between EIF3B expression and clinical features of patients with liver cancer

As indicated in Table II, the vital status of the patients with liver cancer (P<0.001), overall survival (P<0.001; duration ten years) and the histologic grade (P<0.001) were associated with the expression of EIF3B.

Table II

Association between the expression of EIF3B and the clinicopathological characteristics of patients with liver cancer.

		EIF3B expression
Clinical characteristics	No. of patients	High	Low	χ2	P-value
Age (years)				0.089	0.765
<55	117	36 (33.03)	81 (30.80)
≥55	255	73 (66.97)	182 (69.20)
Sex				0.484	0.486
Female	121	32 (29.36)	89 (33.71)
Male	252	77 (70.64)	175 (66.29)
Histological type				1.251	0.534
Fibrolamellar carcinoma	3	0 (0.00)	3 (1.14)
Hepatocellular carcinoma	363	107 (98.17)	256 (96.97)
Hepatocholangiocarcinoma (mixed)	7	2 (1.83)	5 (1.89)
Histologic grade				17.796	<0.001
G1	55	9 (8.33)	46 (17.69)
G2	178	45 (41.67)	133 (51.15)
G3	123	46 (42.59)	77 (29.62)
G4	12	8 (7.41)	4 (1.54)
Stage				4.532	0.209
I	172	43 (40.95)	129 (52.87)
II	87	32 (30.48)	55 (22.54)
III	85	28 (26.67)	57 (23.36)
IV	5	2 (1.9)	3 (1.23)
T classification				7.720	0.102
T1	182	44 (40.37)	138 (52.67)
T2	95	34 (31.19)	61 (23.28)
T3	80	29 (26.61)	51 (19.47)
T4	13	2 (1.83)	11 (4.2)
TX	1	0 (0.00)	1 (0.38)
N classification				1.936	0.379
N0	253	77 (70.64)	176 (66.92)
N1	4	0 (0.00)	4 (1.52)
NX	115	32 (29.36)	83 (31.56)
M classification				2.882	0.236
M0	267	83 (76.15)	184 (69.70)
M1	4	2 (1.83)	2 (0.76)
MX	102	24 (22.02)	78 (29.55)
Radiation therapy				<0.001	>0.999
No	340	92 (97.87)	248 (97.64)
Yes	8	2 (2.13)	6 (2.36)
Residual tumor				5.116	0.163
R0	326	98 (91.59)	228 (88.03)
R1	17	5 (4.67)	12 (4.63)
R2	1	1 (0.93)	0 (0.00)
RX	22	3 (2.80)	19 (7.34)
Vital status				17.505	<0.001
Deceased	130	56 (51.38)	74 (28.03)
Alive	243	53 (48.62)	190 (71.97)
Sample type				0.017	0.895
Primary tumor	371	109(100)	262 (99.24)
Recurrent tumor	2	0 (0.00)	2 (0.76)
Overall survival (ten years)				18.690	<0.001
No	237	50 (47.17)	187 (71.65)
Yes	130	56 (52.83)	74 (28.35)
Relapse-free survival (ten years)				1.018	0.312
No	179	42 (50.6)	137 (57.81)
Yes	141	41 (49.4)	100 (42.19)

EIF3B, eukaryotic translation initiation factor 3 subunit B; G1-4, grade relating to degree of differentiation; T1-4, size and or extension of the primary tumor; TX, tumor could not be assessed; N0, no regional lymph node metastasis; N1, regional lymph node metastasis present; NX, lymph nodes could not be assessed; M0, no distant metastasis; M1, metastasis to distant organs; MX, metastasis could not be assessed; R0, no residual tumor visible under the microscope; R1, residual tumor visible under the microscope; R2, residual tumor visible to the naked eye; RX, residual tumor could not be assessed.

High expression of EIF3B is associated with poor overall survival of patients with liver cancer

Kaplan-Meier analysis indicated that high expression of EIF3B was significantly associated with poor overall survival (P<0.001; Fig. 3). Subgroup analysis provided similar results, particularly in female (P<0.001), younger (<55; P=0.008) and older subjects (≥55; P<0.001), and in patients with R0 (P<0.001), G2 (P=0.003), G3 (P<0.0001), stage I (P<0.001) and stage III (P<0.001).

Figure 3

Kaplan-Meier analysis of the influence of EIF3B expression on overall survival. (A) All patients. Subgroup analysis for (B) females, (C) males, (D) younger patients (<55), (E) older patients (≥55), (F) no lymph node dissection (R0), (G) lymph node dissection (R1/R2/RX), (H-M) histological grade, (H) G1/G2, (I) G3/G4, (J) G1, (K) G2, (L) G3, (M) G4, (N-Q) clinical stage (N) I, (O) II, (P) III, (Q) IV and (R) HCC. EIF3B, eukaryotic translation initiation factor 3 subunit B; HCC, hepatocellular carcinoma; G1-4, grade relating to degree of differentiation; T1-4, size and or extension of the primary tumor; TX, tumor could not be assessed; N0, no regional lymph node metastasis; N1, regional lymph node metastasis present; NX, lymph nodes could not be assessed; M0, no distant metastasis; M1, metastasis to distant organs; MX, metastasis could not be assessed; R0, no residual tumor visible under the microscope; R1, residual tumor visible under the microscope; R2, residual tumor visible to the naked eye; RX, residual tumor could not be assessed. The number of results censored (removed from surviving) is indicated below the survival curve.

As presented in Table III, T classification, stage, residual tumor and EIF3B expression were variables associated with overall survival according to univariate Cox regression analysis. In addition, multivariate Cox regression indicated that high EIF3B expression, T classification and residual tumor were independent risk factors for overall survival of patients with liver cancer [hazard ratio (HR)=2.44, 95% CI=1.71-3.47, P<0.001].

Table III

Summary of univariate and multivariate Cox regression analyses for overall survival duration (ten years).

	Univariate analysis			Multivariate analysis
Parameters	Hazard ratio	95% CI	P-value	Hazard ratio	95% CI	P-value
Age (≥55/<55 years)	1.00	0.69-1.45	0.997
Sex (male/female)	0.80	0.56-1.14	0.220
Histological type (hepatocholangiocarcinoma/Hepatocellular, Hepatocellular /fibrolamellar)	0.99	0.27-3.66	0.986
Histologic grade (G4/G3/G2/G1)	1.04	0.84-1.30	0.698
Stage (IV/III/II/I)	1.38	1.15-1.66	0.001	0.81	0.65-1.01	0.060
T classification (T4/T3/T2/T1/NX)	1.66	1.39-1.99	<0.001	1.91	1.51-2.42	<0.001
N classification (N1/N0/NX)	0.73	0.51-1.05	0.086
M classification (M1/M0/MX)	0.72	0.49-1.04	0.077
Radiation therapy (yes/no)	0.51	0.26-1.03	0.060
Residual tumor classification (RX/R2/R1/R0)	1.42	1.13-1.80	0.003	1.45	1.13-1.87	0.004
EIF3B (high/low)	2.41	1.70-3.42	<0.001	2.44	1.71-3.47	<0.001

EIF3B, eukaryotic translation initiation factor 3 subunit B; G1-4, grade relating to degree of differentiation; T1-4, size and or extension of the primary tumor; TX, tumor could not be assessed; N0, no regional lymph node metastasis; N1, regional lymph node metastasis present; NX, lymph nodes could not be assessed; M0, no distant metastasis; M1, metastasis to distant organs; MX, metastasis could not be assessed; R0, no residual tumor visible under the microscope; R1, residual tumor visible under the microscope; R2, residual tumor visible to the naked eye; RX, residual tumor could not be assessed.

High expression of EIF3B is associated with poor relapse-free survival of patients with liver cancer

Kaplan-Meier analysis indicated that patients with high expression of EIF3B had significantly poorer relapse-free survival (P=0.013; Fig. 4). Subgroup analysis provided similar results, particularly in female (P=0.001) and older (≥55; P=0.032) subjects and in patients with R0 (P=0.020), G2 (P<0.001) and stage III (P=0.004).

Figure 4

Kaplan-Meier analysis of the influence of EIF3B expression on relapse-free survival. (A) All patients. (B-R) Subgroup analysis for (B) females, (C) males, (D) younger patients (<55), (E) older patients (≥55), (F) no lymph node dissection (R0), (G) lymph node dissection (R1/R2/RX), (H-K) histological grade (H) G1, (I) G2, (J) G3, (K) G4, (L-N) clinical stage (L) I, (M) II, (N) III and (O) HCC. EIF3B, eukaryotic translation initiation factor 3 subunit B; HCC, hepatocellular carcinoma; G1-4, grade relating to degree of differentiation; T1-4, size and or extension of the primary tumor; TX, tumor could not be assessed; N0, no regional lymph node metastasis; N1, regional lymph node metastasis present; NX, lymph nodes could not be assessed; M0, no distant metastasis; M1, metastasis to distant organs; MX, metastasis could not be assessed; R0, no residual tumor visible under the microscope; R1, residual tumor visible under the microscope; R2, residual tumor visible to the naked eye; RX, residual tumor could not be assessed. The number of results censored (removed from surviving) is indicated below the survival curve.

As presented in Table IV, T classification, stage, residual tumor and EIF3B expression were variables associated with relapse-free survival according to the univariate Cox regression analysis. In addition, high EIF3B expression, T classification and residual tumor were independent risk factors for relapse-free survival of patients with liver cancer in the multivariate Cox regression analysis (HR=1.54, 95% CI=1.06-2.23, P=0.022).

Table IV

Summary of univariate and multivariate Cox regression analyses or relapse-free survival duration.

	Univariate analysis			Multivariate analysis
Parameters	Hazard ratio	95% CI	P-value	Hazard ratio	95% CI	P-value
Age (≥55/<55 years)	0.90	0.63-1.28	0.550
Sex (male/female)	0.99	0.70-1.41	0.966
Histological type (hepatocholangiocarcinoma/hepatocellular, hepatocellular /fibrolamellar)	2.02	0.66-6.24	0.220
Histologic grade (G4/G3/G2/G1)	0.98	0.80-1.21	0.883
Stage (IV/III/II/I)	1.66	1.38-1.99	<0.001	1.10	0.85-1.42	0.473
T classification (T4/T3/T2/T1/TX)	1.78	1.49-2.12	<0.001	1.67	1.28-2.18	<0.001
N classification (N1/N0/NX)	0.97	0.67-1.40	0.874
M classification (M1/M0/MX)	1.17	0.79-1.74	0.432
Radiation therapy (yes/no)	0.74	0.26-2.16	0.584
Residual tumor classification (RX/R2/R1/R0)	1.28	1.01-1.61	0.042	1.36	1.07-1.73	0.012
EIF3B (high/low)	1.58	1.10-2.28	0.014	1.54	1.06-2.23	0.022

EIF3B, eukaryotic translation initiation factor 3 subunit B; G1-4, grade relating to degree of differentiation; T1-4, size and/or extension of the primary tumor; TX, tumor could not be assessed; N0, no regional lymph node metastasis; N1, regional lymph node metastasis present; NX, lymph nodes could not be assessed; M0, no distant metastasis; M1, metastasis to distant organs; MX, metastasis could not be assessed; R0, no residual tumor visible under the microscope; R1, residual tumor visible under the microscope; R2, residual tumor visible to the naked eye; RX, residual tumor could not be assessed.

Signaling pathways associated with EIF3B

To identify the signaling pathways associated with EIF3B in liver cancer, GSEA was performed between the low EIF3B expression dataset and the high EIF3B expression dataset. The enrichment of the molecular signatures database (MSigDB) determined by GSEA was significantly different (nominal P-value <0.050, false discovery rate <0.250; Table V). As presented in Fig. 5 and Table V, GSEA indicated that MYC-V1 (HALLMARK_MYC_TARGETS_V1 geneset; P=0.009), MYC-V2 (HALLMARK_MYC_TARGETS_V2 geneset; P=0.004) and DNA repair pathways (HALLMARK_DNA_REPAIR geneset; P<0.001) were differentially enriched in high EIF3B expression and low EIF3B expression groups.

Table V

Gene sets enriched in phenotype high.

Molecular signatures database collection	Gene set name	NES	NOM P-value	FDR q-value
h.all.v6.2.symbols.gmt	HALLMARK_MYC_TARGETS_V2	2.154	0.004	0.006
h.all.v6.2.symbols.gmt	HALLMARK_MYC_TARGETS_V1	2.028	0.009	0.010
h.all.v6.2.symbols.gmt	HALLMARK_DNA_REPAIR	2.006	<0.001	0.009

Gene sets with NOM P-value <0.050 and FDR q-value <0.250 were considered as significant. FDR, false discovery rate; NES, normalized enrichment score; NOM, nominal.

Figure 5

Enrichment plots from GSEA. The GSEA results indicated that (A) the MYC-V1, (B) the MYC-V2 and (C) the DNA repair pathway are differentially enriched in high EIF3B expression and low EIF3B expression groups. GSEA, gene set enrichment analysis.

Discussion

Liver cancer is associated with a high mortality rate worldwide; the development of this cancer type may be influenced by viral infection, diet and environmental factors (21-24). In recent years, with the continuous progression of molecular biology and treatments, including chemotherapeutic drugs and surgical technology, the understanding of cancer biology and the treatment of liver cancer have made great progress. However, the prognosis of liver cancer remains poor. The World Health Organization/International Classification of Diseases-10 classifies diseases according to their etiology, pathology, clinical manifestations and anatomical location. Cancer is a gene-associated disease and molecular typing is required to deepen our understanding of the underlying mechanisms of disease development (25). Therefore, novel biomarkers are urgently required. Our research group has been exploring novel biomarkers for a number of years (26-38). The present study focused on EIF3B and indicated that EIF3B is a potential and independent prognostic biomarker for liver cancer.

EIF3B is closely linked to cancer progression. Consistent with previous studies, it was indicated that EIF3B was highly expressed in patients with liver cancer. Although Golob-Schwarzl et al (15) reported that EIF3B was upregulated in HCV-associated HCC, all of their patients were Asians. The patients assessed in the present study were from all over the world and covered other types of liver cancer that may be related to HCV. All of these results indicate that EIF3B has an important role in cancer-associated processes. A previous study suggested that EIF3B is involved in the proliferation and metastasis of gastric cancer (39). In addition, the present study further determined that EIF3B is associated with the histologic grade and survival status of patients with liver cancer. Therefore, in-depth studies using experimental and bioinformatics methods are required.

EIF3B has been indicated to have a marked influence on the prognosis of patients with cancer. In the present study, it was observed that upregulation of EIF3B was associated with poor overall/relapse-free survival of patients with liver cancer. In addition, patients with high EIF3B expression in clear cell renal cell carcinoma, esophageal squamous cell carcinoma and non-small cell lung cancer had a shorter survival time (6,7,11). In order to further explore the association between EIF3B and clinical characteristics of patients with liver cancer, a subgroup analysis was performed. As the TCGA-LIHC dataset does not have a Barcelona Clinic Liver Cancer staging system, TNM staging was used. Kaplan-Meier subgroup analysis indicated that high expression of EIF3B was associated with poor overall survival in the subgroups of females, younger (<55) or older (≥55) patients, R0, G2, G3, stage I and stage III. Furthermore, high expression of EIF3B was associated with poor relapse-free survival in the subgroups of females, older patients, R0, G2 and stage III. However, Tian et al (11) reported that upregulation of EIF3B was associated with tumor depth, TNM stage and lymph node metastasis in patients with esophageal squamous cell carcinoma. However, the results of the present study indicated that EIF3B was related to histological grade and survival status. This may be linked to the heterogeneity of tumor types and individual differences, which may help to select personalized treatments. Owing to the TCGA-LIHC data not including the body mass index and the presence of diabetes mellitus as variables, it is not possible to calculate their association with the prognosis of patients.

In the GSEA analysis, high EIF3B expression was indicated to be associated with MYC-V1, MYC-V2 and DNA repair in liver cancer. The MYC oncogene is an important regulator of liver cancer progression. Previous studies have indicated that MYC is able to promote the proliferation, metastasis and metabolism of liver cancer by regulating signaling pathways including AKT/mTOR and RAS/mitogen-activated protein kinase (40-42). In addition, each replication of DNA in cancer cells may cause a large amount of damage, including DNA substitutions or deletions (43). Therefore, DNA repair mechanisms (damage induction, signal transduction, signal response) are particularly important. This may explain why EIF3B may promote the progression of liver cancer through MYC-V1/V2 and DNA repair pathways.

The present study mainly uncovered the prognostic value of the EIF3B mRNA expression in liver cancer. Along with other studies on EIF3B, the present study contributed to a better understanding of the role of EIF3B, as well as the great possibility for precise prognostication. However, the underlying mechanisms remain to be fully elucidated and require further exploration by scientific research. In the future, the mechanisms of EIF3B will be studied at a deeper level.

In conclusion, the present study investigated the prognostic value of EIF3B in patients with liver cancer. High EIF3B expression was proved to be a potential and independent prognostic biomarker for liver cancer. Future work will include in vivo and in vitro experiments to explore the biological functions of EIF3B and the underlying mechanisms.