GENETIC TESTING AND MOLECULAR BIOMARKERS Volume 18, Number 8, 2014 ª Mary Ann Liebert, Inc. Pp. 580–586 DOI: 10.1089/gtmb.2014.0102
Diagnostic Significance of Serum Osteopontin Level for Pancreatic Cancer: A Meta-Analysis Jian-Jun Li,1 Hong-Yu Li,2 and Fei Gu1
Objective: This meta-analysis aimed to identify the significance of serum osteopontin (OPN) level for the diagnosis of pancreatic cancer (PC). Methods: Through searching the following electronic databases—the Cochrane Library Database (Issue 12, 2013), Web of Science (1945–2013), PubMed (1966–2013), CINAHL (1982–2013), EMBASE (1980–2013), and the Chinese Biomedical Database (CBM) (1982–2013)—related articles were determined without any language restrictions. The STATA statistical software (version 12.0; Stata Corporation, College Station, TX) was chosen to deal with statistical data. Standard mean difference (SMD) and its corresponding 95% confidence interval (95% CI) was calculated. Eleven clinical case–control studies, which recruited 491 PC patients and 481 healthy controls, were selected for statistical analysis. Results: Combined SMD of OPN suggested that the serum OPN level in PC patients was significantly higher than that in healthy controls (SMD = 3.58, 95% CI = 2.42–4.74, p < 0.001). Ethnicity stratified analysis indicated a higher serum OPN level in PC patients compared with control subjects among both Caucasians and Asians (Caucasians: SMD = 2.62, 95% CI = 1.33–3.91, p < 0.001; Asians: SMD = 4.54, 95% CI = 2.80–6.27, p < 0.001; respectively). Conclusion: The main finding of our meta-analysis revealed that an elevated serum OPN level may be used as a promising diagnostic tool for early identification of PC.
Introduction
H
uman pancreatic cancer (PC) is a malignant neoplasm originating from transformed cells arising in tissues forming the pancreas, which is widely regarded as the eighth or ninth most frequent cause of cancer death worldwide (Hidalgo, 2010; Maisonneuve and Lowenfels, 2010). The most common type of PC, accounting for 95% of these tumors, is adenocarcinoma arising within the exocrine component of the pancreas (Hidalgo, 2010). It has been estimated that an estimated 45,200 new invasive cases of PC were diagnosed in both sexes and about 38,460 deaths occurred in the United States in 2013 (Siegel et al., 2013). Furthermore, PC has an extremely poor prognosis, where the 5-year survival after resection of pancreatic adenocarcinoma (PDAC) was *12% and the 10-year relative survival was only 5%, which made PC an important global healthcare problem (Ferrone et al., 2008). In general, the pathogenesis of PC is complicated, and interactions between the genetic and environmental factors may serve as important predictors in the development of PC (Landi, 2009). A wide range of risk factors, including age, cigarette smoking, alcohol consump-
1 2
tion, unhealthy diet, obesity, and diabetes mellitus, have been demonstrated to be involved in PC development (Iodice et al., 2008; Anderson et al., 2009). Recently, epidemiological studies revealed that an increased serum level of osteopontin (OPN) was observed in PC patients, suggesting that the serum OPN level may be implicated in the pathogenesis of PC (Sullivan et al., 2009; Zhivkova-Galunska et al., 2010). The OPN protein, first identified as a bone matrix in 1986 in osteoblasts, is considered as an aspartic acid-rich, N-linked (SIBLING) glycosylated protein, which may be highly phosphorylated on serines and threonines according to the cell type (Scatena et al., 2007; Zhang et al., 2013). Acting as a phosphoglycoprotein produced by various tumor cells, OPN was initially characterized as a phosphoprotein secreted by transformed, malignant epithelial cells (Takahashi et al., 2009). Being one of the most critical extracellular structural proteins and an organic component of bone, OPN has various in vivo functions, such as cell adhesion and migration, immune and inflammatory response, apoptosis, and bone calcification (Ramaiah and Rittling, 2008; Wang and Denhardt, 2008). Evidence accumulated that OPN has been implicated in a number of physiological and pathological processes,
Department of Radiotherapy, The First Affiliated Hospital, China Medical University, Shenyang, People’s Republic of China. Department of Gastroenterology, General Hospital of Shenyang Military Area, Shenyang, People’s Republic of China.
580
SERUM OPN LEVEL IN PC PATIENTS
581
Identification
including fibrosis, neointima formation, arterial occlusion by thrombus, inflammation, and pulmonary hypertension (Cho and Kim, 2009; Inoue and Shinohara, 2011). Detection of plasma OPN level may therefore be useful in the treatment of autoimmune diseases, cancer metastasis, and bone mineralization diseases (Chang et al., 2010; Murugaiyan et al., 2010; Mi et al., 2011). Indeed, numerous studies have been shown that the increased level of OPN was detected in the development and progression of hepatocellular carcinoma, breast cancer, cervical cancer, gastric carcinoma, and so on (Cho et al., 2008; Mi et al., 2009; Shang et al., 2012). More importantly, the elevated serum OPN level could act as a biomarker for metastatic progression and poor survival in patients with pancreatic ductal adenocarcinoma (Brand et al., 2011). Therefore, through taking part in the promotion of cell growth, metastasis, and cell adhesion reduction, the elevated serum levels of OPN may be responsible for the invasion and metastasis of the PC cells and finally lead to the deterioration of tumors (Zhivkova-Galunska et al., 2010). In the past few decades, several previous studies have indicated that an increased serum OPN level may be a biomarker of PC based on its important effect on malignant cell proliferation, tumor metastasis, and survival (Koopmann et al., 2006; Yang et al., 2007); some results are not entirely consistent (Zhu, 2011; Poruk et al., 2013). In this regard, it is worth performing a
Articles identified through electronic database searching (N = 64)
meta-analysis of all available data to assess the value of serum OPN level for the diagnosis of PC patients. Materials and Methods Literature search and selection criteria
The identification of related articles was implemented through searching the following electronic databases without any language restrictions: PubMed, Springer Link, Karger Medical and Scientific Publishers, Chinese Biomedical Database (CBM), Chinese National Knowledge Infrastructure (CNKI), and Google scholar. We employed a search strategy of highly sensitivity in combination with the following keywords and MeSH terms: (‘‘osteopontin’’ or ‘‘OPN’’ or ‘‘sialoprotein 1’’ or ‘‘bone sialoprotein 1’’ or ‘‘secreted phosphoprotein 1’’ or ‘‘uropontin’’ or ‘‘Eta-1’’ or ‘‘2ar’’) and (‘‘pancreatic neoplasms’’ or ‘‘pancreatic cancer’’ or ‘‘pancreatic carcinoma’’ or ‘‘PC’’ or ‘‘pancreatic adenocarcinoma’’ or ‘‘pancreas cancer’’ or ‘‘pancreas carcinoma’’ or ‘‘pancreatic tumor’’). The references of the present articles and reviews were also manually searched for additional potential studies. The eligible studies included in our meta-analysis should meet the following three types of inclusion criteria: (1) must concern the diagnostic significance of serum OPN level for PC; (2) patients included in the current meta-analysis should
Additional articles identified through a manual search (N = 1)
Screening
Articles reviewed for duplicates (N = 65)
Articles after duplicates removed (N = 63) Studies were excluded, due to: (N = 8) Letters, reviews, meta-analysis (N = 11) Not human studies (N = 15) Not related to research topics
Eligibility
Full-text articles assessed for eligibility (N = 29) Studies were excluded, due to: (N = 2) Not case-control or cohort study (N = 5) Not relevant to serum osteopontin (N = 10) Not relevant to Pancreatic cancer
Included
Studies included in qualitative synthesis (N = 12)
Studies included in quantitative synthesis (meta-analysis) (N = 9)
FIG. 1. Flowchart of literature search and study selection. Nine studies were included in this metaanalysis.
582
LI ET AL. 20
Pubmed database All databases
FIG. 2. The distribution of topicrelated literature in the electronic database over the last decade.
Number of articles
15
10
5
0 2000~2001
2002~2003
2004~2005
2006~2007
2008~2009
2010~2011
2012~2013
Publication year
be diagnosed with cytologically or histologically proven PC or PDAC; (3) enough information on serum OPN level should be supplied in eligible articles. Articles that did not meet the inclusion criteria were excluded. If authors published several studies of the same subjects, either the most recent or largest sample size publication was included. Data extraction and methodological assessment
Two authors used a standardized form to independently collect relevant data from each included study. Controversial problems were resolved by discussion and consensus. By reviewing all the articles, two observers determined the inclusion and exclusion criteria. The language of publication, publication year of article, first author’s surname, geographical location, design of study, total number of cases, sample size, the source of controls, cancer type, detection method, and serum OPN level. Two observers separately assessed the methodological quality with the use of the Newcastle–Ottawa scale (NOS) criteria (Stang, 2010). The NOS criteria comprised three aspects: (1) subject selection: 0–4; (2) comparability of subject:
0–2; and (3) clinical outcome: 0–3. NOS scores ranged from 0 to 9, and a score with good quality should be ‡ 7. Statistical analysis
To achieve rigorous statistical analysis, we used the STATA statistical software (version 12.0; Stata Corporation, College Station, TX) to deal with statistical data. Standard mean difference (SMD) and its corresponding 95% confidence interval (95% CI) was calculated. The statistical significance of pooled SMDs was evaluated by the Z-test. Tests for heterogeneity are commonly used to decide on methods for combining studies and for selection of the random-effects model or fixed-effects model. Between-study heterogeneity was assessed by Cochran’s Q-statistic and I2 tests in our meta-analysis (Zintzaras and Ioannidis, 2005). A p-value < 0.05 or I2 > 50% means that these studies were heterogeneous, and then the random-effects model was employed; otherwise, the fixed-effects model was implemented. We also made use of subgroup analysis to explore reasons for the detection of heterogeneity. A sensitivity analysis was also implemented to evaluate the influence of a single study on the overall estimate. Funnel plots and Egger’s
Table 1. Main Characteristics and Methodological Quality of All Eligible Studies First author
Year
Sample size Gender (M/F) Cancer Ethnicity type Case Control Case Control
Poruk Zhu Chen Zhang Yang Chen Yang Koopmann Kolb
2013 2011 2011 2010 2010 2010 2007 2006 2005
Caucasian Asians Asians Asians Asians Caucasian Asians Caucasian Caucasian
PDAC PDAC PDAC PC PC PC PC PC PDAC
86 60 72 57 30 20 46 50 43 27
86 60 150 50 20 25 20 50 20 20
46/40 42/18 45/27 32/25 14/16 — 30/16 22/28 — —
46/40 45/15 — 28/22 10/10 — — 22/28 — —
Age (years) Case
Control
66 (44–87) 53 (35–75) 53 (32–81) 50.7 – 11.6 63 (39–82) — — 67.6 – 10.1 — —
44 (34–94) 51 (40–70) 51 (38–63) 49.1 – 10.5 63 (39–82) — — 67.6 – 10.1 — —
NOS Method Marker score ELISA ECLI ECLI ELISA ELISA ELISA ELISA ELISA ELISA ELISA
OPN OPN OPN OPN OPN OPN OPN OPN OPN OPN
8 8 7 8 6 5 5 7 5
ECLI, electrochemiluminescence immunoassay; ELISA, enzyme-linked immunosorbent assay; F, female; M, male; NOS, Newcastle– Ottawa scale; OPN, osteopontin; PC, pancreatic cancer; PDAC, pancreatic adenocarcinoma.
SERUM OPN LEVEL IN PC PATIENTS
583
Serum OPN level (Case vs. Control)
Included study
SMD (95% CI)
Weight%
Poruk KE (2013)
0.72 (0.41, 1.03)
10.48
Zhu RQ (2011)
3.08 (2.55, 3.61)
10.33
Chen ZG (2011)
3.37 (2.95, 3.79)
10.41
Zhang ZW (2010)
7.20 (6.16, 8.25)
9.73
Yang SQ (2010)
1.32 (0.69, 1.94)
10.25
Chen R (2010)
5.01 (3.80, 6.23)
9.47
Yang YC (2007)
8.30 (6.77, 9.83)
8.92
Koopmann J (2006)
1.09 (0.67, 1.52)
10.42
Kolb A−a (2005)
3.76 (2.91, 4.61)
9.99
2.98 (2.13, 3.82)
10.00
3.58 (2.42, 4.74)
100.00
Kolb A−b (2005) 2
Heterogeneity test (I = 97.4%, P < 0.001) Z test (Z = 6.05, P < 0.001) Random effects analysis −9.83
0
9.83
Serum OPN level
Serum OPN level
(Ethnicity: Case vs. Control)
Included study
FIG. 3. Forest plots for the significance of serum osteopontin (OPN) level for the diagnosis of pancreatic cancer.
SMD (95% CI) Weight%
(Sample size: Case vs. Control)
Included study
SMD (95% CI)
Weight%
Large
Caucasian Poruk KE (2013)
0.72 (0.41, 1.03)
10.48
Poruk KE (2013)
0.72 (0.41, 1.03)
10.48
Chen R (2010)
5.01 (3.80, 6.23)
9.47
Zhu RQ (2011)
3.08 (2.55, 3.61)
10.33
Koopmann J (2006)
1.09 (0.67, 1.52)
10.42
Chen ZG (2011)
3.37 (2.95, 3.79)
10.41
Kolb A−a (2005)
3.76 (2.91, 4.61)
9.99
Zhang ZW (2010)
7.20 (6.16, 8.25)
9.73
Kolb A−b (2005)
2.98 (2.13, 3.82)
10.00
Koopmann J (2006)
1.09 (0.67, 1.52)
10.42
2.62 (1.33, 3.91)
50.36
3.03 (1.45, 4.61)
51.37
Zhu RQ (2011)
3.08 (2.55, 3.61)
10.33
Yang SQ (2010)
1.32 (0.69, 1.94)
10.25
Chen ZG (2011)
3.37 (2.95, 3.79)
10.41
Chen R (2010)
5.01 (3.80, 6.23)
9.47
Zhang ZW (2010)
7.20 (6.16, 8.25)
9.73
Yang YC (2007)
8.30 (6.77, 9.83)
8.92
Yang SQ (2010)
1.32 (0.69, 1.94)
10.25
Kolb A−a (2005)
3.76 (2.91, 4.61)
9.99
Yang YC (2007)
8.30 (6.77, 9.83)
8.92
Kolb A−b (2005)
2.98 (2.13, 3.82)
10.00
49.64
(I 2
4.19 (2.25, 6.13)
48.63
(I 2 =
3.58 (2.42, 4.74)
100.00
Heterogeneity test (I2 = 95.9%, P < 0.001) Z test (Z = 3.99, P < 0.001)
Subtotal (I−squared = 98.3%, P < 0.001) Z test (Z = 3.77, P < 0.001) Small
Asians
(I2 =
4.54 (2.80, 6.27)
Heterogeneity test 97.0%, P < 0.001) Z test (Z = 5.12, P < 0.001) (I2 =
3.58 (2.42, 4.74)
Heterogeneity test 97.4%, P < 0.001) Z test (Z = 6.05, P < 0.001) Random effects analysis −9.83
0
100.00
= 95.4%, P < 0.001) Subtotal Z test (Z = 4.24, P < 0.001) Overall 97.4%, P < 0.001) Z test (Z = 6.05, P < 0.001) Random effects analysis −9.83
9.83
9.83
Serum OPN level
Serum OPN level (Cancer type: Case vs. Control)
Included study
0
SMD (95% CI) Weight%
PDAC
(Method: Case vs. Control)
Included study
SMD (95% CI) Weight%
ELISA
Poruk KE (2013)
0.72 (0.41, 1.03)
10.48
Poruk KE (2013)
0.72 (0.41, 1.03)
Zhu RQ (2011)
3.08 (2.55, 3.61)
10.33
Zhang ZW (2010)
7.20 (6.16, 8.25)
9.73
Chen ZG (2011)
3.37 (2.95, 3.79)
10.41
Yang SQ (2010)
1.32 (0.69, 1.94)
10.25
Kolb A−a (2005)
3.76 (2.91, 4.61)
9.99
Chen R (2010)
5.01 (3.80, 6.23)
9.47
09 Kolb A−b (2005)
2.98 (2.13, 3.82)
10.00
Yang YC (2007)
8.30 (6.77, 9.83)
8.92
2.76 (1.39, 4.14)
51.22
Koopmann J (2006)
1.09 (0.67, 1.52)
10.42
Kolb A−a (2005)
3.76 (2.91, 4.61)
9.99
Kolb A−b (2005)
2.98 (2.13, 3.82)
10.00
3.70 (2.23, 5.17)
79.25
Heterogeneity test (I2 = 97.2%, P < 0.001) Z test (Z = 3.94, P < 0.001) PC
Heterogeneity test (I 2 = 97.6%, P < 0.001) Z test (Z = 4.93, P < 0.001)
10.48
Zhang ZW (2010)
7.20 (6.16, 8.25)
9.73
Yang SQ (2010)
1.32 (0.69, 1.94)
10.25
Chen R (2010)
5.01 (3.80, 6.23)
9.47
Yang YC (2007)
8.30 (6.77, 9.83)
8.92
Zhu RQ (2011)
3.08 (2.55, 3.61)
10.33
Koopmann J (2006)
1.09 (0.67, 1.52)
10.42
Chen ZG (2011)
3.37 (2.95, 3.79)
10.41
4.52 (1.97, 7.08)
48.78
3.26 (2.93, 3.59)
20.75
3.58 (2.42, 4.74)
100.00
Heterogeneity test (I2 = 98.0%, P < 0.001) Z test (Z = 3.47, P = 0.001) (I2 =
3.58 (2.42, 4.74)
Heterogeneity test 97.4%, P < 0.001) Z test (Z = 6.05, P < 0.001) Random effects analysis −9.83
0
9.83
100.00
ECLI
Heterogeneity test (I 2 = 0.0%, P < 0.001) Z test (Z = 19.32, P < 0.001) (I 2 =
Heterogeneity test 97.4%, P < 0.001) Z test (Z = 6.05, P < 0.001) Random effects analysis −9.83
0
9.83
FIG. 4. Subgroup analyses by ethnicity, sample size, cancer type, and detection method for the diagnosis of pancreatic cancer.
584
LI ET AL.
Serum OPN level (Case vs. Control) Lower CI Limit
Estimate
Upper CI Limit
Poruk KE (2013)
Zhu RQ (2011)
Chen ZG (2011)
Zhang ZW (2010)
FIG. 5. Sensitivity analysis of the summary odds ratio coefficients for the diagnosis of pancreatic cancer.
Yang SQ (2010)
Chen R (2010)
Yang YC (2007)
Koopmann J (2006)
Kolb A−a (2005)
Kolb A−b (2005) 2.42
linear regression test were adapted to examine the presence of publication bias (Peters et al., 2006). Results Characteristics of included studies
A total of 65 articles relevant to the searched keywords were initially identified. We reviewed the titles and abstracts of all the articles and 34 articles were excluded; full texts and data integrity were then reviewed and another 17 articles were finally excluded. Another three studies were also excluded due to lack of data integrity (Fig. 1). Finally, nine clinical case– control studies with a total of 491 PC patients and 481 healthy controls were selected for statistical analysis (Kolb et al., 2005; Koopmann et al., 2006; Yang et al., 2007, 2010; Chen et al., 2010; Zhang et al., 2010; Chen and Li, 2011; Zhu, 2011; Poruk et al., 2013). The range of publication year of eligible studies was from 2005 to 2013 (Fig. 2). Overall, five studies were conducted among Asians and the other four studies among Caucasians. Two kinds of genotyping methods were adopted in eligible studies, including enzyme-linked immunosorbent assay and electrochemiluminescence immunoassay. NOS scores of all included studies were ‡ 5. The main characteristics and methodological quality of all included studies are listed in Table 1. Quantitative data synthesis
Due to the presence of apparent heterogeneity between studies (all p < 0.05), we implemented the random-effects model. The results of the current meta-analysis revealed that the serum OPN level in PC patients was significantly higher than that in healthy subjects (SMD = 3.58, 95% CI = 2.42– 4.74, p < 0.001) (Fig. 3).
3.58
4.74
5.21
Ethnicity stratified analysis findings showed that the elevated serum OPN level was significantly related to the development in PC patients among both Caucasians and Asians (Caucasians: SMD = 2.62, 95% CI = 1.33–3.91, p < 0.001; Asians: SMD = 4.54, 95% CI = 2.80–6.27, p < 0.001; respectively) (Fig. 4). We also performed a subgroup analysis by sample size, cancer type, and detection method; we observed a significant association between the increased serum OPN level and the development of PC in all these groups (all p < 0.005) (Fig. 4). To assess the influence of each individual study on the pooled SMDs, we performed a sensitivity analysis by omitting individual studies. The results of sensitivity analysis
Serum OPN level (Case vs. Control)
8
(Egger’s test: t = 3.71, P = 0.006)
6
SMD
2.00
4
2
0 0
0.2
0.4
0.6
0.8
SE [SMD]
FIG. 6. Funnel plot of publication biases for the diagnosis of pancreatic cancer.
SERUM OPN LEVEL IN PC PATIENTS
indicated that the removal of any single study could not affect the overall pooled SMDs (Fig. 5). There was strong evidence for asymmetry presented in the funnel plots (Fig. 6). The presence of publication bias was also strongly demonstrated by Egger’s test (t = 3.71, p = 0.006). Discussion
We aimed to evaluate the potential significance of serum level of OPN for diagnostics. Our results showed that the serum OPN level in patients with PC was significantly higher than in the control participants, suggesting that the level of OPN may be essential as a therapeutic approach for predicting the invasion and metastasis of PC cells. It is noteworthy that OPN is a crucial adhesive bone matrix protein that plays an important role in immune cell recruitment, wound healing, and tissue remodeling (Takenaka et al., 2013). An increased expression of OPN is found to occur not only in pathologic conditions such as inflammation or ischemia but also in many types of malignancies (Sreekanthreddy et al., 2010; Rud et al., 2013). However, up to now, there is only limited information about the role of OPN in pancreatic diseases, not to mention the possible mechanisms of increased OPN expression and induction in PC. As a matter of fact, OPN itself contains several adhesive amino acid sequences, which would contribute to the interaction with cell surface receptors like CD44 and integrins (Ramaiah and Rittling, 2008). With respect to this, OPN is therefore known to play a critical role in cell adhesion and metastasis based on one of its binding domains on the arginyl-glycyl-aspartic acid sequence that interacts with integrins and the extracellular matrix (Narisawa et al., 2013). For the above reasons, the elevated level of OPN could stimulate cell migration and tumor growth by acting as a proangiogenic factor and by strengthening invasiveness and metastasis of tumor cells through binding to CD44 and certain integrins and, therefore, resulting in the invasion and metastasis of human tumor cells (Kolb et al., 2005). We therefore hypothesized that the intrinsic OPN excess production may be at least, in part, conducive to the aggressive and invasive behavior of human PC. This result agreed with the conclusions of the study conducted by Poruk et al. (2013) who showed a significant association between the increased serum level of OPN and the invasiveness of PC cells as well as the growth of metastases, supporting that OPN may serve as a serum marker for PC. Similarly, Kolb et al. (2005) also provided further evidence of the importance of OPN in the biology of PC and that blockade of OPN might be useful as a therapeutic approach to inhibit invasion and metastasis of PC cells. In our meta-analysis, we had dealt with numerous heterogeneity problems. Further subgroup analysis by ethnicity identified that the increased level of OPN may be significantly associated with a decreased overall survival rate in PC patients among both Caucasian and Asian populations. The present meta-analysis indicates that ethnicity difference may not have an impact on the heterogeneity of outcomes in the pathogenesis of PC. Although our meta-analysis was a practical way to generate a more powerful estimate of true effect size with less random error than individual studies, it did come with some potential limitations. First, we did not take into account
585
unpublished articles and abstracts since a myriad of needed information could not be successfully obtained. In this regard, due to the small number of studies, our results did not include all the data from all trials to evaluate the relationship of serum OPN level with the pathogenesis of PC. Nevertheless, our meta-analysis managed to overcome the limits of size or scope in individual studies to acquire more reliable and general information from each study. A second limitation of our meta-analysis is the fact that the results of meta-analysis might lack reliability to some extent since it is a retrospective study, which may induce potential publication bias. Especially, we enrolled eligible English studies, only which may relate to the occurrence of potential biases because of excluding parts of qualified studies based on language criteria. Another potential limitation is that our meta-analysis may still be underpowered to acquire original data from the included studies. Despite the above limitations, this is the first example of meta-analysis on the association of serum OPN level with the development of PC. More importantly, our meta-analysis uses a statistical approach to combine the results from multiple studies. To achieve strong objectivity, all the research methods were carried out using strict inclusion and exclusion criteria. Besides, inconsistency of results was rigorously quantified and analyzed in our meta-analysis, leading to a more reliable conclusion. In conclusion, the present meta-analysis suggests that an elevated serum OPN level may be used as a promising diagnostic tool for early identification of PC. Nevertheless, due to several limitations addressed previously, further clinical research8 with more integral data and larger sample size is warranted to obtain a more generally applicable statistical analysis. Acknowledgments
This work was supported by the Scientific Research Fund of Liaoning Province Education Department (No. L2010627). The authors would like to acknowledge the reviewers for their helpful comments on this article. Author Disclosure Statement
No competing financial interests exist. References
Anderson LN, Cotterchio M, Gallinger S (2009) Lifestyle, dietary, and medical history factors associated with pancreatic cancer risk in Ontario, Canada. Cancer Causes Control 20:825–834. Brand RE, Nolen BM, Zeh HJ, et al. (2011) Serum biomarker panels for the detection of pancreatic cancer. Clin Cancer Res 17:805–816. Chang IC, Chiang TI, Yeh KT, et al. (2010) Increased serum osteopontin is a risk factor for osteoporosis in menopausal women. Osteoporos Int 21:1401–1409. Chen R, Crispin DA, Pan S, et al. (2010) Pilot study of blood biomarker candidates for detection of pancreatic cancer. Pancreas 39:981–988. Chen ZG, Li QJ (2011) Association the detection of Opn, Cea, Ca125, and Ca19-9 and the diagnosis of pancreatic cancer. J Henan Med Col Staff Workers 23:181–182.
586
Cho H, Hong SW, Oh YJ, et al. (2008) Clinical significance of osteopontin expression in cervical cancer. J Cancer Res Clin Oncol 134:909–917. Cho HJ, Kim HS (2009) Osteopontin: a multifunctional protein at the crossroads of inflammation, atherosclerosis, and vascular calcification. Curr Atheroscler Rep 11:206–213. Ferrone CR, Brennan MF, Gonen M, et al. (2008) Pancreatic adenocarcinoma: the actual 5-year survivors. J Gastrointest Surg 12:701–706. Hidalgo M (2010) Pancreatic cancer. N Engl J Med 362:1605– 1617. Inoue M, Shinohara ML (2011) Intracellular osteopontin (iOPN) and immunity. Immunol Res 49:160–172. Iodice S, Gandini S, Maisonneuve P, et al. (2008) Tobacco and the risk of pancreatic cancer: a review and meta-analysis. Langenbecks Arch Surg 393:535–545. Kolb A, Kleeff J, Guweidhi A, et al. (2005) Osteopontin influences the invasiveness of pancreatic cancer cells and is increased in neoplastic and inflammatory conditions. Cancer Biol Ther 4:740–746. Koopmann J, Rosenzweig CN, Zhang Z, et al. (2006) Serum markers in patients with resectable pancreatic adenocarcinoma: macrophage inhibitory cytokine 1 versus CA19-9. Clin Cancer Res 12:442–446. Landi S (2009) Genetic predisposition and environmental risk factors to pancreatic cancer: a review of the literature. Mutat Res 681:299–307. Maisonneuve P, Lowenfels AB (2010) Epidemiology of pancreatic cancer: an update. Dig Dis 28:645–656. Mi Z, Bhattacharya SD, Kim VM, et al. (2011) Osteopontin promotes CCL5-mesenchymal stromal cell-mediated breast cancer metastasis. Carcinogenesis 32:477–487. Mi Z, Guo H, Russell MB, et al. (2009) RNA aptamer blockade of osteopontin inhibits growth and metastasis of MDAMB231 breast cancer cells. Mol Ther 17:153–161. Murugaiyan G, Mittal A, Weiner HL (2010) Identification of an IL-27/osteopontin axis in dendritic cells and its modulation by IFN-gamma limits IL-17-mediated autoimmune inflammation. Proc Natl Acad Sci U S A 107:11495–11500. Narisawa S, Yadav MC, Millan JL (2013) In vivo overexpression of tissue-nonspecific alkaline phosphatase increases skeletal mineralization and affects the phosphorylation status of osteopontin. J Bone Miner Res 28:1587–1598. Peters JL, Sutton AJ, Jones DR, et al. (2006) Comparison of two methods to detect publication bias in meta-analysis. JAMA 295:676–680. Poruk KE, Firpo MA, Scaife CL, et al. (2013) Serum osteopontin and tissue inhibitor of metalloproteinase 1 as diagnostic and prognostic biomarkers for pancreatic adenocarcinoma. Pancreas 42:193–197. Ramaiah SK, Rittling S (2008) Pathophysiological role of osteopontin in hepatic inflammation, toxicity, and cancer. Toxicol Sci 103:4–13. Rud AK, Boye K, Oijordsbakken M, et al. (2013) Osteopontin is a prognostic biomarker in non-small cell lung cancer. BMC Cancer 13:540. Scatena M, Liaw L, Giachelli CM (2007) Osteopontin: a multifunctional molecule regulating chronic inflammation and vascular disease. Arterioscler Thromb Vasc Biol 27:2302–2309.
LI ET AL.
Shang S, Plymoth A, Ge S, et al. (2012) Identification of osteopontin as a novel marker for early hepatocellular carcinoma. Hepatology 55:483–490. Siegel R, Naishadham D, Jemal A (2013) Cancer statistics, 2013. CA Cancer J Clin 63:11–30. Sreekanthreddy P, Srinivasan H, Kumar DM, et al. (2010) Identification of potential serum biomarkers of glioblastoma: serum osteopontin levels correlate with poor prognosis. Cancer Epidemiol Biomarkers Prev 19:1409–1422. Stang A (2010) Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses. Eur J Epidemiol 25:603–605. Sullivan J, Blair L, Alnajar A, et al. (2009) Expression of a prometastatic splice variant of osteopontin, OPNC, in human pancreatic ductal adenocarcinoma. Surgery 146:232–240. Takahashi F, Kumasaka T, Nagaoka T, et al. (2009) Osteopontin expression in pulmonary tumor thrombotic microangiopathy caused by gastric carcinoma. Pathol Int 59:752–756. Takenaka M, Hanagiri T, Shinohara S, et al. (2013) Serum level of osteopontin as a prognostic factor in patients who underwent surgical resection for non-small-cell lung cancer. Clin Lung Cancer 14:288–294. Wang KX, Denhardt DT (2008) Osteopontin: role in immune regulation and stress responses. Cytokine Growth Factor Rev 19:333–345. Yang SQ, Li XM, He F, et al. (2010) Clinical significance of detection of serum osteopontin in patient with pancreatic cancer. Labor Med Clin 7:1281–1282. Yang YC, Zhao YP, Liao Q, et al. (2007) Evaluation of osteopontin as serum tumor marker for pancreatic cancer. Chin J Practical Surg 27:821–823. Zhang X, Eirin A, Lerman A, et al. (2013) Osteopontin: an emerging therapeutic target in uraemic vascular disease. Cardiovasc Res 98:332–333. Zhang ZM, Guo JM, Wang XB, et al. (2010) Measurement of serum osteopontin in patients with pancreatic cancer and its clinical significance. Zhejiang Med J 32:644–645. Zhivkova-Galunska M, Adwan H, Eyol E, et al. (2010) Osteopontin but not osteonectin favors the metastatic growth of pancreatic cancer cell lines. Cancer Biol Ther 10:54–64. Zhu RQ (2011) Association the detection of Opn, Cea, Ca125, and Ca19-9 and the diagnosis of pancreatic cancer. Chin Manipulation Rehabil Med 2:49. Zintzaras E, Ioannidis JP (2005) HEGESMA: genome search meta-analysis and heterogeneity testing. Bioinformatics 21:3672–3673.
Address correspondence to: Jian-Jun Li, MD Department of Radiotherapy The First Affiliated Hospital China Medical University No. 155 Nanjing Street Heping District Shenyang 110001 People’s Republic of China E-mail: [email protected]
Diagnostic Significance of Serum Osteopontin Level for Pancreatic Cancer: A Meta-Analysis Jian-Jun Li,1 Hong-Yu Li,2 and Fei Gu1
Objective: This meta-analysis aimed to identify the significance of serum osteopontin (OPN) level for the diagnosis of pancreatic cancer (PC). Methods: Through searching the following electronic databases—the Cochrane Library Database (Issue 12, 2013), Web of Science (1945–2013), PubMed (1966–2013), CINAHL (1982–2013), EMBASE (1980–2013), and the Chinese Biomedical Database (CBM) (1982–2013)—related articles were determined without any language restrictions. The STATA statistical software (version 12.0; Stata Corporation, College Station, TX) was chosen to deal with statistical data. Standard mean difference (SMD) and its corresponding 95% confidence interval (95% CI) was calculated. Eleven clinical case–control studies, which recruited 491 PC patients and 481 healthy controls, were selected for statistical analysis. Results: Combined SMD of OPN suggested that the serum OPN level in PC patients was significantly higher than that in healthy controls (SMD = 3.58, 95% CI = 2.42–4.74, p < 0.001). Ethnicity stratified analysis indicated a higher serum OPN level in PC patients compared with control subjects among both Caucasians and Asians (Caucasians: SMD = 2.62, 95% CI = 1.33–3.91, p < 0.001; Asians: SMD = 4.54, 95% CI = 2.80–6.27, p < 0.001; respectively). Conclusion: The main finding of our meta-analysis revealed that an elevated serum OPN level may be used as a promising diagnostic tool for early identification of PC.
Introduction
H
uman pancreatic cancer (PC) is a malignant neoplasm originating from transformed cells arising in tissues forming the pancreas, which is widely regarded as the eighth or ninth most frequent cause of cancer death worldwide (Hidalgo, 2010; Maisonneuve and Lowenfels, 2010). The most common type of PC, accounting for 95% of these tumors, is adenocarcinoma arising within the exocrine component of the pancreas (Hidalgo, 2010). It has been estimated that an estimated 45,200 new invasive cases of PC were diagnosed in both sexes and about 38,460 deaths occurred in the United States in 2013 (Siegel et al., 2013). Furthermore, PC has an extremely poor prognosis, where the 5-year survival after resection of pancreatic adenocarcinoma (PDAC) was *12% and the 10-year relative survival was only 5%, which made PC an important global healthcare problem (Ferrone et al., 2008). In general, the pathogenesis of PC is complicated, and interactions between the genetic and environmental factors may serve as important predictors in the development of PC (Landi, 2009). A wide range of risk factors, including age, cigarette smoking, alcohol consump-
1 2
tion, unhealthy diet, obesity, and diabetes mellitus, have been demonstrated to be involved in PC development (Iodice et al., 2008; Anderson et al., 2009). Recently, epidemiological studies revealed that an increased serum level of osteopontin (OPN) was observed in PC patients, suggesting that the serum OPN level may be implicated in the pathogenesis of PC (Sullivan et al., 2009; Zhivkova-Galunska et al., 2010). The OPN protein, first identified as a bone matrix in 1986 in osteoblasts, is considered as an aspartic acid-rich, N-linked (SIBLING) glycosylated protein, which may be highly phosphorylated on serines and threonines according to the cell type (Scatena et al., 2007; Zhang et al., 2013). Acting as a phosphoglycoprotein produced by various tumor cells, OPN was initially characterized as a phosphoprotein secreted by transformed, malignant epithelial cells (Takahashi et al., 2009). Being one of the most critical extracellular structural proteins and an organic component of bone, OPN has various in vivo functions, such as cell adhesion and migration, immune and inflammatory response, apoptosis, and bone calcification (Ramaiah and Rittling, 2008; Wang and Denhardt, 2008). Evidence accumulated that OPN has been implicated in a number of physiological and pathological processes,
Department of Radiotherapy, The First Affiliated Hospital, China Medical University, Shenyang, People’s Republic of China. Department of Gastroenterology, General Hospital of Shenyang Military Area, Shenyang, People’s Republic of China.
580
SERUM OPN LEVEL IN PC PATIENTS
581
Identification
including fibrosis, neointima formation, arterial occlusion by thrombus, inflammation, and pulmonary hypertension (Cho and Kim, 2009; Inoue and Shinohara, 2011). Detection of plasma OPN level may therefore be useful in the treatment of autoimmune diseases, cancer metastasis, and bone mineralization diseases (Chang et al., 2010; Murugaiyan et al., 2010; Mi et al., 2011). Indeed, numerous studies have been shown that the increased level of OPN was detected in the development and progression of hepatocellular carcinoma, breast cancer, cervical cancer, gastric carcinoma, and so on (Cho et al., 2008; Mi et al., 2009; Shang et al., 2012). More importantly, the elevated serum OPN level could act as a biomarker for metastatic progression and poor survival in patients with pancreatic ductal adenocarcinoma (Brand et al., 2011). Therefore, through taking part in the promotion of cell growth, metastasis, and cell adhesion reduction, the elevated serum levels of OPN may be responsible for the invasion and metastasis of the PC cells and finally lead to the deterioration of tumors (Zhivkova-Galunska et al., 2010). In the past few decades, several previous studies have indicated that an increased serum OPN level may be a biomarker of PC based on its important effect on malignant cell proliferation, tumor metastasis, and survival (Koopmann et al., 2006; Yang et al., 2007); some results are not entirely consistent (Zhu, 2011; Poruk et al., 2013). In this regard, it is worth performing a
Articles identified through electronic database searching (N = 64)
meta-analysis of all available data to assess the value of serum OPN level for the diagnosis of PC patients. Materials and Methods Literature search and selection criteria
The identification of related articles was implemented through searching the following electronic databases without any language restrictions: PubMed, Springer Link, Karger Medical and Scientific Publishers, Chinese Biomedical Database (CBM), Chinese National Knowledge Infrastructure (CNKI), and Google scholar. We employed a search strategy of highly sensitivity in combination with the following keywords and MeSH terms: (‘‘osteopontin’’ or ‘‘OPN’’ or ‘‘sialoprotein 1’’ or ‘‘bone sialoprotein 1’’ or ‘‘secreted phosphoprotein 1’’ or ‘‘uropontin’’ or ‘‘Eta-1’’ or ‘‘2ar’’) and (‘‘pancreatic neoplasms’’ or ‘‘pancreatic cancer’’ or ‘‘pancreatic carcinoma’’ or ‘‘PC’’ or ‘‘pancreatic adenocarcinoma’’ or ‘‘pancreas cancer’’ or ‘‘pancreas carcinoma’’ or ‘‘pancreatic tumor’’). The references of the present articles and reviews were also manually searched for additional potential studies. The eligible studies included in our meta-analysis should meet the following three types of inclusion criteria: (1) must concern the diagnostic significance of serum OPN level for PC; (2) patients included in the current meta-analysis should
Additional articles identified through a manual search (N = 1)
Screening
Articles reviewed for duplicates (N = 65)
Articles after duplicates removed (N = 63) Studies were excluded, due to: (N = 8) Letters, reviews, meta-analysis (N = 11) Not human studies (N = 15) Not related to research topics
Eligibility
Full-text articles assessed for eligibility (N = 29) Studies were excluded, due to: (N = 2) Not case-control or cohort study (N = 5) Not relevant to serum osteopontin (N = 10) Not relevant to Pancreatic cancer
Included
Studies included in qualitative synthesis (N = 12)
Studies included in quantitative synthesis (meta-analysis) (N = 9)
FIG. 1. Flowchart of literature search and study selection. Nine studies were included in this metaanalysis.
582
LI ET AL. 20
Pubmed database All databases
FIG. 2. The distribution of topicrelated literature in the electronic database over the last decade.
Number of articles
15
10
5
0 2000~2001
2002~2003
2004~2005
2006~2007
2008~2009
2010~2011
2012~2013
Publication year
be diagnosed with cytologically or histologically proven PC or PDAC; (3) enough information on serum OPN level should be supplied in eligible articles. Articles that did not meet the inclusion criteria were excluded. If authors published several studies of the same subjects, either the most recent or largest sample size publication was included. Data extraction and methodological assessment
Two authors used a standardized form to independently collect relevant data from each included study. Controversial problems were resolved by discussion and consensus. By reviewing all the articles, two observers determined the inclusion and exclusion criteria. The language of publication, publication year of article, first author’s surname, geographical location, design of study, total number of cases, sample size, the source of controls, cancer type, detection method, and serum OPN level. Two observers separately assessed the methodological quality with the use of the Newcastle–Ottawa scale (NOS) criteria (Stang, 2010). The NOS criteria comprised three aspects: (1) subject selection: 0–4; (2) comparability of subject:
0–2; and (3) clinical outcome: 0–3. NOS scores ranged from 0 to 9, and a score with good quality should be ‡ 7. Statistical analysis
To achieve rigorous statistical analysis, we used the STATA statistical software (version 12.0; Stata Corporation, College Station, TX) to deal with statistical data. Standard mean difference (SMD) and its corresponding 95% confidence interval (95% CI) was calculated. The statistical significance of pooled SMDs was evaluated by the Z-test. Tests for heterogeneity are commonly used to decide on methods for combining studies and for selection of the random-effects model or fixed-effects model. Between-study heterogeneity was assessed by Cochran’s Q-statistic and I2 tests in our meta-analysis (Zintzaras and Ioannidis, 2005). A p-value < 0.05 or I2 > 50% means that these studies were heterogeneous, and then the random-effects model was employed; otherwise, the fixed-effects model was implemented. We also made use of subgroup analysis to explore reasons for the detection of heterogeneity. A sensitivity analysis was also implemented to evaluate the influence of a single study on the overall estimate. Funnel plots and Egger’s
Table 1. Main Characteristics and Methodological Quality of All Eligible Studies First author
Year
Sample size Gender (M/F) Cancer Ethnicity type Case Control Case Control
Poruk Zhu Chen Zhang Yang Chen Yang Koopmann Kolb
2013 2011 2011 2010 2010 2010 2007 2006 2005
Caucasian Asians Asians Asians Asians Caucasian Asians Caucasian Caucasian
PDAC PDAC PDAC PC PC PC PC PC PDAC
86 60 72 57 30 20 46 50 43 27
86 60 150 50 20 25 20 50 20 20
46/40 42/18 45/27 32/25 14/16 — 30/16 22/28 — —
46/40 45/15 — 28/22 10/10 — — 22/28 — —
Age (years) Case
Control
66 (44–87) 53 (35–75) 53 (32–81) 50.7 – 11.6 63 (39–82) — — 67.6 – 10.1 — —
44 (34–94) 51 (40–70) 51 (38–63) 49.1 – 10.5 63 (39–82) — — 67.6 – 10.1 — —
NOS Method Marker score ELISA ECLI ECLI ELISA ELISA ELISA ELISA ELISA ELISA ELISA
OPN OPN OPN OPN OPN OPN OPN OPN OPN OPN
8 8 7 8 6 5 5 7 5
ECLI, electrochemiluminescence immunoassay; ELISA, enzyme-linked immunosorbent assay; F, female; M, male; NOS, Newcastle– Ottawa scale; OPN, osteopontin; PC, pancreatic cancer; PDAC, pancreatic adenocarcinoma.
SERUM OPN LEVEL IN PC PATIENTS
583
Serum OPN level (Case vs. Control)
Included study
SMD (95% CI)
Weight%
Poruk KE (2013)
0.72 (0.41, 1.03)
10.48
Zhu RQ (2011)
3.08 (2.55, 3.61)
10.33
Chen ZG (2011)
3.37 (2.95, 3.79)
10.41
Zhang ZW (2010)
7.20 (6.16, 8.25)
9.73
Yang SQ (2010)
1.32 (0.69, 1.94)
10.25
Chen R (2010)
5.01 (3.80, 6.23)
9.47
Yang YC (2007)
8.30 (6.77, 9.83)
8.92
Koopmann J (2006)
1.09 (0.67, 1.52)
10.42
Kolb A−a (2005)
3.76 (2.91, 4.61)
9.99
2.98 (2.13, 3.82)
10.00
3.58 (2.42, 4.74)
100.00
Kolb A−b (2005) 2
Heterogeneity test (I = 97.4%, P < 0.001) Z test (Z = 6.05, P < 0.001) Random effects analysis −9.83
0
9.83
Serum OPN level
Serum OPN level
(Ethnicity: Case vs. Control)
Included study
FIG. 3. Forest plots for the significance of serum osteopontin (OPN) level for the diagnosis of pancreatic cancer.
SMD (95% CI) Weight%
(Sample size: Case vs. Control)
Included study
SMD (95% CI)
Weight%
Large
Caucasian Poruk KE (2013)
0.72 (0.41, 1.03)
10.48
Poruk KE (2013)
0.72 (0.41, 1.03)
10.48
Chen R (2010)
5.01 (3.80, 6.23)
9.47
Zhu RQ (2011)
3.08 (2.55, 3.61)
10.33
Koopmann J (2006)
1.09 (0.67, 1.52)
10.42
Chen ZG (2011)
3.37 (2.95, 3.79)
10.41
Kolb A−a (2005)
3.76 (2.91, 4.61)
9.99
Zhang ZW (2010)
7.20 (6.16, 8.25)
9.73
Kolb A−b (2005)
2.98 (2.13, 3.82)
10.00
Koopmann J (2006)
1.09 (0.67, 1.52)
10.42
2.62 (1.33, 3.91)
50.36
3.03 (1.45, 4.61)
51.37
Zhu RQ (2011)
3.08 (2.55, 3.61)
10.33
Yang SQ (2010)
1.32 (0.69, 1.94)
10.25
Chen ZG (2011)
3.37 (2.95, 3.79)
10.41
Chen R (2010)
5.01 (3.80, 6.23)
9.47
Zhang ZW (2010)
7.20 (6.16, 8.25)
9.73
Yang YC (2007)
8.30 (6.77, 9.83)
8.92
Yang SQ (2010)
1.32 (0.69, 1.94)
10.25
Kolb A−a (2005)
3.76 (2.91, 4.61)
9.99
Yang YC (2007)
8.30 (6.77, 9.83)
8.92
Kolb A−b (2005)
2.98 (2.13, 3.82)
10.00
49.64
(I 2
4.19 (2.25, 6.13)
48.63
(I 2 =
3.58 (2.42, 4.74)
100.00
Heterogeneity test (I2 = 95.9%, P < 0.001) Z test (Z = 3.99, P < 0.001)
Subtotal (I−squared = 98.3%, P < 0.001) Z test (Z = 3.77, P < 0.001) Small
Asians
(I2 =
4.54 (2.80, 6.27)
Heterogeneity test 97.0%, P < 0.001) Z test (Z = 5.12, P < 0.001) (I2 =
3.58 (2.42, 4.74)
Heterogeneity test 97.4%, P < 0.001) Z test (Z = 6.05, P < 0.001) Random effects analysis −9.83
0
100.00
= 95.4%, P < 0.001) Subtotal Z test (Z = 4.24, P < 0.001) Overall 97.4%, P < 0.001) Z test (Z = 6.05, P < 0.001) Random effects analysis −9.83
9.83
9.83
Serum OPN level
Serum OPN level (Cancer type: Case vs. Control)
Included study
0
SMD (95% CI) Weight%
PDAC
(Method: Case vs. Control)
Included study
SMD (95% CI) Weight%
ELISA
Poruk KE (2013)
0.72 (0.41, 1.03)
10.48
Poruk KE (2013)
0.72 (0.41, 1.03)
Zhu RQ (2011)
3.08 (2.55, 3.61)
10.33
Zhang ZW (2010)
7.20 (6.16, 8.25)
9.73
Chen ZG (2011)
3.37 (2.95, 3.79)
10.41
Yang SQ (2010)
1.32 (0.69, 1.94)
10.25
Kolb A−a (2005)
3.76 (2.91, 4.61)
9.99
Chen R (2010)
5.01 (3.80, 6.23)
9.47
09 Kolb A−b (2005)
2.98 (2.13, 3.82)
10.00
Yang YC (2007)
8.30 (6.77, 9.83)
8.92
2.76 (1.39, 4.14)
51.22
Koopmann J (2006)
1.09 (0.67, 1.52)
10.42
Kolb A−a (2005)
3.76 (2.91, 4.61)
9.99
Kolb A−b (2005)
2.98 (2.13, 3.82)
10.00
3.70 (2.23, 5.17)
79.25
Heterogeneity test (I2 = 97.2%, P < 0.001) Z test (Z = 3.94, P < 0.001) PC
Heterogeneity test (I 2 = 97.6%, P < 0.001) Z test (Z = 4.93, P < 0.001)
10.48
Zhang ZW (2010)
7.20 (6.16, 8.25)
9.73
Yang SQ (2010)
1.32 (0.69, 1.94)
10.25
Chen R (2010)
5.01 (3.80, 6.23)
9.47
Yang YC (2007)
8.30 (6.77, 9.83)
8.92
Zhu RQ (2011)
3.08 (2.55, 3.61)
10.33
Koopmann J (2006)
1.09 (0.67, 1.52)
10.42
Chen ZG (2011)
3.37 (2.95, 3.79)
10.41
4.52 (1.97, 7.08)
48.78
3.26 (2.93, 3.59)
20.75
3.58 (2.42, 4.74)
100.00
Heterogeneity test (I2 = 98.0%, P < 0.001) Z test (Z = 3.47, P = 0.001) (I2 =
3.58 (2.42, 4.74)
Heterogeneity test 97.4%, P < 0.001) Z test (Z = 6.05, P < 0.001) Random effects analysis −9.83
0
9.83
100.00
ECLI
Heterogeneity test (I 2 = 0.0%, P < 0.001) Z test (Z = 19.32, P < 0.001) (I 2 =
Heterogeneity test 97.4%, P < 0.001) Z test (Z = 6.05, P < 0.001) Random effects analysis −9.83
0
9.83
FIG. 4. Subgroup analyses by ethnicity, sample size, cancer type, and detection method for the diagnosis of pancreatic cancer.
584
LI ET AL.
Serum OPN level (Case vs. Control) Lower CI Limit
Estimate
Upper CI Limit
Poruk KE (2013)
Zhu RQ (2011)
Chen ZG (2011)
Zhang ZW (2010)
FIG. 5. Sensitivity analysis of the summary odds ratio coefficients for the diagnosis of pancreatic cancer.
Yang SQ (2010)
Chen R (2010)
Yang YC (2007)
Koopmann J (2006)
Kolb A−a (2005)
Kolb A−b (2005) 2.42
linear regression test were adapted to examine the presence of publication bias (Peters et al., 2006). Results Characteristics of included studies
A total of 65 articles relevant to the searched keywords were initially identified. We reviewed the titles and abstracts of all the articles and 34 articles were excluded; full texts and data integrity were then reviewed and another 17 articles were finally excluded. Another three studies were also excluded due to lack of data integrity (Fig. 1). Finally, nine clinical case– control studies with a total of 491 PC patients and 481 healthy controls were selected for statistical analysis (Kolb et al., 2005; Koopmann et al., 2006; Yang et al., 2007, 2010; Chen et al., 2010; Zhang et al., 2010; Chen and Li, 2011; Zhu, 2011; Poruk et al., 2013). The range of publication year of eligible studies was from 2005 to 2013 (Fig. 2). Overall, five studies were conducted among Asians and the other four studies among Caucasians. Two kinds of genotyping methods were adopted in eligible studies, including enzyme-linked immunosorbent assay and electrochemiluminescence immunoassay. NOS scores of all included studies were ‡ 5. The main characteristics and methodological quality of all included studies are listed in Table 1. Quantitative data synthesis
Due to the presence of apparent heterogeneity between studies (all p < 0.05), we implemented the random-effects model. The results of the current meta-analysis revealed that the serum OPN level in PC patients was significantly higher than that in healthy subjects (SMD = 3.58, 95% CI = 2.42– 4.74, p < 0.001) (Fig. 3).
3.58
4.74
5.21
Ethnicity stratified analysis findings showed that the elevated serum OPN level was significantly related to the development in PC patients among both Caucasians and Asians (Caucasians: SMD = 2.62, 95% CI = 1.33–3.91, p < 0.001; Asians: SMD = 4.54, 95% CI = 2.80–6.27, p < 0.001; respectively) (Fig. 4). We also performed a subgroup analysis by sample size, cancer type, and detection method; we observed a significant association between the increased serum OPN level and the development of PC in all these groups (all p < 0.005) (Fig. 4). To assess the influence of each individual study on the pooled SMDs, we performed a sensitivity analysis by omitting individual studies. The results of sensitivity analysis
Serum OPN level (Case vs. Control)
8
(Egger’s test: t = 3.71, P = 0.006)
6
SMD
2.00
4
2
0 0
0.2
0.4
0.6
0.8
SE [SMD]
FIG. 6. Funnel plot of publication biases for the diagnosis of pancreatic cancer.
SERUM OPN LEVEL IN PC PATIENTS
indicated that the removal of any single study could not affect the overall pooled SMDs (Fig. 5). There was strong evidence for asymmetry presented in the funnel plots (Fig. 6). The presence of publication bias was also strongly demonstrated by Egger’s test (t = 3.71, p = 0.006). Discussion
We aimed to evaluate the potential significance of serum level of OPN for diagnostics. Our results showed that the serum OPN level in patients with PC was significantly higher than in the control participants, suggesting that the level of OPN may be essential as a therapeutic approach for predicting the invasion and metastasis of PC cells. It is noteworthy that OPN is a crucial adhesive bone matrix protein that plays an important role in immune cell recruitment, wound healing, and tissue remodeling (Takenaka et al., 2013). An increased expression of OPN is found to occur not only in pathologic conditions such as inflammation or ischemia but also in many types of malignancies (Sreekanthreddy et al., 2010; Rud et al., 2013). However, up to now, there is only limited information about the role of OPN in pancreatic diseases, not to mention the possible mechanisms of increased OPN expression and induction in PC. As a matter of fact, OPN itself contains several adhesive amino acid sequences, which would contribute to the interaction with cell surface receptors like CD44 and integrins (Ramaiah and Rittling, 2008). With respect to this, OPN is therefore known to play a critical role in cell adhesion and metastasis based on one of its binding domains on the arginyl-glycyl-aspartic acid sequence that interacts with integrins and the extracellular matrix (Narisawa et al., 2013). For the above reasons, the elevated level of OPN could stimulate cell migration and tumor growth by acting as a proangiogenic factor and by strengthening invasiveness and metastasis of tumor cells through binding to CD44 and certain integrins and, therefore, resulting in the invasion and metastasis of human tumor cells (Kolb et al., 2005). We therefore hypothesized that the intrinsic OPN excess production may be at least, in part, conducive to the aggressive and invasive behavior of human PC. This result agreed with the conclusions of the study conducted by Poruk et al. (2013) who showed a significant association between the increased serum level of OPN and the invasiveness of PC cells as well as the growth of metastases, supporting that OPN may serve as a serum marker for PC. Similarly, Kolb et al. (2005) also provided further evidence of the importance of OPN in the biology of PC and that blockade of OPN might be useful as a therapeutic approach to inhibit invasion and metastasis of PC cells. In our meta-analysis, we had dealt with numerous heterogeneity problems. Further subgroup analysis by ethnicity identified that the increased level of OPN may be significantly associated with a decreased overall survival rate in PC patients among both Caucasian and Asian populations. The present meta-analysis indicates that ethnicity difference may not have an impact on the heterogeneity of outcomes in the pathogenesis of PC. Although our meta-analysis was a practical way to generate a more powerful estimate of true effect size with less random error than individual studies, it did come with some potential limitations. First, we did not take into account
585
unpublished articles and abstracts since a myriad of needed information could not be successfully obtained. In this regard, due to the small number of studies, our results did not include all the data from all trials to evaluate the relationship of serum OPN level with the pathogenesis of PC. Nevertheless, our meta-analysis managed to overcome the limits of size or scope in individual studies to acquire more reliable and general information from each study. A second limitation of our meta-analysis is the fact that the results of meta-analysis might lack reliability to some extent since it is a retrospective study, which may induce potential publication bias. Especially, we enrolled eligible English studies, only which may relate to the occurrence of potential biases because of excluding parts of qualified studies based on language criteria. Another potential limitation is that our meta-analysis may still be underpowered to acquire original data from the included studies. Despite the above limitations, this is the first example of meta-analysis on the association of serum OPN level with the development of PC. More importantly, our meta-analysis uses a statistical approach to combine the results from multiple studies. To achieve strong objectivity, all the research methods were carried out using strict inclusion and exclusion criteria. Besides, inconsistency of results was rigorously quantified and analyzed in our meta-analysis, leading to a more reliable conclusion. In conclusion, the present meta-analysis suggests that an elevated serum OPN level may be used as a promising diagnostic tool for early identification of PC. Nevertheless, due to several limitations addressed previously, further clinical research8 with more integral data and larger sample size is warranted to obtain a more generally applicable statistical analysis. Acknowledgments
This work was supported by the Scientific Research Fund of Liaoning Province Education Department (No. L2010627). The authors would like to acknowledge the reviewers for their helpful comments on this article. Author Disclosure Statement
No competing financial interests exist. References
Anderson LN, Cotterchio M, Gallinger S (2009) Lifestyle, dietary, and medical history factors associated with pancreatic cancer risk in Ontario, Canada. Cancer Causes Control 20:825–834. Brand RE, Nolen BM, Zeh HJ, et al. (2011) Serum biomarker panels for the detection of pancreatic cancer. Clin Cancer Res 17:805–816. Chang IC, Chiang TI, Yeh KT, et al. (2010) Increased serum osteopontin is a risk factor for osteoporosis in menopausal women. Osteoporos Int 21:1401–1409. Chen R, Crispin DA, Pan S, et al. (2010) Pilot study of blood biomarker candidates for detection of pancreatic cancer. Pancreas 39:981–988. Chen ZG, Li QJ (2011) Association the detection of Opn, Cea, Ca125, and Ca19-9 and the diagnosis of pancreatic cancer. J Henan Med Col Staff Workers 23:181–182.
586
Cho H, Hong SW, Oh YJ, et al. (2008) Clinical significance of osteopontin expression in cervical cancer. J Cancer Res Clin Oncol 134:909–917. Cho HJ, Kim HS (2009) Osteopontin: a multifunctional protein at the crossroads of inflammation, atherosclerosis, and vascular calcification. Curr Atheroscler Rep 11:206–213. Ferrone CR, Brennan MF, Gonen M, et al. (2008) Pancreatic adenocarcinoma: the actual 5-year survivors. J Gastrointest Surg 12:701–706. Hidalgo M (2010) Pancreatic cancer. N Engl J Med 362:1605– 1617. Inoue M, Shinohara ML (2011) Intracellular osteopontin (iOPN) and immunity. Immunol Res 49:160–172. Iodice S, Gandini S, Maisonneuve P, et al. (2008) Tobacco and the risk of pancreatic cancer: a review and meta-analysis. Langenbecks Arch Surg 393:535–545. Kolb A, Kleeff J, Guweidhi A, et al. (2005) Osteopontin influences the invasiveness of pancreatic cancer cells and is increased in neoplastic and inflammatory conditions. Cancer Biol Ther 4:740–746. Koopmann J, Rosenzweig CN, Zhang Z, et al. (2006) Serum markers in patients with resectable pancreatic adenocarcinoma: macrophage inhibitory cytokine 1 versus CA19-9. Clin Cancer Res 12:442–446. Landi S (2009) Genetic predisposition and environmental risk factors to pancreatic cancer: a review of the literature. Mutat Res 681:299–307. Maisonneuve P, Lowenfels AB (2010) Epidemiology of pancreatic cancer: an update. Dig Dis 28:645–656. Mi Z, Bhattacharya SD, Kim VM, et al. (2011) Osteopontin promotes CCL5-mesenchymal stromal cell-mediated breast cancer metastasis. Carcinogenesis 32:477–487. Mi Z, Guo H, Russell MB, et al. (2009) RNA aptamer blockade of osteopontin inhibits growth and metastasis of MDAMB231 breast cancer cells. Mol Ther 17:153–161. Murugaiyan G, Mittal A, Weiner HL (2010) Identification of an IL-27/osteopontin axis in dendritic cells and its modulation by IFN-gamma limits IL-17-mediated autoimmune inflammation. Proc Natl Acad Sci U S A 107:11495–11500. Narisawa S, Yadav MC, Millan JL (2013) In vivo overexpression of tissue-nonspecific alkaline phosphatase increases skeletal mineralization and affects the phosphorylation status of osteopontin. J Bone Miner Res 28:1587–1598. Peters JL, Sutton AJ, Jones DR, et al. (2006) Comparison of two methods to detect publication bias in meta-analysis. JAMA 295:676–680. Poruk KE, Firpo MA, Scaife CL, et al. (2013) Serum osteopontin and tissue inhibitor of metalloproteinase 1 as diagnostic and prognostic biomarkers for pancreatic adenocarcinoma. Pancreas 42:193–197. Ramaiah SK, Rittling S (2008) Pathophysiological role of osteopontin in hepatic inflammation, toxicity, and cancer. Toxicol Sci 103:4–13. Rud AK, Boye K, Oijordsbakken M, et al. (2013) Osteopontin is a prognostic biomarker in non-small cell lung cancer. BMC Cancer 13:540. Scatena M, Liaw L, Giachelli CM (2007) Osteopontin: a multifunctional molecule regulating chronic inflammation and vascular disease. Arterioscler Thromb Vasc Biol 27:2302–2309.
LI ET AL.
Shang S, Plymoth A, Ge S, et al. (2012) Identification of osteopontin as a novel marker for early hepatocellular carcinoma. Hepatology 55:483–490. Siegel R, Naishadham D, Jemal A (2013) Cancer statistics, 2013. CA Cancer J Clin 63:11–30. Sreekanthreddy P, Srinivasan H, Kumar DM, et al. (2010) Identification of potential serum biomarkers of glioblastoma: serum osteopontin levels correlate with poor prognosis. Cancer Epidemiol Biomarkers Prev 19:1409–1422. Stang A (2010) Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses. Eur J Epidemiol 25:603–605. Sullivan J, Blair L, Alnajar A, et al. (2009) Expression of a prometastatic splice variant of osteopontin, OPNC, in human pancreatic ductal adenocarcinoma. Surgery 146:232–240. Takahashi F, Kumasaka T, Nagaoka T, et al. (2009) Osteopontin expression in pulmonary tumor thrombotic microangiopathy caused by gastric carcinoma. Pathol Int 59:752–756. Takenaka M, Hanagiri T, Shinohara S, et al. (2013) Serum level of osteopontin as a prognostic factor in patients who underwent surgical resection for non-small-cell lung cancer. Clin Lung Cancer 14:288–294. Wang KX, Denhardt DT (2008) Osteopontin: role in immune regulation and stress responses. Cytokine Growth Factor Rev 19:333–345. Yang SQ, Li XM, He F, et al. (2010) Clinical significance of detection of serum osteopontin in patient with pancreatic cancer. Labor Med Clin 7:1281–1282. Yang YC, Zhao YP, Liao Q, et al. (2007) Evaluation of osteopontin as serum tumor marker for pancreatic cancer. Chin J Practical Surg 27:821–823. Zhang X, Eirin A, Lerman A, et al. (2013) Osteopontin: an emerging therapeutic target in uraemic vascular disease. Cardiovasc Res 98:332–333. Zhang ZM, Guo JM, Wang XB, et al. (2010) Measurement of serum osteopontin in patients with pancreatic cancer and its clinical significance. Zhejiang Med J 32:644–645. Zhivkova-Galunska M, Adwan H, Eyol E, et al. (2010) Osteopontin but not osteonectin favors the metastatic growth of pancreatic cancer cell lines. Cancer Biol Ther 10:54–64. Zhu RQ (2011) Association the detection of Opn, Cea, Ca125, and Ca19-9 and the diagnosis of pancreatic cancer. Chin Manipulation Rehabil Med 2:49. Zintzaras E, Ioannidis JP (2005) HEGESMA: genome search meta-analysis and heterogeneity testing. Bioinformatics 21:3672–3673.
Address correspondence to: Jian-Jun Li, MD Department of Radiotherapy The First Affiliated Hospital China Medical University No. 155 Nanjing Street Heping District Shenyang 110001 People’s Republic of China E-mail: [email protected]
