Author Affiliations: Department of Pathology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea (Ahn S, Hyeon J and Park CK)

Corresponding Author: Cheol-Keun Park, MD, PhD, Department of Pathology, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Irwon-dong, Gangnam-gu, Seoul 135-710, Korea (Tel: 82-2-34102766; Fax: 82-2-34106396; Email: ckpark@skku.edu)

 

© 2013, Hepatobiliary Pancreat Dis Int. All rights reserved.

doi: 10.1016/S1499-3872(13)60046-6

 

Contributors: PCK proposed the study. AS and PCK performed research and wrote the first draft. HJ collected and analyzed the data. All authors contributed to the design and interpretation of the study and to further drafts. PCK is the guarantor.

Funding: None.

Ethical approval: This study was approved by the Institutional Review Board of Samsung Medical Center (Seoul, Korea).

Competing interest: No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subjects of this article.

 

 

BACKGROUND: Notch signaling is critical to physiologic angiogenesis and has been implicated in tumor angiogenesis and metastasis. Notch signaling was reported to exert either oncogenic or tumor-suppressive function in hepatocellular carcinoma (HCC) tumorigenesis. However, the prognostic significance of Notch receptors in HCC remains uncertain. In this study, we investigated the roles of Notch receptors in the prognosis of HCC.

 

METHODS: We investigated the expressions of Notch receptors in tumor tissue microarrays of 288 patients with primary HCC who had undergone curative resection using immunohistochemistry. Additionally, prognostic factors of HCC were examined by univariate and multivariate analyses. The median follow-up period was 97.1 months. Tumor recurrence was detected in 189 patients (65.6%), and 99 (34.4%) died of HCC.

 

RESULTS: Cytoplasmic expression of Notch1, cytoplasmic expression of Notch3, coexistent nuclear expression of Notch3, and cytoplasmic Notch4 overexpression were observed in 145 (50.3%), 60 (20.8%), 17 (5.9%), and 172 (59.7%) of the 288 HCCs, respectively. Multivariate analyses revealed that Notch1 expression (P=0.029), Edmondson grade III (P=0.038), and higher BCLC stage (P<0.001) were independent predictors of shorter disease-free survival. Cytoplasmic Notch3 expression tended to be an independent predictor of shorter disease-free survival (P=0.055). Notch1 expression (P=0.039), Notch4 overexpression (P=0.012), and higher BCLC stage (P<0.001) were independent predictors of shorter disease-specific survival. On univariate analysis, Notch1 expression tended to show an unfavorable influence on disease-specific survival (P=0.063) and Notch4 overexpression did not show an unfavorable influence on disease-specific survival (P=0.103).

 

CONCLUSIONS: Notch1 expression might be an independent predictor of both shorter disease-free survival and shorter disease-specific survival in HCC patients after curative resection. Notch4 overexpression might be an independent predictor of shorter disease-specific survival. Notch1 could be used as an immunohistochemical biomarker to detect patients with a high-risk of recurrence. Notch1 and Notch4 could be used as immunohistochemical biomarkers to detect patients with a shorter disease-specific survival.

 

(Hepatobiliary Pancreat Dis Int 2013;12:286-294)

 

KEY WORDS: Notch1; Notch4; hepatocellular carcinoma; recurrence; survival

The prognosis after surgical resection of hepatocellular carcinoma (HCC) remains poor because of a high rate of recurrence and lack of effective adjuvant therapy.^[1] Tumor recurrence complicates more than 70% of cases at five years,^[2] and the five-year survival rate is 60%-70%.^[1] Many signaling pathways and molecules are proved to be involved in HCC carcinogenesis. Molecules involved in cancer progression might serve as markers for early detection of recurrence and metastasis after curative hepatectomy.^[3] It is crucial to identify patients with a high-risk of recurrence and to develop more effective and targeted treatment strategies to improve disease outcome.

 

Notch receptors (Notch1, Notch2, Notch3, and Notch4) are involved in cell-fate determination during development, and stem cell renewal and differentiation of proliferating cells in many adult tissues.^[4] The Notch receptor is cleaved upon ligand-binding, and the released intracellular domain of Notch activates related transcriptional factors.^[5] Notch signaling is critical to physiologic angiogenesis and has been implicated in tumor angiogenesis and metastasis.^[6] Notch signaling is aberrantly activated due to upregulation of JAG1/Notch1 in breast cancer,^[7] amplification and overexpression of Notch3 in ovarian cancer,^[8] and chromosomal translocation of Notch3 in lung cancer.^[9] Expression of an activated Notch4 oncoprotein disrupts normal epithelial morphogenesis and induces an invasive phenotype in mammary epithelial cells.^[10] On the other hand, Notch signaling is anti-oncogenic for squamous cell carcinoma of the uterine cervix^[11] and skin.^[12] High expression of Notch1 is associated with poor overall survival in colorectal cancer patients, while high expression of Notch2 may predict good survival.^[13] Notch2 has a weaker transactivation domain, so the down-regulation of Notch2 may make Notch1 and Notch4 take advantage of more ligands, thereby leading to a stronger activation of Notch signaling.^[14]

 

Notch1 signaling was reported to exert either oncogenic or tumor-suppressive function in HCC tumori-genesis. Ning et al^[15] reported that down-regulation of Notch1 signaling inhibited HCC cell growth. It has been reported that Notch1 signaling could inhibit HCC cell growth by arresting cell cycle and inducing apoptosis in vitro and in vivo.^{[16, 17]} Lim et al^[18] reported that Snail/Notch1 activation increased the invasiveness of HCC cells in the presence of wild type p53, and Notch1 decreased the invasiveness of HCC cells in the absence of wild type p53. Recent reports^{[19, 20]} have shown overexpressions of Notch1, Notch3, and Notch4 proteins in human HCC tissues. Notch2 was down-regulated in HCC compared with adjacent nontumor liver.^[21] However, the prognostic significance of Notch receptors in HCC remained uncertain. In this study, we investigated the expression of Notch receptors to evaluate their prognostic roles in 288 HCC patients with long-term follow-up and extensive information on clinicopathologic characteristics.

 

 

Primary HCC tissues were collected from the 288 consecutive patients who were treated with curative hepatectomy at Samsung Medical Center, Seoul, Korea from July 2000 to May 2006 (237 males and 51 females; mean age 52.6 years, ranged 17-76). Two hundred and eighteen (75.7%) patients were infected with hepatitis B, 26 (9.0%) with hepatitis C, and 4 (1.4%) coinfected with hepatitis B and hepatitis C. No viral marker was recognized in 40 (13.9%) patients. We defined curative resection as complete resection of all tumor nodules with clear microscopic resection margins and no residual tumors as indicated by a computed tomography scan one month after surgery. None of the patients received preoperative chemotherapy. This study was approved by the Institutional Review Board of Samsung Medical Center.

 

Tissue samples were fixed in 10% formaldehyde and embedded in paraffin for histopathologic examination. Histopathologic features of HCCs examined were tumor size, histological differentiation, microvascular invasion, major portal vein invasion, intrahepatic metastasis, multicentric occurrence, and non-tumor liver pathology. Differentiation was graded histologically using the Edmondson and Steiner's criteria.^[22] Microvascular invasion was considered present when at least one or more endothelial cells or the tunica media of the vessel surrounded a neoplastic cell group. Intrahepatic metastasis and multicentric occurrence were defined according to the previously reported criteria.^[23] Liver cirrhosis was defined as a diffuse process characterized by fibrosis and the conversion of the normal liver architecture into structurally abnormal nodules. Chronic hepatitis was defined as liver disorders of varying causes and severity in which hepatic inflammation and necrosis continue for at least 6 months. Stage was determined according to both the American Joint Committee on Cancer (AJCC) staging system^[24] and Barcelona Clinic Liver Cancer (BCLC) staging classification.^[25]

 

We followed up all patients at least once every three months after surgery until December 31, 2010, and serum α-fetoprotein levels were assessed and computed tomography scans were performed. When tumor recurrence was suspected, precise diagnostic imaging was performed using magnetic resonance imaging. Disease-free survival was defined from the date of resection until the detection of tumor recurrence. While HCC is the cause of death in most patients with the disease, some patients die of liver failure or other causes in the absence of progressive HCC (30 of the 129 deaths in this study were due to non-HCC causes). We chose HCC-related mortality (disease-specific death) as the clinical endpoint for survival analysis, defined as: 1) tumor occupying more than 80% of the liver, 2) portal venous tumor thrombus (PVTT) proximal to the second bifurcation, 3) obstructive jaundice due to the tumor, 4) distant metastases, or 5) variceal hemorrhage with PVTT proximal to the first bifurcation.^[26] The median follow-up period was 97.1 months (range 40-126). Tumor recurrence was detected in 189 patients (65.6%), and 99 (34.4%) died of HCC.

 

All histologic sections were examined by two pathologists (Ahn S and Park CK) and representative tumor areas free from necrosis or hemorrhage were pre-marked in formalin-fixed paraffin-embedded blocks. Two 2.0-mm-diameter tissue cores were obtained from donor blocks and arranged in recipient paraffin blocks. Two cores of normal liver tissue from 12 patients with metastatic colonic carcinoma of the liver were included in each array block. Each tissue array block contained up to 60 tissue cores.

 

Horseradish peroxidase staining was used to visualize antigens on consecutive 4-µm thick sections. The sections were deparaffinized, hydrated, and immersed in peroxidase-blocking solution (Dako, Glostrup, Denmark) to inhibit endogenous peroxidase. For antigen retrieval, microwave pretreatment was performed with a 0.01 mol/L citrate buffer (pH 6.0) for 30 minutes. The sections were incubated overnight at 4 �� with monoclonal antibodies to Notch1 (#3608, 1:400; Cell Signaling Technology Inc., Beverly, MA), Notch2 (#4530, 1:400; Cell Signaling Technology Inc.), Notch3 (#5276, 1:200; Cell Signaling Technology Inc.), and Notch4 (#2423, 1:400; Cell Signaling Technology Inc.). These sections were then incubated in DakoREAL EnVision/HRP rabbit/mouse detection reagent (Dako) for 20 minutes at room temperature. Staining was visualized using the diaminobenzidine color substrate. The slides were counterstained with Mayer's hematoxylin. In negative controls, isotype-matched irrelevant antibodies were used to substitute for the primary antibodies.

 

We used a scoring method based on intensity and proportion of stained cells as reported previously.^{[21, 27]} The percentage of positive tumor cells was determined semi-quantitatively and each sample was scored on a scale of 0-4 (0, <1%; 1, 1%-25%; 2, 26%-50%; 3, 51%-75%; 4, 76%-100%). Staining intensity was determined as 0 (negative), 1 (weak), 2 (moderate), and 3 (strong). The immunoreactive score of each tumor was calculated by the sum of these two parameters. The total score was graded as negative (0), low (1-3), or high (4-7). All stained sections were assessed by two independent pathologists (Ahn S and Park CK) without knowledge of clinicopathologic features, and any differences in interpretation were resolved by consensus. Duplicate tissue cores for each tumor showed high levels of homogeneity for both intensity and percentage of stained cells. In cases of differences between duplicate tissue cores, the higher score was taken as the total score.

 

Statistical analyses were performed using SPSS software (SPSS Inc., Chicago, IL., USA). The Chi-square and Fisher's exact test were used for comparisons of variables. Survival curves were estimated using the Kaplan-Meier method and differences in survival distributions were evaluated by the log-rank test. Univariate and multivariate analyses were based on the Cox proportional hazards regression model. P values less than 0.05 were regarded as statistically significant.

 

 

All normal liver tissues were negative for both Notch1 and Notch3, while Notch4 was detected in 20%-50% of normal hepatocyte cytoplasm with weak staining intensity. Notch1, Notch3, and Notch4 were detected on the endothelial cells of both blood vessels in normal liver tissues and neovessels inside HCCs. In HCC, immunoreactivity for Notch1 was observed in the cytoplasm and membrane of tumor cells, Notch3 in the cytoplasm of tumor cells with or without nuclear expression, and Notch4 in the cytoplasm of tumor cells. There were no significant differences in staining intensity regardless of storage condition of paraffin blocks. We regarded Notch1 and Notch3 as positive when the tumor showed low or high immunoreactivity, and Notch4 as positive when the tumor showed high immunoreactivity. Cytoplasmic expression of Notch1, cytoplasmic expression of Notch3, and coexistent nuclear expression of Notch3 were observed in 145 (50.3%), 60 (20.8%), and 17 (5.9%) of the 288 HCCs, respectively. High immunoreactivity for Notch4 was observed in 172 of the 288 HCCs (59.7%) (Fig. 1). Notch1 expression was significantly associated with higher Edmondson grade (P=0.005) and higher α-fetoprotein level (P=0.010). Notch4 overexpression was significantly associated with lower Edmondson grade (P=0.002), no microvascular invasion (P=0.001), no intrahepatic metastasis (P=0.007), lower AJCC T-stage (P=0.006), and lower α-fetoprotein level (P<0.001). No significant correlations were observed between Notch3 expression and any clinicopathologic parameters (Table 1). Notch2 could not be effectively analyzed because of the high background noise of commercial anti-Notch2 antibody with formalin-fixed paraffin embedded tissues.

 

The disease-free survival and disease-specific survival rates for the 288 HCC patients were 42.7% and 78.2% at three years, 36.3% and 71.4% at five years, 30.1% and 67.1% at seven years, and 27.9% and 60.8% at nine years, respectively. On univariate analyses, larger tumor size, Edmondson grade III, microvascular invasion, major portal vein invasion, intrahepatic metastasis, higher AJCC T-stage, higher BCLC stage, higher α-fetoprotein level, and lower albumin level showed unfavorable influences on disease-free survival and disease-specific survival (Table 2). Notch1 expression showed an unfavorable influence on disease-free survival (P=0.028) and tended to show an unfavorable influence on disease-specific survival (P=0.063). The five-year disease-free survival rate of the Notch1-positive group was significantly lower than that of the Notch1-negative group (31.4% versus 41.4%) (Fig. 2A). The median disease-free survival of the Notch1-positive group and Notch1-negative group was 16.5 and 31.8 months, respectively. The five-year disease-specific survival rate of the Notch1-positive group was lower than that of the Notch1-negative group (67.4% versus 75.3%) (Fig. 2B). The median disease-specific survival of the Notch1-positive group and Notch1-negative group was 66.1 and 87.1 months, respectively. Cytoplasmic Notch3 expression showed an unfavorable influence on disease-free survival (P=0.049), but it was not a prognostic factor for disease-specific survival. The five-year disease-free survival rate of the cytoplasmic Notch3-positive group was significantly lower than that of the cytoplasmic Notch3-negative group (30.5% versus 37.9%) (Fig. 2C). The median disease-free survival of the cytoplasmic Notch3-positive group and cytoplasmic Notch3-negative group was 15.4 and 24.9 months, respectively. However, coexistent nuclear Notch3 expression or Notch4 overexpression was not a prognostic factor for disease-free survival or disease-specific survival (Fig. 2E, F). The five-year disease-specific survival rate was 69.0% for the Notch4-positive group and 75.0% for the Notch4-negtive group. The median disease-specific survival of the Notch4-positive group and Notch4-negative group was 75.2 and 83.0 months, respectively.

 

As tumor size, vascular invasion, intrahepatic metastasis, AJCC stage, and albumin serum level were associated with BCLC stage, and coexistent nuclear Notch3 expression was observed in only 5.9% of HCCs, we did not make multiple analyses with the indices to avoid potential bias. On multivariate analyses, Edmondson grade III (P=0.038), higher BCLC stage (P<0.001), and Notch1 expression (P=0.029) were independent predictors of shorter disease-free survival. Cytoplasmic Notch3 expression tended to be an independent predictor of shorter disease-free survival (P=0.055). Notch1-positive patients were more likely to suffer from recurrence than Notch1-negative patients (hazard ratio=1.399). Cytoplasmic Notch3-positive patients tended to be more likely to suffer from recurrence than cytoplasmic Notch3-negative patients (hazard ratio=1.394). Higher BCLC stage (P<0.001), Notch1 expression (P=0.039), and Notch4 overexpression (P=0.012) were independent predictors of shorter disease-specific survival. Notch1-positive patients were more likely to suffer from disease-specific death than Notch1-negative patients (hazard ratio=1.565). Notch4-positive patients were more likely to suffer from disease-specific death than Notch4-negative patients (hazard ratio=1.761) (Table 3).

 

 

This study demonstrated overexpressions of Notch1, Notch3, and Notch4 proteins in HCC, as compared to normal liver, which are consistent with previous reports.^{[19, 20]} Additionally, the prognostic factors of HCC were examined by univariate and multivariate analyses. We collected a large number of clinical HCC tissues with long-term follow-up data and extensive information on clinicopathologic characteristics. Thus, we were able to determine the roles of Notch receptors on patient survival independent of other prognostic factors of clinical outcome. As a result, Notch1 expression was an independent predictor of shorter disease-free survival, Notch3 expression tended to be an independent predictor of shorter disease-free survival, and Notch1 expression and Notch4 overexpression were independent predictors of shorter disease-specific survival. The roles of aberrant Notch signaling in tumors were found to be linked to tumor initiation and maintenance,^[28] tumor progression,^[29] and tumor angiogenesis.^[30] Notch receptors may affect the invasion and metastasis of certain tumors. A recent study showed that an increased Notch1 mRNA level was associated with AJCC staging and tumor venous invasion of HCC, and that knock-down of Notch1 reversed HCC tumor metastasis in a mouse model.^[31] Joo et al^[32] observed that Notch1 expression was significantly related to the cervical lymph node metastasis and the depth of invasion in tongue cancer patients. Yoon et al^[33] reported that Notch1 expression was correlated with advanced TNM stage in extrahepatic cholangiocarcinoma and gallbladder carcinoma, and Notch3 was correlated with advanced T stage. Ding et al^[34] reported that Notch4 expression was significantly higher in salivary adenoid cystic carcinomas with metastasis and recurrence than in those without metastasis and recurrence. Giovannini et al^[35] reported that Notch3, Jagged1, Delta1, and HES-1 were all expressed in HepG2 cell line derived from human HCC. In this study, coexistent nuclear Notch3 expression was not a prognostic factor of HCC. A recent report showed that cytoplasmic Notch3 expression correlated with the progression of extrahepatic cholangiocarcinoma, while coexistent nuclear Notch3 expression did not correlate with it.^[33] Further study is needed to examine the mechanism of action regarding nuclear Notch3 expression in HCC prognosis.

 

Our findings indicate that Notch1 and Notch4 are potential new prognostic markers for HCC after curative resection, and could help clinicians identify patients at high-risk of recurrence, or with shorter disease-specific survival. Disruptions of the fundamental signaling pathways that enable tumors to grow and invade could represent an elegant therapeutic approach. A recent report showed that new agents targeting Notch pathways inhibited tumor growth and reduced angiogenesis.^[36] Moreover, undetectable or very occasional expression of Notch1 in normal liver tissues suggests that targeting Notch1 for HCC therapy may not damage liver tissue.

 

This study demonstrates for the first time that Notch1 expression as an independent predictor of both shorter disease-free survival and shorter disease-specific survival, and Notch4 overexpression as an independent predictor of shorter disease-specific survival after curative resection in a large number of HCC patients with long-term follow-up. Notch1 could be used as an immunohistochemical biomarker to detect patients with a high-risk of recurrence. Notch1 and Notch4 could be used as immunohistochemical biomarkers to detect patients with a shorter disease-specific survival. Prospective studies with larger patient populations are needed to further investigate the value of Notch1 and Notch4 as prognostic predictors.

 

 

1 Llovet JM, Schwartz M, Mazzaferro V. Resection and liver transplantation for hepatocellular carcinoma. Semin Liver Dis 2005;25:181-200. PMID: 15918147

2 Poon RT. Prevention of recurrence after resection of hepatocellular carcinoma: a daunting challenge. Hepatology 2011;54:757-759. PMID: 21793027

3 Ho MC, Lin JJ, Chen CN, Chen CC, Lee H, Yang CY, et al. A gene expression profile for vascular invasion can predict the recurrence after resection of hepatocellular carcinoma: a microarray approach. Ann Surg Oncol 2006;13:1474-1484. PMID: 17009164

4 Lai EC. Notch signaling: control of cell communication and cell fate. Development 2004;131:965-973. PMID: 14973298

5 Miele L. Notch signaling. Clin Cancer Res 2006;12:1074-1079. PMID: 16489059

6 Leong KG, Karsan A. Recent insights into the role of Notch signaling in tumorigenesis. Blood 2006;107:2223-2233. PMID: 16291593

7 Reedijk M, Odorcic S, Chang L, Zhang H, Miller N, McCready DR, et al. High-level coexpression of JAG1 and NOTCH1 is observed in human breast cancer and is associated with poor overall survival. Cancer Res 2005;65:8530-8537. PMID: 16166334

8 Park JT, Li M, Nakayama K, Mao TL, Davidson B, Zhang Z, et al. Notch3 gene amplification in ovarian cancer. Cancer Res 2006;66:6312-6318. PMID: 16778208

9 Dang TP, Gazdar AF, Virmani AK, Sepetavec T, Hande KR, Minna JD, et al. Chromosome 19 translocation, overexpression of Notch3, and human lung cancer. J Natl Cancer Inst 2000;92: 1355-1357. PMID: 10944559

10 Soriano JV, Uyttendaele H, Kitajewski J, Montesano R. Expression of an activated Notch4(int-3) oncoprotein disrupts morphogenesis and induces an invasive phenotype in mammary epithelial cells in vitro. Int J Cancer 2000;86:652-659. PMID: 10797286

11 Radtke F, Raj K. The role of Notch in tumorigenesis: oncogene or tumour suppressor? Nat Rev Cancer 2003;3:756-767. PMID: 14570040

12 Proweller A, Tu L, Lepore JJ, Cheng L, Lu MM, Seykora J, et al. Impaired notch signaling promotes de novo squamous cell carcinoma formation. Cancer Res 2006;66:7438-7444. PMID: 16885339

13 Chu D, Zhang Z, Zhou Y, Wang W, Li Y, Zhang H, et al. Notch1 and Notch2 have opposite prognostic effects on patients with colorectal cancer. Ann Oncol 2011;22:2440-2447. PMID: 21378202

14 Garcia A, Kandel JJ. Notch: a key regulator of tumor angiogenesis and metastasis. Histol Histopathol 2012;27:151-156. PMID: 22207549

15 Ning L, Wentworth L, Chen H, Weber SM. Down-regulation of Notch1 signaling inhibits tumor growth in human hepatocellular carcinoma. Am J Transl Res 2009;1:358-366. PMID: 19956448

16 Wang C, Qi R, Li N, Wang Z, An H, Zhang Q, et al. Notch1 signaling sensitizes tumor necrosis factor-related apoptosis-inducing ligand-induced apoptosis in human hepatocellular carcinoma cells by inhibiting Akt/Hdm2-mediated p53 degradation and up-regulating p53-dependent DR5 expression. J Biol Chem 2009;284:16183-16190. PMID: 19376776

17 Qi R, An H, Yu Y, Zhang M, Liu S, Xu H, et al. Notch1 signaling inhibits growth of human hepatocellular carcinoma through induction of cell cycle arrest and apoptosis. Cancer Res 2003;63:8323-8329. PMID: 14678992

18 Lim SO, Park YM, Kim HS, Quan X, Yoo JE, Park YN, et al. Notch1 differentially regulates oncogenesis by wildtype p53 overexpression and p53 mutation in grade III hepatocellular carcinoma. Hepatology 2011;53:1352-1362. PMID: 21480340

19 Gramantieri L, Giovannini C, Lanzi A, Chieco P, Ravaioli M, Venturi A, et al. Aberrant Notch3 and Notch4 expression in human hepatocellular carcinoma. Liver Int 2007;27:997-1007. PMID: 17696940

20 Giovannini C, Gramantieri L, Chieco P, Minguzzi M, Lago F, Pianetti S, et al. Selective ablation of Notch3 in HCC enhances doxorubicin's death promoting effect by a p53 dependent mechanism. J Hepatol 2009;50:969-979. PMID: 19304334

21 Gao J, Song Z, Chen Y, Xia L, Wang J, Fan R, et al. Deregulated expression of Notch receptors in human hepatocellular carcinoma. Dig Liver Dis 2008;40:114-121. PMID: 17920003

22 Edmondson HA, Steiner PE. Primary carcinoma of the liver: a study of 100 cases among 48 900 necropsies. Cancer 1954;7:462-503. PMID: 13160935

23 Kumada T, Nakano S, Takeda I, Sugiyama K, Osada T, Kiriyama S, et al. Patterns of recurrence after initial treatment in patients with small hepatocellular carcinoma. Hepatology 1997;25:87-92. PMID: 8985270

24 Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti III A. AJCC cancer staging manual, 7th ed. Chicago, IL: Springer;2010.

25 Llovet JM, Brú C, Bruix J. Prognosis of hepatocellular carcinoma: the BCLC staging classification. Semin Liver Dis 1999;19:329-338. PMID: 10518312

26 Hoshida Y, Villanueva A, Kobayashi M, Peix J, Chiang DY, Camargo A, et al. Gene expression in fixed tissues and outcome in hepatocellular carcinoma. N Engl J Med 2008; 359:1995-2004. PMID: 18923165

27 Kim SH, Ahn S, Park CK. Smad3 and its phosphoisoforms are prognostic predictors of hepatocellular carcinoma after curative hepatectomy. Hepatobiliary Pancreat Dis Int 2012;11:51-59. PMID: 22251470

28 Mullendore ME, Koorstra JB, Li YM, Offerhaus GJ, Fan X, Henderson CM, et al. Ligand-dependent Notch signaling is involved in tumor initiation and tumor maintenance in pancreatic cancer. Clin Cancer Res 2009;15:2291-2301. PMID: 19258443

29 Balint K, Xiao M, Pinnix CC, Soma A, Veres I, Juhasz I, et al. Activation of Notch1 signaling is required for beta-catenin-mediated human primary melanoma progression. J Clin Invest 2005;115:3166-3176. PMID: 16239965

30 Dufraine J, Funahashi Y, Kitajewski J. Notch signaling regulates tumor angiogenesis by diverse mechanisms. Oncogene 2008;27:5132-5137. PMID: 18758482

31 Wang XQ, Zhang W, Lui EL, Zhu Y, Lu P, Yu X, et al. Notch1-Snail1-E-cadherin pathway in metastatic hepatocellular carcinoma. Int J Cancer 2012;131:E163-172. PMID: 22052196

32 Joo YH, Jung CK, Kim MS, Sun DI. Relationship between vascular endothelial growth factor and Notch1 expression and lymphatic metastasis in tongue cancer. Otolaryngol Head Neck Surg 2009;140:512-518. PMID: 19328339

33 Yoon HA, Noh MH, Kim BG, Han JS, Jang JS, Choi SR, et al. Clinicopathological significance of altered Notch signaling in extrahepatic cholangiocarcinoma and gallbladder carcinoma. World J Gastroenterol 2011;17:4023-4030. PMID: 22046092

34 Ding LC, She L, Zheng DL, Huang QL, Wang JF, Zheng FF, et al. Notch-4 contributes to the metastasis of salivary adenoid cystic carcinoma. Oncol Rep 2010;24:363-368. PMID: 20596622

35 Giovannini C, Lacchini M, Gramantieri L, Chieco P, Bolondi L. Notch3 intracellular domain accumulates in HepG2 cell line. Anticancer Res 2006;26:2123-2127. PMID: 16827154

36 Wu Y, Cain-Hom C, Choy L, Hagenbeek TJ, de Leon GP, Chen Y, et al. Therapeutic antibody targeting of individual Notch receptors. Nature 2010;464:1052-1057. PMID: 20393564