Abstract
Background/Aim: Epithelial cell adhesion molecule (EpCAM) is expressed in hepatic progenitor cells and hepatocellular carcinoma (HCC), and is considered a marker of liver cancer stem cells. Materials and Methods: A total of 262 patients were enrolled who had undergone surgical resection for HCC, with immunohistochemical staining results for EpCAM. The immunohistochemical expression of EpCAM and other stemness-related markers was evaluated as prognosticators of tumor recurrence and survival in patients who underwent surgical resection for HCC. Results: A multivariate Cox regression analysis showed that tumor size [hazard ratio (HR)=2.26, p=0.005], intrahepatic metastasis (HR=2.31, p=0.011), and EpCAM positivity (HR=1.74, p=0.038) were associated with tumor recurrence. In a Kaplan–Meier survival analysis, patients with EpCAM-positive tumors had a significantly higher tumor recurrence rate and a reduced overall survival compared to those with EpCAM-negative tumors. Conclusion: Immunohistochemical expression of EpCAM was identified as a poor prognosticator of recurrence and survival after surgical resection in patients with HCC.
Hepatocellular carcinoma (HCC) is one of the most frequently occurring cancers and the third leading cause of cancer-related death in Asia (1). A high prevalence of hepatitis B (HBV) and hepatitis C virus infections strongly predisposes the population to chronic liver disease and subsequent development of HCC in Asian countries (2-4). Although treatments for HCC have evolved, only those limited to early HCC have satisfactory treatment outcomes, and the mortality rate remains high despite the application of surgical resection, liver transplantation, and various other treatment modalities. Curative resection is widely considered as the first choice of therapy for patients with resectable HCC, however, the relatively high postoperative recurrence of HCC remains an unsolved problem, as up to 70% of patients will experience a recurrence within 5 years after curative resection (5-7). Prognostic predictions regarding recurrence and mortality should be made soon after surgical resection, as this information can help to direct the postoperative management strategies in patients undergoing surgical resection.
Two basic models have been proposed to explain the initiation and development of tumors in vivo. In the first model, the accumulation of mutations in somatic cells leads to uncontrolled proliferation and, finally, to carcinogenic transformation (8-10). The second model assumes the existence of a hierarchical tumor cell arrangement, wherein only a subpopulation of cells harbors the capacity for tumorigenesis (11, 12). In the latter model, the cells of interest possess two main characteristics, self-renewal and multipotency, and express stemness-related markers such as epithelial cell adhesion molecule (EpCAM), cytokeratin (CK) 7 and 19, glypican-3 and cluster of differentiation (CD)133 (13, 14).
EpCAM is expressed on a subset of normal epithelial cells and can be overexpressed on malignant cells from a variety of different tumor entities. This overexpression is even more pronounced on so-called tumor-initiating cells associated with many carcinomas (15). EpCAM is also expressed on hepatic progenitor cells and fetal hepatoblasts, and is considered a biomarker of hepatic cancer stem cell. Notably, EpCAM-positive HCC cells derived from cell lines and tumor specimens were found to be highly invasive and tumorigenic, suggesting that EpCAM might be a useful biomarker of HCC-initiating cells (16-18).
Flow diagram of study enrollment, allocation, and analysis.
Given the importance of prognostic prediction for patients with resected HCC, clinicians should identify those who are at a high risk of recurrence following treatment in order to consider the need for additional therapies or more careful follow-up. Although previous studies have described various prognostic indicators in patients with HCCs, reliable prognosticators are limited in number. In this study, the immunohistochemical (IHC) expression of stemness-related markers as prognosticators of tumor recurrence and survival was evaluated in patients who underwent surgical resection for HCC, with a particular focus on the potential of EpCAM as a predictive biomarker.
Materials and Methods
Patients. Four hundred and eighty patients were identified, who had undergone surgical resection for primary HCC at Ajou University Hospital (Suwon, Republic of Korea) between January 2011 and January 2017. Among them, 262 patients who had postoperative pathological findings of EpCAM expression were included in this study. In addition to EpCAM, the following stemness-related markers were evaluated: alpha-fetoprotein (AFP; n=248), CK19 (n=140), CK7 (n=182), glypican-3 (n=245) and CD34 (n=195). Data on CD34 IHC were excluded from the final analysis because all but six patients were positive for this marker (Figure 1). Patients were followed-up at 3-month intervals in the first year after surgery, at 4-month intervals in the second year, and then every six months thereafter. A contrast-enhanced abdomen computed tomography or magnetic resonance imaging scan was performed every visit at the out-patient department during the follow-up period.
All processes regarding access to their medical records were conducted in accordance with the guidelines of institutional Review Board (approval no. AJIRB-MED-MDB-17-005).
Samples and pathological data. Tumor size, presence of vascular invasion, tumor stage and serum AFP level, in addition to patient age and sex, were analyzed. Resected surgical specimens were subjected to IHC to identify the protein expression of EpCAM, AFP, CK19, CK7, and glypican-3. Tumors were graded using the criteria proposed by Edmondson and Steiner: I, Well-differentiated; II, moderately differentiated; III, poorly differentiated; and IV, undifferentiated (19). Tumor staging was based on the modified Union for International Cancer Control (UICC) staging system (20, 21). The analysis included patients who were followed-up for at least 6 months and for whom recurrence and death had been confirmed. The mean follow-up period was 29.9±17.15 (median=26, range=6-75) months.
IHC staining. HCC specimens collected from patients who had undergone resection were fixed in 10% formalin and embedded in paraffin. Four-micrometer-thick sections of the paraffin-embedded tissue blocks were used for IHC studies. All slides were scored by two pathologists blinded to the patients' clinical information (Kim YB and Roh J). The proportion of cell staining was used to evaluate the IHC results, ranging from 0 to 100%. The extent of positivity was assigned a score of 0 when no positive cells were observed; other results were semi-quantitatively scored as follows: negative, <5%; weak (1+), 5-30%; moderate (2+), 30-60%; and strong staining (3+), >60%. The expression of EpCAM and other stemness-related markers was considered positive if >10% of cells received a final score of moderate or strong staining (Figure 2).
Immunohistochemical staining of resected hepatocellular carcinoma (HCC) specimens for epithelial cell adhesion molecule (EpCAM) and other stemness-related markers. A: Hematoxylin and eosin-stained HCC (magnification: ×100). B: Specimen positive for EpCAM expression (magnification: ×200). C: Specimen negative for EpCAM expression (magnification: ×200). D: Specimen positive for CK19 expression (magnification: ×200). E: Specimen positive for CK7 expression (magnification: ×200). F: Specimen positive for glypican-3 expression (magnification: ×200).
Statistical analysis. The clinicopathological characteristics of the patients with HCC were evaluated using the chi-square test and Fisher's exact test. A multivariate Cox regression hazards model with forward stepwise entry was used to identify independent predictors of survival. The classification accuracy was measured using the area under the receiver-operating characteristic curve (AUROC). Using this curve, the cut-off value was determined as the point where the value of the sum yielded the greatest sensitivity and specificity, and was used to estimate the sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy. Kaplan–Meier survival curves were used to evaluate the cumulative recurrence and overall survival outcomes, and statistical differences in survival curves between the low-risk and high-risk groups were determined using the log-rank test. Recurrence and overall survival time were defined as the interval between the dates of curative resection and first recurrence or death, respectively. All statistical analyses were performed using the open-source statistical programming environment R (R software program ver. 3.1.2; R Project for Statistical Computing, Vienna, Austria). A p-value of less than 0.05 was considered statistically significant.
Results
Baseline clinicopathological characteristics. Of the 262 patients included in the final analysis, 219 (83.6%) were male and the mean age was 55.2 (median: 54, range: 31-78) years. The main etiology of the HCC was hepatitis B virus (HBV) infection (82.4%). The mean tumor size was 4.6 cm (median: 3.2, range: 1.3-9.0), and 190 patients (72.5%) presented with tumors smaller than 5 cm. Two hundred and thirty-seven patients (90.5%) had a single tumor mass.
Seventy-eight patients (30.0%) had a serum AFP level of more than 200 ng/ml. Among the study subjects, 179 (68.3%) had liver cirrhosis. IHC revealed that 123 out of 262 patients (46.9%) harbored EpCAM-positive tumors. Additionally, 112 (45.2%), 50 (35.7%), 99 (54.4%) and 206 (78.6%) specimens were positive for AFP, CK19, CK7, and glypican-3, respectively. Other clinical data and pathological results are listed in Table I.
Factors predictive of HCC recurrence after surgical resection. For the univariate analyses, a Cox proportional hazards model was used to investigate factors predictive of HCC recurrence after surgical resection. Of the investigated variables, tumor size ≥5 cm), tumor stage >2, microvessel invasion, intrahepatic metastasis, serum AFP ≥200 ng/ml, and EpCAM positivity were identified as significant predictors of HCC recurrence (p<0.05). In the multivariate analysis, however, tumor size [hazard ratio (HR)=3.69, 95% CI=1.310-10.409, p=0.013], intrahepatic metastasis (HR=5.06, 95% CI=1.387-18.490, p=0.014), and EpCAM positivity (HR=2.94, 95% CI=1.054-8.173, p=0.039) remained significant predictors of HCC recurrence (Table II). EpCAM positivity was the only significant stemness-related immunohistochemical marker identified as a predictor of HCC recurrence in the multivariate analysis.
Clinicopathological characteristics of patients who had undergone resection for hepatocellular carcinoma.
Clinicopathological characteristics and outcomes according to IHC expression of EpCAM. To determine the clinical significance of EpCAM after surgical resection for HCC, the patients were divided into two groups according to EpCAM expression (Table III). The two groups did not differ in terms of sex, age, etiology of HCC, tumor size, tumor multiplicity, tumor stage, presence of cirrhosis, or serum total bilirubin levels. However, significant intergroup differences were observed in serum AFP level (p<0.001) and various pathological factors (microvessel invasion, p=0.003; portal vein invasion, p=0.006; intrahepatic metastasis, p=0.041; Edmonson grade, p=0.002) (Table III).
The cumulative recurrence rate was higher among those with EpCAM-positive tumors than those with EpCAM-negative tumors (2-year cumulative recurrence: 44.0% vs. 21.5%, p<0.001). Furthermore, patients with EpCAM-positive tumors had a worse overall survival rate compared to their counterparts with EpCAM-negative tumors (2-year overall survival: 82.1% vs. 92.4%, p=0.018; Figure 3).
EpCAM and serum AFP level as a predictor of HCC recurrence and survival. A high serum AFP level was confirmed to be a significant risk factor for recurrence in univariate analysis, thus, we next conducted a recurrence and survival analysis after stratifying patients based on the combined EpCAM and serum AFP (≥200 ng/ml). Patients with a high serum AFP level and EpCAM-positive tumors had a significantly higher 2-year cumulative recurrence rate compared to the other patients (45.9% vs. 28.3%, p=0.026). Additionally, these patients had a tendency to have a lower survival rate (2-year cumulative survival rates: 81.7% vs. 89.0%, p=0.069) (Figure 4).
Diagnostic accuracy of EpCAM and serum AFP level as a predictor of HCC recurrence. The value of IHC EpCAM expression, alone and combined with serum AFP level (≥200 ng/ml), as a predictor of HCC recurrence after surgical resection was evaluated. For EpCAM alone, the AUROC for recurrence was 0.607 (95% CI=0.545-0.670, p=0.004); the sensitivity, specificity, positive predictive value, and negative predictive value were 0.609, 0606, 0.455, and 0.741, respectively. Even though a benefit was expected from combining EpCAM expression and serum AFP level, the AUROC value for this combination was only 0.572 (95% CI=0.510-0.633, p=0.054), with a sensitivity, specificity, positive predictive value and negative predictive value of 0.293, 0.851, 0.519 and 0.688, respectively (Figure 5).
Kaplan–Meier analyses of recurrence (A) and overall survival (B) among patients with epithelial cell adhesion molecule (EpCAM)-positive (HCC) vs. those with EpCAM-negative hepatocellular carcinomas.
Univariate and multivariate analyses of clinicopathological parameters associated with cumulative recurrence of hepatocellular carcinoma.
Discussion
In our study, we found that tumor expression of EpCAM, alone and in combination with the serum AFP level, may serve as a prognostic marker in patients with HCC who have undergone surgical resection. Specifically, the IHC expression of EpCAM was found to be associated with a higher risk of recurrence and shorter overall survival in patients with HCC. Additionally, the combination of positive EpCAM expression and a high serum AFP level was associated with rapid recurrence after surgical resection.
EpCAM is expressed on several types of epithelial cells and gastrointestinal carcinomas, where it acts as a homotypic calcium-independent cell adhesion molecule (22, 23). EpCAM strongly influences both the cell cycle and proliferation and contributes to up-regulation of the proto-oncogenes c-MYC and cyclins A and E (23). Furthermore, the inhibition of EpCAM expression was shown to reduce proliferation and metabolism in human cancer cells (24). In our study, the results obtained from a relatively simple IHC analysis of resected surgical specimens were consistent with these earlier findings. The EpCAM-positive group had a higher recurrence rate and lower survival rate compared to their EpCAM-negative HCC counterparts. These results suggest that a useful analysis of the resected surgical specimen can be conducted and that a patient's prognosis might be sufficiently predicted even from a preoperative biopsy.
A recent study compared peritumoral EpCAM expression and CD13 as a prognostic marker in patients who underwent curative resection of HBV-related HCC, using samples of liver parenchyma located within 20 mm of the tumor (25). The authors concluded that peritumoral EpCAM and CD13 expression were associated with a poor prognosis, but EpCAM may be a better prognostic marker than CD13 in HBV-related patients with HCC (25).
Another recent study showed an association between the preoperative presence of EpCAM-expressing circulating tumor cells (CTC) and T-regulatory cell level with the recurrence of HCC after radical hepatic resection (26). In that study, the group with a high level of CTCs had a significantly higher 1-year recurrence rate compared with the group with lower CTC level (50.0% vs. 10.3%, p=0.004) in patients with HCC (26). Another investigation reported a correlation of survival outcomes with EpCAM-positive CTCs from patients with HCC. Patients with EpCAM-positive CTCs had a significantly reduced overall survival compared to patients without these cells (p=0.017), and the presence of CTCs was also found to correlate with a high serum AFP level (p=0.050) (27). The same group also reported a correlation of EpCAM-positive CTCs with a high recurrence rate (HR=2.3, p=0.027) and shorter recurrence-free survival among patients who underwent curative resection for HCC (5.0±1.5 vs. 12.0±2.6 months, p=0.039) (28). To our knowledge, there are no previous studies addressing the correlation between the IHC expression of EpCAM and the level of EpCAM-positive CTCs, thus further studies are needed to determine the relationship between these two measures.
CK19 and CK7 are induced in biliary marker-positive hepatic progenitor cells undergoing carcinogenesis (29, 30). These markers are not always expressed in HCCs. A previous study divided patients with HCC into three groups according to the CK19 and glypican-3 IHC results: CK19+/glypican-3+, CK19−/glypican-3+, CK19−/glypican-3−. The researchers recorded a significantly lower cumulative recurrence-free survival rate among the CK19+/glypican-3+ group relative to the other groups (31). Furthermore, Uenishi et al. reported that both CK19 and CK7 were associated with tumor-free survival, although only the association with CK19 remained significant in a multivariate analysis (risk ratio=2.2, p=0.0082) (32). Although we also analyzed the associations of stemness-related markers (AFP, CK19, CK7, and glypican-3) with recurrence and overall survival, no differences in recurrence and survival rates were found. In contrast to previous studies, the Cox proportional analysis failed to identify these markers as significant risk factors for recurrence in patients with HCC who underwent surgical resection.
Clinicopathological characteristics and hepatocellular carcinoma recurrence according to epithelial cell adhesion molecule (EpCAM) expression.
Kaplan–Meier analyses of recurrence (A) and overall survival (B) among patients with epithelial cell adhesion molecule (EpCAM)-positive hepatocellular carcinomas and high expression of alpha-fetoprotein (AFP >200 ng/ml) vs. other patients with hepatocellular carcinoma.
Receiver-operating characteristic curve of epithelial cell adhesion molecule (EpCAM) expression for the detection of hepatocellular carcinoma recurrence. A: EpCAM only. B: EpCAM combined with serum alpha-fetoprotein (AFP) level.
Serum AFP is a known prognostic marker in patients with HCC. We investigated whether the serum AFP level can be a complimentary prognostic marker when combined with IHC expression of EpCAM of resected HCC specimens. However, better prediction of recurrence or survival was not obtained when an elevated serum AFP level (>200 ng/ml) was combined with EpCAM expression compared to use of EpCAM expression alone.
In the current study, analysis yielded an AUROC of EpCAM for recurrence of 0.607. Among patients with urothelial carcinoma, IHC-detected EpCAM expression was reported to be associated with worse overall survival (p=0.033) and an AUROC of 0.604 (33). Given that both studies divided patients into EpCAM-positive and negative groups, the AUROC values would be expected to improve if patients were classified by IHC staining intensity with specific grade.
Our study had certain limitations. Firstly, our IHC findings could not be compared to those of CTCs from patient blood samples, given the retrospective design. As noted, previous studies have identified EpCAM-positive CTCs as a poor prognostic factor. Future studies should determine whether CTC findings correlate with IHC results. Secondly, this study included many patients with short postoperative follow-up periods after surgery (mean follow-up period=29.9 months). A longer follow-up period may increase the value of EpCAM as a prognostic marker. Finally, we evaluated only patients for whom EpCAM IHC data were available, rather than all patients who underwent surgical resection. We would expect more informative results from data that included all patients, regarding the usefulness of other stemness-related markers.
In conclusion, we identified EpCAM expression as a marker of poor prognosis of recurrence and survival in those undergoing surgical resection for HCC even after clinicopathological factors for prognosis were taken into account. We expect that IHC expression of EpCAM might be used to predict outcomes in patients with HCC who underwent percutaneous biopsy as well as hepatic resection.
Acknowledgements
This study was supported by a grant from the Korean Health R & D Project, Ministry of Health and Welfare, Republic of Korea (HI16C2011).
Footnotes
Conflicts of Interests
The Authors declare that they have no competing interest in regard to this study.
- Received May 11, 2018.
- Revision received June 5, 2018.
- Accepted June 13, 2018.
- Copyright© 2018, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.