- Email: [email protected]
Tumor-stroma ratio(TSR) as a potential novel predictor of prognosis in digestive system cancers: A meta-analysis
Clinica Chimica Acta 472 (2017) 64–68
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/cca
Tumor-stroma ratio(TSR) as a potential novel predictor of prognosis in digestive system cancers: A meta-analysis
MARK
Runjin Zhang1, Wei Song1, Kai Wang, Shubing Zou⁎ Department of hepatobiliary surgery, The Second Affiliated Hospital of Nanchang University, Nanchang, China
A R T I C L E I N F O
A B S T R A C T
Keywords: Tumor-stroma ratio (TSR) Digestive system cancers Prognosis
Main problem: The tumor-stroma ratio (TSR) has been reported as a prognosis predictor in multiple cancers. The aim of this meta-analysis was to investigate the potential value of TSR as a prognostic predictor of cancer in the digestive system. Methods: We searched PubMed, Embase, Elsevier and Web of Science. All studies exploring the association of TSR with overall survival (OS) or disease-free survival (DFS), and lymph node metastasis (LNM) were identified. Results: In total, eight studies were eligible for analysis, and they included 1959 patients. Meta-analysis showed that the low TSR in the tumor could predict poor overall survival (OS) in multiple cancers (pooled Hazard Ratio [HR]: 2.15, 95%CI: 1.80–2.57, P < 0.00001, fixed effects). For disease-free survival (DFS), low TSR was also a significant predictor (pooled Hazard Ratio [HR]: 2.31, 95%CI: 1.88–2.83, P < 0.00001, fixed effects). In addition, low TSR was correlated with tumor stage. Discussion: The tumor-stroma ratio (TSR) may potentially serve as a poor prognostic predictor for the metastasis and prognosis of cancer.
1. Introduction
2. Evidence acquisition
Cancer is a large public health problem and a major cause of deaths worldwide. In 2015, 1,658,370 individuals developed cancer, and 589,430 people died of cancer in the United States, according to the American National Center for Health Statistics [1]. Furthermore, digestive system cancers are the deadliest forms of cancer, including many organs cancer in our body [2]. Reflecting the complex initiation and progression mechanism, the 5-year survival rate is low in most tumors. Consequently, we developed a novel prognostic marker to assist clinical application in multiple cancers. The tumor-stroma ratio (TSR) is the proportion of tumor cells in the stroma. Tumors are complex tissues that are composed of carcinoma cells and surrounding stroma; thus, the tumor stroma is associated with tumor initiation, progression, and metastasis, and it holds prognostic value [3–4]. Recently, the tumor to stroma ratio of (TSR) has been demonstrated to be a novel and practical prognostic predictor in many neoplasms, such as breast cancer [5], esophageal cancer [6], colon cancer [7], and hepatocellular carcinoma [8]. Those studies have shown that a low TSR is a poor prognostic marker [5.6.7.8]. So we need performed a meta-analysis to explore the relation of TSR with OS and DFS.
2.1. Search strategies
⁎
1
Corresponding author. E-mail address: [email protected] (S. Zou). These authors have contributed equally to this work.
http://dx.doi.org/10.1016/j.cca.2017.07.003 Received 23 February 2017; Received in revised form 3 July 2017; Accepted 3 July 2017 Available online 17 July 2017 0009-8981/ © 2017 Published by Elsevier B.V.
We searched MEDLINE, Embase, Elsevier and Web of science. Search terms included “carcinoma-stroma ratio (CSR)”, “tumor-stroma ratio (TSR)”, “tumor-stroma percentage (TSP)”, “cancer”, “tumor”, “carcinoma”, “neoplasms”, “survival”, “prognosis”, and “outcome”. Systematic reviews and meta-analyses were manually retrieved and served as references for the included studies. The language was restricted to English. The literature search was completed on November 31, 2016. 2.2. Inclusion criteria The inclusion criteria included 1) a clear relationship between TSR and overall survival or TSR and disease-free survival; 2) multivariate analysis as the analysis method; 3) studies reported on hazard ratio (HR) and 95% confidence interval (CI); and 4) studies published in English.
Clinica Chimica Acta 472 (2017) 64–68
R. Zhang et al.
and were included in the meta-analysis [6–8,10–14]; this process is shown in Fig. 1.
2.3. Exclusion criteria The exclusion criteria included 1) metastatic cancers in organs not related to the primary cancers; 2) studies investigating the structure, mechanism and functions; 3) studies in which it was impossible to obtain available data; 4) duplicate publications; and 5) non-Englishlanguage studies.
3.2. Characteristics of the included studies A total of eight studies that included 1959 patients were enrolled in the present meta-analysis. The number of patients in each study ranged from 81 to 710. Among these studies, three studies were conducted in the UK, two were conducted in the Netherlands, two were conducted in China, and one was conducted in the Republic of Korea. Among these eight studies, four studies on colorectal cancer, one study on esophageal adenocarcinoma, one study on esophageal squamous cell carcinoma, one study on hepatocellular carcinoma, and one study on gastric signet ring cell carcinoma were observed. In total, eight studies focused on OS, and five studies focused on DFS. The hazard ratios (HRs) and corresponding 95% confidence intervals were obtained though multivariate analysis. Most of studies were divided into two groups: high stroma ratio and low stroma ratio of TSR, and the cutoff values were all set to 50%. Concerning the methodological quality of these studies, the NOS scores of all included studies were ≥6. Detailed patient characteristics are shown in Table 1.
2.4. Data extraction According to the aforementioned inclusion and exclusion criteria, two investigators reviewed all eligible studies and extracted the necessary data. Any disagreement was resolved by discussions among all coauthors. The following information was collected: name of the first author, year of publication, country, number of patients, tumor type, stage, overall survival (OS), disease-free survival (DFS), analysis methods, hazard ratio (HR) with 95% confidence interval (CI), and follow-ups. 2.5. Statistical analysis The meta-analysis was conducted using the RevMan version 5.3 software (Cochrane Collaboration). Heterogeneity was measured using Q and I2 tests, and I2 > 50% and P < 0.10 indicated heterogeneity [9]. When I2 < 50%, no statistically significant heterogeneity was considered. The fixed-effects model was used to pool the results, and the random-effects model was used. Pooled hazard ratios (HRs) were extracted from the published data; the log HR and standard error (SE) were used to aggregate the survival results. P < 0.05 was defined as statistically significant.
4. Meta-analysis 4.1. Association between TSR and OS Eight eligible studies showed that the low TSR, reflecting rich stroma, was associated with the high risk of OS (pooled Hazard Ratio [HR]: 2.15, 95%CI: 1.80–2.57, P < 0.00001, fixed effects, I2 = 0%, Ph = 0.58) (Table 3, Fig. 2). In the subgroup analysis, potential heterogeneity was detected using the region, sample size, cancer type, clinical stage, and NOS score (Table 2). And we found that rich stroma had significantly poor prognostic in all subgroups.
3. Evidence synthesis 3.1. Search results
4.2. Association between TSR and DFS A total of 322 articles were found in the selected databases. A total of 25 articles were excluded as duplication, and 282 articles were excluded after screening the titles and abstracts, leaving 15 articles for further evaluation. As a result, 8 articles fulfilled the inclusion criteria
Five studies described the relation between TSR and DFS. An analysis of 1402 patients to evaluate this association showed that a low TSR was significantly correlated with the poor outcome of DFS (pooled
Fig. 1. Flow diagram of the study selection process. 332 studies were preretrieved in accordance with the established search strategies. Of these articles, 25 articles were excluded due to duplication and 282 were excluded by screening titles and abstracts, leaving 15 articles for further evaluation. As a result, 8 articles, fulfilled the inclusion criteria, were involved in the meta-analysis.
65
Clinica Chimica Acta 472 (2017) 64–68
6
6
7
6
7
7
8
6
≥ 50% I–III 95 Until Mar 2011 2007 Asia 82/13 60 (42–77) China Wang.K 2012
CC UK Park.JH 2014
SRCC Korea Lee D 2016
HCC China Lv Z 2015
CC UK West.NP 2010
OC
≥ 50% 331 160/171
Europe
1997–2008
NR
I–III
≥ 50% II–III 175 Median 72.5 2005–2008 Asia 94/81
≥ 50% I–IV 300 Until Aug 2014 2006–2013 Asia 262/38
≥ 47% I–IV 145 Median 52 (26–74) 1990–1995 Europe 58/87
≥ 50% II–III 710 Median 58 (2–76) Europe UK Huijbers.A 2013
CC
65 (25–86) 69 (61–76) 54 (27–80) 49 (23–85) NR
438/272
2002–2004
≥ 50% I–IV 81 Median 23 (3–220) 1990–2004 Europe 66/15 64(37–82) OC Netherlands Courrech Staal E.F 2011
We identified several elemental factors to explore the relationship of TSR with clinicopathological features (Table 3). The low TSR in tumor tissue was associated with the size and stage of the tumor. Pooled data revealed that a low TSR was significantly associated with tumor stage (I + II/III + IV, OR: 0.47, 95%CI: 0.26–0.84, P = 0.01, random effects, I2 = 70, Ph = 0.005). In contrast, no significant association was discovered in gender (male/female), differentiation (moderate + well/ poor), tumor invasion (T1 + T2/T3 + T4), lymph node metastasis (pos/neg), and venous invasion (pos/neg).
CC: colon cancer or colorectal cancer; OC: oesophageal cancer; HCC: hepatocellular carcinoma; SRCC: signet ring cell carcinomas; OS: overall survival; DFS: disease-free survival; MV: multivariate; NR: none reported.
MV Reported
MV
MV Reported
Reported OS DFS
MV
MV Reported
Reported
OS DFS OS DFS OS
MV Reported
MV Reported
OS FDS OS
MV Reported
OS DFS OS
33 27.0% 12 14.8% 207 29.2% 35 24.1% 75 25.0% 64 36.6% 250 75.5% 30 31.6% Netherlands Mesker WE 2007
CC
NR
72/50
Europe
1980–2001
NR
122
I–III
≥ 50%
HR estimate Stroma rich (%) Cut off Clinical stage No. of patients Follow-up (months) Duration Region Gender (M/F) Age (years) Cancer type Country Study (author, years)
Table 1 Characteristics of the eligible studies in the meta-analysis.
Hazard Ratio [HR]: 2.31, 95%CI: 1.88–2.83, P < 0.001; fixed effects, I2 = 0%, Ph = 0.57) (Fig. 3). 4.3. Correlation between TSR and clinicopathological features
Survival analysis
Analysis
Quality
R. Zhang et al.
5. Discussion The digestive system, which is one of the most complicated systems in the human body. Different locations in the digestive system show different types of cancers. Therefore, the tumor mechanism in the digestive system during development is extremely complex. Thus, we need a simple, convenient, and effective method to predict the prognosis in digestive system tumors. The TSR (ratio value of tumor to stroma in tumor tissue) was first reported to determine the prognosis of colorectal cancer in 2007 [12] as a simple, cheap, convenient and useful method in clinical studies. Thereafter, increasing studies have shown the value of the TSR as a prognosis predictor in various cancers. In these studies, the patients were divided into two groups based on a cutoff value of TSR = 50%, which was considered the most significant prognostic value. However, the prognostic value of TSR in digestive system tumor remains unclear. The present study is a meta-analysis to explore the correlation between TSR and overall survival in digestive system cancers. The present meta-analysis, which included eight studies and 1959 patients, showed that a low TSR was associated with poor overall survival in digestive system tumors (pooled Hazard Ratio [HR]: 2.15, 95%CI: 1.80–2.57, P < 0.00001, fixed effects, I2 = 0%, Ph = 0.58). When the analysis was restricted to the survival outcome of DFS, a positive result was observed in the present meta-analysis (pooled Hazard Ratio [HR]: 2.31, 95%CI: 1.88–2.83, P < 0.001; fixed effects). Thus, TSR could also be an independent predictor of OS/DFS in patients with digestive system cancers. The meta-analysis was shown a low heterogeneity of OS (I2 = 0%, Ph = 0.58) and DFS (I2 = 0%, Ph = 0.57). If necessary, a multicenter analysis and large patient sample size were used to study the problems in digestive system cancer. We also analyzed the association between TSR and the clinicopathological of OS on digestive system cancers. However, we still do not fully understand the mechanism underlying the promoting effect of stroma in tumors. According to the results, a high proportion of stroma was significantly associated with tumor size and tumor stage, which reflects poor prognosis in digestive system cancers. The formation of tumor-activated stroma results in the disruption of the epithelial tissue, the immune evasion of malignant cells, and the invasion of tumors, regarded as tumor stromatogenesis [15]. In normal tissue, the stroma acts as a protective area in tumor stromatogenesis. However, in tumor tissue, the stroma is a major part of the tumor microenvironment that could accelerate tumor progression [16]. The main composition of extracellular matrix (ECM) in the tumor stroma can change the microenvironment of tumors to promote cancer cell metastasis [17], such as the overexpression of amino acids, which are rich in cysteine [18], and matrix metalloproteinases [19], which can facilitate tumorigenesis and invasion. Cancer-associated fibroblasts (CAF), which are the predominant cells within the tumor tissue, are key regulators between stroma and cancer cells [20]. A variety of soluble factors that are secreted by CAFs act in a paracrine manner and affect not only cancer cells but also other cell types [21]. In different periods, 66
Clinica Chimica Acta 472 (2017) 64–68
R. Zhang et al.
Fig. 2. Forest plots of the overall survival (OS). Hazard ratios (HRs) for each trial are represented by the squares, and the horizontal lines crossing the square stand for the 95% confidence intervals (CIs). The diamonds represent the estimated pooled effect of the overall outcome for OS. Table 2 Pooled hazard ratios(HRs) for OS according to subgroup analysis. Subgroup
No. of studies
No. of patients
Effects model
HR(95%CI)
Z
P
I2(%)
Ph
Overall Region Asia Europe Sample size < 100 ≥ 100 Cancer type CC HCC OC SRCC Clinical stage I–III I–IV II–III NOS ≥7 <6
8
1959
Fixed
2.15(1.80, 2.57)
8.39
< 0.00001
0
0.58
3 5
570 1389
Fixed Fixed
2.77(2.01, 3.84) 1.92(1.55, 2.38)
6.17 5.98
< 0.00001 < 0.00001
0 0
0.69 0.83
2 6
176 1783
Fixed Fixed
2.62(1.74, 3.94) 2.05(1.68, 2.50)
4.64 7.07
< 0.00001 < 0.00001
37 0
0.21 0.71
4 1 2 1
1308 300 176 175
Fixed – Fixed –
1.90(1.51, 2.55(1.44, 2.62(1.74, 2.50(1.48,
2.39) 4.52) 3.94) 4.24)
5.44 3.21 4.64 3.4
< 0.00001 0.001 < 0.00001 0.0007
0 – 37 –
0.70 – 0.21 –
3 3 2
548 526 885
Fixed Fixed Fixed
2.39(1.76, 3.26) 2.22(1.58, 3.13) 1.91(1.44, 2.54)
5.54 4.59 4.46
< 0.00001 < 0.00001 < 0.00001
27 0 30
0.25 0.84 0.23
4 4
1111 848
Fixed Fixed
1.96(1.54, 2.48) 2.43(1.85, 3.19)
5.56 6.40
< 0.00001 < 0.00001
0 0
0.68 0.43
CC: colon cancer or colorectal cancer; OC: oesophageal cancer; HCC: hepatocellular carcinoma; SRCC: signet ring cell carcinomas; NOS: Newcastle-Ottawa Scale.
Fig. 3. Forest plots of the disease-free survival (DFS). Hazard ratios (HRs) for each trial are represented by the squares, and the horizontal lines crossing the square stand for the 95% confidence intervals (CIs). The diamonds represent the estimated pooled effect of the overall outcome for DFS.
cells have frequently been associated with the development of highgrade malignancies and poor prognoses [26]. Although this meta-analysis revealed a positive conclusion, it does have several limitations. Firstly, this study found a low heterogeneity, but it does have many biases. One of the biases is the method to evaluate the TSR. Although those researchers have standard evaluation methods, but it still can't avoid personal subjectivity. So we need a more scientific and rigorous method to evaluate the TSR. Secondly, the sample size was small. Only eight studies, involving 1595 patients, were
CAFs also play different roles. CAFs can inhibit cancer cells by enhancing the formation of gap junctions during the early stages. However, during later stages, CAFs play an accelerated role in tumor growth and progression through several tumor-secreted factors [22,23]. Moreover, many growth factors and chemokines, such as nuclear factor κB [24] and tumor necrosis factor α [25], are secreted into the stroma by stromal cells or cancer cells and act as chemoattractants for other noncancer cells. The mechanism by which stroma cells affect cancer growth and progression remains incompletely understood. However, stroma 67
Clinica Chimica Acta 472 (2017) 64–68
R. Zhang et al.
Table 3 Meta-analysis of the tumor-stroma ratio and clinicopathological features of digestive system cancer. Characteristics
No. of studies
No. of patients
OR(95%CI)
Z
P
I2(%)
Ph
Gender (male/female) Differentiation (moderate + well/poor) Tumor size (≥ 5 cm/ < 5 cm) Tumor stage (I + II/III + IV) Tumor invasion (T1 + T2/T3 + T4) Lymph node metastasis (pos/neg) Venous invasion (pos/neg)
7 5 2 6 6 5 2
1247 631 246 999 947 829 120
0.87(0.66, 0.81(0.57, 1.69(1.13, 0.47(0.26, 0.51(0.19, 0.80(0.49, 0.63(0.38,
1.0 1.22 2.53 2.53 1.38 0.92 1.83
0.32 0.22 0.01 0.01 0.17 0.36 0.07
29 36 45 70 81 52 0
0.21 0.18 0.18 0.005 < 0.0001 0.08 0.84
1.15) 1.14) 2.53) 0.84) 1.33) 1.29) 1.03)
1457–1461. [7] A. Huijbers, R.A. Tollenaar, G.W. v Pelt, E.C. Zeestraten, S. Dutton, McConkey CC, E. Domingo, V.T. Smit, R. Midgley, B.F. Warren, E.C. Johnstone, D.J. Kerr, W.E. Mesker, The proportion of tumor stroma as a strong prognosticator for stage II and III colon cancer patients: validation in the VICTOR trial, Ann. Oncol. 24 (2013) 179–185. [8] Z. Lv, X. Cai, X. Weng, H. Xiao, C. Du, J. Cheng, L. Zhou, H. Xie, K. Sun, J. Wu, S. Zheng, Tumorstromaratio is a prognostic factor for survival in hepatocellular carcinoma patients after liver resection or transplantation, Surgery 158 (2015) 142–150. [9] J.P. Higgins, S.G. Thompson, Quantifying heterogeneity in a meta-analysis, Stat. Med. 21 (2002) 1539–1558. [10] E.F. Courrech Staal, V.T. Smit, M.L. van Velthuysen, J.M. Spitzer-Naaykens, M.W. Wouters, W.E. Mesker, R.A. Tollenaar, J.W. van Sandick, Reproducibility and validation of tumour stroma ratio scoring on oesophageal adenocarcinoma biopsies, Eur. J. Cancer 47 (3) (2011 Feb) 375–382. [11] N.P. West, M. Dattani, P. McShane, G. Hutchins, J. Grabsch, W. Mueller, D. Treanor, P. Quirke, H. Grabsch, The proportion of tumour cells is an independent predictor for survival in colorectal cancer patients, Br. J. Cancer 102 (10) (2010 May 11) 1519–1523. [12] W.E. Mesker, J.M. Junggeburt, K. Szuhai, P. de Heer, H. Morreau, H.J. Tanke, R.A. Tollenaar, The carcinoma-stromal ratio of colon carcinoma is an independent factor for survival compared to lymph node status and tumor stage, Cell. Oncol. 29 (5) (2007) 387–398. [13] D. Lee, I.H. Ham, S.Y. Son, S.U. Han, Y.B. Kim, H. Hur, Intratumor stromal proportion predicts aggressive phenotype of gastric signet ring cell carcinomas, Gastric Cancer (2016 Nov 17) (Epub ahead of print). [14] J.H. Park, C.H. Richards, D.C. McMillan, P.G. Horgan, C.S. Roxburgh, The relationship between tumour stroma percentage, the tumour microenvironment and survival in patients with primary operable colorectal cancer, Ann. Oncol. 25 (3) (2014 Mar) 644–651, http://dx.doi.org/10.1093/annonc/mdt593. [15] A. Giatromanolaki, E. Sivridis, M.I. Koukourakis, The pathology of tumor stromatogenesis, Cancer Biol Ther. 6 (2007) 639–645. [16] M.J. Bissell, D. Radisky, Putting tumours in context, Nat. Rev. Cancer 1 (2001) 46–54. [17] D.X. Nguyen, P.D. Bos, J. Massague, Metastasis: from dissemination to organ-specific colonization, Nat. Rev. Cancer 9 (2009) 274–284. [18] R.A. Brekken, P. Puolakkainen, D.C. Graves, G. Workman, S.R. Lubkin, E.H. Sage, Enhanced growth of tumors in SPARC null mice is associated with changes in the ECM, J. Clin. Invest. 111 (2003) 487–495. [19] H. Luo, G. Tu, Z. Liu, M. Liu, Cancer-associated fibroblasts: a multifaceted driver of breast cancer progression, Cancer Lett. 361 (2015) 155–163. [20] G. Lorusso, C. Rüegg, The tumor microenvironment and its contribution to tumor evolution toward metastasis, Histochem. Cell Biol. 130 (2008) 1091–1103. [21] J. Massagué, TGFbeta in cancer, Cell 134 (2008) 215–230. [22] P. Cirri, P. Chiarugi, Cancar-associated fibroblasts and tumor cells: a diabolic liaison driving cancer progression, Cancer Metastasis Rev. 31 (2012) 195–208. [23] R.M. Bremnes, T. Donnem, S. Al-Saad, K. Al-Shibli, S. Andersen, R. Sirera, C. Camps, I. Marinez, L.T. Busund, The role of tumor stroma in cancer progression and prognosis: emphasis on carcinoma-associated fibroblasts and non-small cell lung cancer, J. Thorac. Oncol. 6 (2011) 209–217. [24] A.J. Schottelius, H. Dinter, Cytokines. NF-kappaB, microenvironment, intestinal inflammation and cancer, Cancer Treat. Res. 130 (2006) 67–87. [25] R. Kalluri, M. Zeisberg, Fibroblasts in cancer, Nat. Rev. Cancer 6 (2006) 392–401. [26] M. Shimoda, K.T. Mellody, A. Orimo, Carcinoma-associated fibroblasts area ratelimiting determinant for tumour progression, Semin. Cell Dev. Biol. 21 (2010) 19–25.
included; thus, the reliability of the results may be weakened. Thirdly, the search language only included English. Finally, the patient level was not included in the present analysis. In conclusion, the present meta-analysis revealed that low TSR stroma could be a poor prognosis predictor, including OS and DFS, in digestive system cancers. The specific relation between TSR and prognosis is clear, based on the results of the present study, so a clear mechanism will likely be revealed in the near future. Authors' contributions Conception and design: RJZ, WS, SBZ. Acquisition of data: RJZ, WS. Analysis and interpretation of data: RJZ, WS, SBZ. Writing and final approval of the manuscript: RJZ, WS, SBZ. Acknowledgments The authors gratefully acknowledge the reviewers for their helpful comments on this paper. Funding No funding was provided for this study. Conflicts of interest The authors declare that there are no conflicts of interest. References [1] R.L. Siegel, K.D. Miller, A. Jemal, Cancer statistics, 2015, CA Cancer J. Clin. 65 (2015) 5–29. [2] G. Bjelakovic, D. Nikolova, R.G. Simonetti, C. Gluud, Antioxidant supplements for preventing gastrointestinal cancers, Cochrane Database Syst. Rev. 3 (2008) CD 004183. [3] O. De Wever, M. Mareel, Role of tissue stroma in cancer cell invasion, J. Pathol. 200 (2003) 429–447. [4] G. Finak, N. Bertos, F. Pepin, S. Sadekova, M. Souleimanova, H. Zhao, et al., Stromal gene expression predicts clinical outcome in breast cancer, Nat. Med. 14 (2008) 518–527, http://dx.doi.org/10.1038/nm1764. [5] E.M. De Kruijf, J.G. van Nes, C.J. van de Velde, H. Putter, V.T. Smit, G.J. Liefers, P.J. Kuppen, R.A. Tollenaar, W.E. Mesker, Tumor-stroma ratio in the primary tumor is a prognostic factor in early breast cancer patients, especially in triple-negative carcinoma patients, Breast Cancer Res. Treat. 125 (2011) 687–696. [6] K. Wang, W. Ma, J. Wang, L. Yu, X. Zhang, Z. Wang, B. Tan, N. Wang, B. Bai, S. Yang, H. Liu, S. Zhu, Y. Cheng, Tumor-stroma ratio is an independent predictor for survival in esophageal squamous cell carcinoma, J. Thorac. Oncol. 7 (2012)
68
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/cca
Tumor-stroma ratio(TSR) as a potential novel predictor of prognosis in digestive system cancers: A meta-analysis
MARK
Runjin Zhang1, Wei Song1, Kai Wang, Shubing Zou⁎ Department of hepatobiliary surgery, The Second Affiliated Hospital of Nanchang University, Nanchang, China
A R T I C L E I N F O
A B S T R A C T
Keywords: Tumor-stroma ratio (TSR) Digestive system cancers Prognosis
Main problem: The tumor-stroma ratio (TSR) has been reported as a prognosis predictor in multiple cancers. The aim of this meta-analysis was to investigate the potential value of TSR as a prognostic predictor of cancer in the digestive system. Methods: We searched PubMed, Embase, Elsevier and Web of Science. All studies exploring the association of TSR with overall survival (OS) or disease-free survival (DFS), and lymph node metastasis (LNM) were identified. Results: In total, eight studies were eligible for analysis, and they included 1959 patients. Meta-analysis showed that the low TSR in the tumor could predict poor overall survival (OS) in multiple cancers (pooled Hazard Ratio [HR]: 2.15, 95%CI: 1.80–2.57, P < 0.00001, fixed effects). For disease-free survival (DFS), low TSR was also a significant predictor (pooled Hazard Ratio [HR]: 2.31, 95%CI: 1.88–2.83, P < 0.00001, fixed effects). In addition, low TSR was correlated with tumor stage. Discussion: The tumor-stroma ratio (TSR) may potentially serve as a poor prognostic predictor for the metastasis and prognosis of cancer.
1. Introduction
2. Evidence acquisition
Cancer is a large public health problem and a major cause of deaths worldwide. In 2015, 1,658,370 individuals developed cancer, and 589,430 people died of cancer in the United States, according to the American National Center for Health Statistics [1]. Furthermore, digestive system cancers are the deadliest forms of cancer, including many organs cancer in our body [2]. Reflecting the complex initiation and progression mechanism, the 5-year survival rate is low in most tumors. Consequently, we developed a novel prognostic marker to assist clinical application in multiple cancers. The tumor-stroma ratio (TSR) is the proportion of tumor cells in the stroma. Tumors are complex tissues that are composed of carcinoma cells and surrounding stroma; thus, the tumor stroma is associated with tumor initiation, progression, and metastasis, and it holds prognostic value [3–4]. Recently, the tumor to stroma ratio of (TSR) has been demonstrated to be a novel and practical prognostic predictor in many neoplasms, such as breast cancer [5], esophageal cancer [6], colon cancer [7], and hepatocellular carcinoma [8]. Those studies have shown that a low TSR is a poor prognostic marker [5.6.7.8]. So we need performed a meta-analysis to explore the relation of TSR with OS and DFS.
2.1. Search strategies
⁎
1
Corresponding author. E-mail address: [email protected] (S. Zou). These authors have contributed equally to this work.
http://dx.doi.org/10.1016/j.cca.2017.07.003 Received 23 February 2017; Received in revised form 3 July 2017; Accepted 3 July 2017 Available online 17 July 2017 0009-8981/ © 2017 Published by Elsevier B.V.
We searched MEDLINE, Embase, Elsevier and Web of science. Search terms included “carcinoma-stroma ratio (CSR)”, “tumor-stroma ratio (TSR)”, “tumor-stroma percentage (TSP)”, “cancer”, “tumor”, “carcinoma”, “neoplasms”, “survival”, “prognosis”, and “outcome”. Systematic reviews and meta-analyses were manually retrieved and served as references for the included studies. The language was restricted to English. The literature search was completed on November 31, 2016. 2.2. Inclusion criteria The inclusion criteria included 1) a clear relationship between TSR and overall survival or TSR and disease-free survival; 2) multivariate analysis as the analysis method; 3) studies reported on hazard ratio (HR) and 95% confidence interval (CI); and 4) studies published in English.
Clinica Chimica Acta 472 (2017) 64–68
R. Zhang et al.
and were included in the meta-analysis [6–8,10–14]; this process is shown in Fig. 1.
2.3. Exclusion criteria The exclusion criteria included 1) metastatic cancers in organs not related to the primary cancers; 2) studies investigating the structure, mechanism and functions; 3) studies in which it was impossible to obtain available data; 4) duplicate publications; and 5) non-Englishlanguage studies.
3.2. Characteristics of the included studies A total of eight studies that included 1959 patients were enrolled in the present meta-analysis. The number of patients in each study ranged from 81 to 710. Among these studies, three studies were conducted in the UK, two were conducted in the Netherlands, two were conducted in China, and one was conducted in the Republic of Korea. Among these eight studies, four studies on colorectal cancer, one study on esophageal adenocarcinoma, one study on esophageal squamous cell carcinoma, one study on hepatocellular carcinoma, and one study on gastric signet ring cell carcinoma were observed. In total, eight studies focused on OS, and five studies focused on DFS. The hazard ratios (HRs) and corresponding 95% confidence intervals were obtained though multivariate analysis. Most of studies were divided into two groups: high stroma ratio and low stroma ratio of TSR, and the cutoff values were all set to 50%. Concerning the methodological quality of these studies, the NOS scores of all included studies were ≥6. Detailed patient characteristics are shown in Table 1.
2.4. Data extraction According to the aforementioned inclusion and exclusion criteria, two investigators reviewed all eligible studies and extracted the necessary data. Any disagreement was resolved by discussions among all coauthors. The following information was collected: name of the first author, year of publication, country, number of patients, tumor type, stage, overall survival (OS), disease-free survival (DFS), analysis methods, hazard ratio (HR) with 95% confidence interval (CI), and follow-ups. 2.5. Statistical analysis The meta-analysis was conducted using the RevMan version 5.3 software (Cochrane Collaboration). Heterogeneity was measured using Q and I2 tests, and I2 > 50% and P < 0.10 indicated heterogeneity [9]. When I2 < 50%, no statistically significant heterogeneity was considered. The fixed-effects model was used to pool the results, and the random-effects model was used. Pooled hazard ratios (HRs) were extracted from the published data; the log HR and standard error (SE) were used to aggregate the survival results. P < 0.05 was defined as statistically significant.
4. Meta-analysis 4.1. Association between TSR and OS Eight eligible studies showed that the low TSR, reflecting rich stroma, was associated with the high risk of OS (pooled Hazard Ratio [HR]: 2.15, 95%CI: 1.80–2.57, P < 0.00001, fixed effects, I2 = 0%, Ph = 0.58) (Table 3, Fig. 2). In the subgroup analysis, potential heterogeneity was detected using the region, sample size, cancer type, clinical stage, and NOS score (Table 2). And we found that rich stroma had significantly poor prognostic in all subgroups.
3. Evidence synthesis 3.1. Search results
4.2. Association between TSR and DFS A total of 322 articles were found in the selected databases. A total of 25 articles were excluded as duplication, and 282 articles were excluded after screening the titles and abstracts, leaving 15 articles for further evaluation. As a result, 8 articles fulfilled the inclusion criteria
Five studies described the relation between TSR and DFS. An analysis of 1402 patients to evaluate this association showed that a low TSR was significantly correlated with the poor outcome of DFS (pooled
Fig. 1. Flow diagram of the study selection process. 332 studies were preretrieved in accordance with the established search strategies. Of these articles, 25 articles were excluded due to duplication and 282 were excluded by screening titles and abstracts, leaving 15 articles for further evaluation. As a result, 8 articles, fulfilled the inclusion criteria, were involved in the meta-analysis.
65
Clinica Chimica Acta 472 (2017) 64–68
6
6
7
6
7
7
8
6
≥ 50% I–III 95 Until Mar 2011 2007 Asia 82/13 60 (42–77) China Wang.K 2012
CC UK Park.JH 2014
SRCC Korea Lee D 2016
HCC China Lv Z 2015
CC UK West.NP 2010
OC
≥ 50% 331 160/171
Europe
1997–2008
NR
I–III
≥ 50% II–III 175 Median 72.5 2005–2008 Asia 94/81
≥ 50% I–IV 300 Until Aug 2014 2006–2013 Asia 262/38
≥ 47% I–IV 145 Median 52 (26–74) 1990–1995 Europe 58/87
≥ 50% II–III 710 Median 58 (2–76) Europe UK Huijbers.A 2013
CC
65 (25–86) 69 (61–76) 54 (27–80) 49 (23–85) NR
438/272
2002–2004
≥ 50% I–IV 81 Median 23 (3–220) 1990–2004 Europe 66/15 64(37–82) OC Netherlands Courrech Staal E.F 2011
We identified several elemental factors to explore the relationship of TSR with clinicopathological features (Table 3). The low TSR in tumor tissue was associated with the size and stage of the tumor. Pooled data revealed that a low TSR was significantly associated with tumor stage (I + II/III + IV, OR: 0.47, 95%CI: 0.26–0.84, P = 0.01, random effects, I2 = 70, Ph = 0.005). In contrast, no significant association was discovered in gender (male/female), differentiation (moderate + well/ poor), tumor invasion (T1 + T2/T3 + T4), lymph node metastasis (pos/neg), and venous invasion (pos/neg).
CC: colon cancer or colorectal cancer; OC: oesophageal cancer; HCC: hepatocellular carcinoma; SRCC: signet ring cell carcinomas; OS: overall survival; DFS: disease-free survival; MV: multivariate; NR: none reported.
MV Reported
MV
MV Reported
Reported OS DFS
MV
MV Reported
Reported
OS DFS OS DFS OS
MV Reported
MV Reported
OS FDS OS
MV Reported
OS DFS OS
33 27.0% 12 14.8% 207 29.2% 35 24.1% 75 25.0% 64 36.6% 250 75.5% 30 31.6% Netherlands Mesker WE 2007
CC
NR
72/50
Europe
1980–2001
NR
122
I–III
≥ 50%
HR estimate Stroma rich (%) Cut off Clinical stage No. of patients Follow-up (months) Duration Region Gender (M/F) Age (years) Cancer type Country Study (author, years)
Table 1 Characteristics of the eligible studies in the meta-analysis.
Hazard Ratio [HR]: 2.31, 95%CI: 1.88–2.83, P < 0.001; fixed effects, I2 = 0%, Ph = 0.57) (Fig. 3). 4.3. Correlation between TSR and clinicopathological features
Survival analysis
Analysis
Quality
R. Zhang et al.
5. Discussion The digestive system, which is one of the most complicated systems in the human body. Different locations in the digestive system show different types of cancers. Therefore, the tumor mechanism in the digestive system during development is extremely complex. Thus, we need a simple, convenient, and effective method to predict the prognosis in digestive system tumors. The TSR (ratio value of tumor to stroma in tumor tissue) was first reported to determine the prognosis of colorectal cancer in 2007 [12] as a simple, cheap, convenient and useful method in clinical studies. Thereafter, increasing studies have shown the value of the TSR as a prognosis predictor in various cancers. In these studies, the patients were divided into two groups based on a cutoff value of TSR = 50%, which was considered the most significant prognostic value. However, the prognostic value of TSR in digestive system tumor remains unclear. The present study is a meta-analysis to explore the correlation between TSR and overall survival in digestive system cancers. The present meta-analysis, which included eight studies and 1959 patients, showed that a low TSR was associated with poor overall survival in digestive system tumors (pooled Hazard Ratio [HR]: 2.15, 95%CI: 1.80–2.57, P < 0.00001, fixed effects, I2 = 0%, Ph = 0.58). When the analysis was restricted to the survival outcome of DFS, a positive result was observed in the present meta-analysis (pooled Hazard Ratio [HR]: 2.31, 95%CI: 1.88–2.83, P < 0.001; fixed effects). Thus, TSR could also be an independent predictor of OS/DFS in patients with digestive system cancers. The meta-analysis was shown a low heterogeneity of OS (I2 = 0%, Ph = 0.58) and DFS (I2 = 0%, Ph = 0.57). If necessary, a multicenter analysis and large patient sample size were used to study the problems in digestive system cancer. We also analyzed the association between TSR and the clinicopathological of OS on digestive system cancers. However, we still do not fully understand the mechanism underlying the promoting effect of stroma in tumors. According to the results, a high proportion of stroma was significantly associated with tumor size and tumor stage, which reflects poor prognosis in digestive system cancers. The formation of tumor-activated stroma results in the disruption of the epithelial tissue, the immune evasion of malignant cells, and the invasion of tumors, regarded as tumor stromatogenesis [15]. In normal tissue, the stroma acts as a protective area in tumor stromatogenesis. However, in tumor tissue, the stroma is a major part of the tumor microenvironment that could accelerate tumor progression [16]. The main composition of extracellular matrix (ECM) in the tumor stroma can change the microenvironment of tumors to promote cancer cell metastasis [17], such as the overexpression of amino acids, which are rich in cysteine [18], and matrix metalloproteinases [19], which can facilitate tumorigenesis and invasion. Cancer-associated fibroblasts (CAF), which are the predominant cells within the tumor tissue, are key regulators between stroma and cancer cells [20]. A variety of soluble factors that are secreted by CAFs act in a paracrine manner and affect not only cancer cells but also other cell types [21]. In different periods, 66
Clinica Chimica Acta 472 (2017) 64–68
R. Zhang et al.
Fig. 2. Forest plots of the overall survival (OS). Hazard ratios (HRs) for each trial are represented by the squares, and the horizontal lines crossing the square stand for the 95% confidence intervals (CIs). The diamonds represent the estimated pooled effect of the overall outcome for OS. Table 2 Pooled hazard ratios(HRs) for OS according to subgroup analysis. Subgroup
No. of studies
No. of patients
Effects model
HR(95%CI)
Z
P
I2(%)
Ph
Overall Region Asia Europe Sample size < 100 ≥ 100 Cancer type CC HCC OC SRCC Clinical stage I–III I–IV II–III NOS ≥7 <6
8
1959
Fixed
2.15(1.80, 2.57)
8.39
< 0.00001
0
0.58
3 5
570 1389
Fixed Fixed
2.77(2.01, 3.84) 1.92(1.55, 2.38)
6.17 5.98
< 0.00001 < 0.00001
0 0
0.69 0.83
2 6
176 1783
Fixed Fixed
2.62(1.74, 3.94) 2.05(1.68, 2.50)
4.64 7.07
< 0.00001 < 0.00001
37 0
0.21 0.71
4 1 2 1
1308 300 176 175
Fixed – Fixed –
1.90(1.51, 2.55(1.44, 2.62(1.74, 2.50(1.48,
2.39) 4.52) 3.94) 4.24)
5.44 3.21 4.64 3.4
< 0.00001 0.001 < 0.00001 0.0007
0 – 37 –
0.70 – 0.21 –
3 3 2
548 526 885
Fixed Fixed Fixed
2.39(1.76, 3.26) 2.22(1.58, 3.13) 1.91(1.44, 2.54)
5.54 4.59 4.46
< 0.00001 < 0.00001 < 0.00001
27 0 30
0.25 0.84 0.23
4 4
1111 848
Fixed Fixed
1.96(1.54, 2.48) 2.43(1.85, 3.19)
5.56 6.40
< 0.00001 < 0.00001
0 0
0.68 0.43
CC: colon cancer or colorectal cancer; OC: oesophageal cancer; HCC: hepatocellular carcinoma; SRCC: signet ring cell carcinomas; NOS: Newcastle-Ottawa Scale.
Fig. 3. Forest plots of the disease-free survival (DFS). Hazard ratios (HRs) for each trial are represented by the squares, and the horizontal lines crossing the square stand for the 95% confidence intervals (CIs). The diamonds represent the estimated pooled effect of the overall outcome for DFS.
cells have frequently been associated with the development of highgrade malignancies and poor prognoses [26]. Although this meta-analysis revealed a positive conclusion, it does have several limitations. Firstly, this study found a low heterogeneity, but it does have many biases. One of the biases is the method to evaluate the TSR. Although those researchers have standard evaluation methods, but it still can't avoid personal subjectivity. So we need a more scientific and rigorous method to evaluate the TSR. Secondly, the sample size was small. Only eight studies, involving 1595 patients, were
CAFs also play different roles. CAFs can inhibit cancer cells by enhancing the formation of gap junctions during the early stages. However, during later stages, CAFs play an accelerated role in tumor growth and progression through several tumor-secreted factors [22,23]. Moreover, many growth factors and chemokines, such as nuclear factor κB [24] and tumor necrosis factor α [25], are secreted into the stroma by stromal cells or cancer cells and act as chemoattractants for other noncancer cells. The mechanism by which stroma cells affect cancer growth and progression remains incompletely understood. However, stroma 67
Clinica Chimica Acta 472 (2017) 64–68
R. Zhang et al.
Table 3 Meta-analysis of the tumor-stroma ratio and clinicopathological features of digestive system cancer. Characteristics
No. of studies
No. of patients
OR(95%CI)
Z
P
I2(%)
Ph
Gender (male/female) Differentiation (moderate + well/poor) Tumor size (≥ 5 cm/ < 5 cm) Tumor stage (I + II/III + IV) Tumor invasion (T1 + T2/T3 + T4) Lymph node metastasis (pos/neg) Venous invasion (pos/neg)
7 5 2 6 6 5 2
1247 631 246 999 947 829 120
0.87(0.66, 0.81(0.57, 1.69(1.13, 0.47(0.26, 0.51(0.19, 0.80(0.49, 0.63(0.38,
1.0 1.22 2.53 2.53 1.38 0.92 1.83
0.32 0.22 0.01 0.01 0.17 0.36 0.07
29 36 45 70 81 52 0
0.21 0.18 0.18 0.005 < 0.0001 0.08 0.84
1.15) 1.14) 2.53) 0.84) 1.33) 1.29) 1.03)
1457–1461. [7] A. Huijbers, R.A. Tollenaar, G.W. v Pelt, E.C. Zeestraten, S. Dutton, McConkey CC, E. Domingo, V.T. Smit, R. Midgley, B.F. Warren, E.C. Johnstone, D.J. Kerr, W.E. Mesker, The proportion of tumor stroma as a strong prognosticator for stage II and III colon cancer patients: validation in the VICTOR trial, Ann. Oncol. 24 (2013) 179–185. [8] Z. Lv, X. Cai, X. Weng, H. Xiao, C. Du, J. Cheng, L. Zhou, H. Xie, K. Sun, J. Wu, S. Zheng, Tumorstromaratio is a prognostic factor for survival in hepatocellular carcinoma patients after liver resection or transplantation, Surgery 158 (2015) 142–150. [9] J.P. Higgins, S.G. Thompson, Quantifying heterogeneity in a meta-analysis, Stat. Med. 21 (2002) 1539–1558. [10] E.F. Courrech Staal, V.T. Smit, M.L. van Velthuysen, J.M. Spitzer-Naaykens, M.W. Wouters, W.E. Mesker, R.A. Tollenaar, J.W. van Sandick, Reproducibility and validation of tumour stroma ratio scoring on oesophageal adenocarcinoma biopsies, Eur. J. Cancer 47 (3) (2011 Feb) 375–382. [11] N.P. West, M. Dattani, P. McShane, G. Hutchins, J. Grabsch, W. Mueller, D. Treanor, P. Quirke, H. Grabsch, The proportion of tumour cells is an independent predictor for survival in colorectal cancer patients, Br. J. Cancer 102 (10) (2010 May 11) 1519–1523. [12] W.E. Mesker, J.M. Junggeburt, K. Szuhai, P. de Heer, H. Morreau, H.J. Tanke, R.A. Tollenaar, The carcinoma-stromal ratio of colon carcinoma is an independent factor for survival compared to lymph node status and tumor stage, Cell. Oncol. 29 (5) (2007) 387–398. [13] D. Lee, I.H. Ham, S.Y. Son, S.U. Han, Y.B. Kim, H. Hur, Intratumor stromal proportion predicts aggressive phenotype of gastric signet ring cell carcinomas, Gastric Cancer (2016 Nov 17) (Epub ahead of print). [14] J.H. Park, C.H. Richards, D.C. McMillan, P.G. Horgan, C.S. Roxburgh, The relationship between tumour stroma percentage, the tumour microenvironment and survival in patients with primary operable colorectal cancer, Ann. Oncol. 25 (3) (2014 Mar) 644–651, http://dx.doi.org/10.1093/annonc/mdt593. [15] A. Giatromanolaki, E. Sivridis, M.I. Koukourakis, The pathology of tumor stromatogenesis, Cancer Biol Ther. 6 (2007) 639–645. [16] M.J. Bissell, D. Radisky, Putting tumours in context, Nat. Rev. Cancer 1 (2001) 46–54. [17] D.X. Nguyen, P.D. Bos, J. Massague, Metastasis: from dissemination to organ-specific colonization, Nat. Rev. Cancer 9 (2009) 274–284. [18] R.A. Brekken, P. Puolakkainen, D.C. Graves, G. Workman, S.R. Lubkin, E.H. Sage, Enhanced growth of tumors in SPARC null mice is associated with changes in the ECM, J. Clin. Invest. 111 (2003) 487–495. [19] H. Luo, G. Tu, Z. Liu, M. Liu, Cancer-associated fibroblasts: a multifaceted driver of breast cancer progression, Cancer Lett. 361 (2015) 155–163. [20] G. Lorusso, C. Rüegg, The tumor microenvironment and its contribution to tumor evolution toward metastasis, Histochem. Cell Biol. 130 (2008) 1091–1103. [21] J. Massagué, TGFbeta in cancer, Cell 134 (2008) 215–230. [22] P. Cirri, P. Chiarugi, Cancar-associated fibroblasts and tumor cells: a diabolic liaison driving cancer progression, Cancer Metastasis Rev. 31 (2012) 195–208. [23] R.M. Bremnes, T. Donnem, S. Al-Saad, K. Al-Shibli, S. Andersen, R. Sirera, C. Camps, I. Marinez, L.T. Busund, The role of tumor stroma in cancer progression and prognosis: emphasis on carcinoma-associated fibroblasts and non-small cell lung cancer, J. Thorac. Oncol. 6 (2011) 209–217. [24] A.J. Schottelius, H. Dinter, Cytokines. NF-kappaB, microenvironment, intestinal inflammation and cancer, Cancer Treat. Res. 130 (2006) 67–87. [25] R. Kalluri, M. Zeisberg, Fibroblasts in cancer, Nat. Rev. Cancer 6 (2006) 392–401. [26] M. Shimoda, K.T. Mellody, A. Orimo, Carcinoma-associated fibroblasts area ratelimiting determinant for tumour progression, Semin. Cell Dev. Biol. 21 (2010) 19–25.
included; thus, the reliability of the results may be weakened. Thirdly, the search language only included English. Finally, the patient level was not included in the present analysis. In conclusion, the present meta-analysis revealed that low TSR stroma could be a poor prognosis predictor, including OS and DFS, in digestive system cancers. The specific relation between TSR and prognosis is clear, based on the results of the present study, so a clear mechanism will likely be revealed in the near future. Authors' contributions Conception and design: RJZ, WS, SBZ. Acquisition of data: RJZ, WS. Analysis and interpretation of data: RJZ, WS, SBZ. Writing and final approval of the manuscript: RJZ, WS, SBZ. Acknowledgments The authors gratefully acknowledge the reviewers for their helpful comments on this paper. Funding No funding was provided for this study. Conflicts of interest The authors declare that there are no conflicts of interest. References [1] R.L. Siegel, K.D. Miller, A. Jemal, Cancer statistics, 2015, CA Cancer J. Clin. 65 (2015) 5–29. [2] G. Bjelakovic, D. Nikolova, R.G. Simonetti, C. Gluud, Antioxidant supplements for preventing gastrointestinal cancers, Cochrane Database Syst. Rev. 3 (2008) CD 004183. [3] O. De Wever, M. Mareel, Role of tissue stroma in cancer cell invasion, J. Pathol. 200 (2003) 429–447. [4] G. Finak, N. Bertos, F. Pepin, S. Sadekova, M. Souleimanova, H. Zhao, et al., Stromal gene expression predicts clinical outcome in breast cancer, Nat. Med. 14 (2008) 518–527, http://dx.doi.org/10.1038/nm1764. [5] E.M. De Kruijf, J.G. van Nes, C.J. van de Velde, H. Putter, V.T. Smit, G.J. Liefers, P.J. Kuppen, R.A. Tollenaar, W.E. Mesker, Tumor-stroma ratio in the primary tumor is a prognostic factor in early breast cancer patients, especially in triple-negative carcinoma patients, Breast Cancer Res. Treat. 125 (2011) 687–696. [6] K. Wang, W. Ma, J. Wang, L. Yu, X. Zhang, Z. Wang, B. Tan, N. Wang, B. Bai, S. Yang, H. Liu, S. Zhu, Y. Cheng, Tumor-stroma ratio is an independent predictor for survival in esophageal squamous cell carcinoma, J. Thorac. Oncol. 7 (2012)
68