Prognostic and predictive significance of thymidylate synthase protein ...

2 downloads 0 Views 3MB Size Report
Abstract. BACKGROUND: It remains controversial whether thymidylate synthase (TS) protein expression is associated with survival for patients with non-small ...
65

Cancer Biomarkers 15 (2015) 65–78 DOI 10.3233/CBM-140432 IOS Press

Prognostic and predictive significance of thymidylate synthase protein expression in non-small cell lung cancer: A systematic review and meta-analysis Qingyun Liua,1 , Zubin Yub,1 , Ying Xianga , Na Wua , Long Wua , Bin Xua , Liang Wangc , Ping Yangd , Yafei Lia,∗ and Li Baie,∗ a

Department of Epidemiology, College of Preventive Medicine, Third Military Medical University, Chongqing, China b Department of Thoracic Surgery, Xinqiao Hospital, Third Military Medical University, Chongqing, China c Department of Pathology, MCW Cancer Center, Medical College of Wisconsin, Milwaukee, WI, USA d Health Sciences Research, Mayo Clinic College of Medicine, Rochester, MI, USA e Department of Respiratory Disease, Xinqiao Hospital, Third Military Medical University, Chongqing, China

Abstract. BACKGROUND: It remains controversial whether thymidylate synthase (TS) protein expression is associated with survival for patients with non-small cell lung cancer (NSCLC). OBJECTIVE: To evaluate prognostic and predictive significance of tumor TS protein level in NSCLC. METHODS: Electronic searches were performed for relevant studies in PubMed, EMBASE, Web of Science, and Chinese Biomedical Literature Database. Hazard ratios (HRs) for overall survival (OS) and progression-free survival (PFS) were pooled for meta-analysis. Subgroup and sensitivity analyses were performed. Publication bias was evaluated by funnel plot and Begg’s test. RESULTS: Twenty-four studies, including 2280 patients, were eligible. This analysis showed that patients with low TS expression had statistically significantly longer OS and PFS than those with high TS (HR = 0.51 and HR = 0.49, respectively). Based on TS-targeted drug use status, TS expression was significantly associated with OS in pemetrexed (HR = 0.42) and 5Fluorouracil subgroups (HR = 0.34), but not in no TS-targeted drug subgroup. There were similar results for PFS analyses. Sensitivity analysis indicated that the results were robust. Begg’s test did not reveal any publication bias. CONCLUSION: Low TS protein expression is a favorable predictive factor for better OS/PFS in NSCLC patients treated with TS-targeted drugs. Prognostic value of TS protein expression needs further validation. Keywords: Non-small cell lung cancer, thymidylate synthase, prognosis, predictive factor

1. Introduction 1 These

authors contributed equally to this work. authors: Yafei Li, Department of Epidemiology, College of Preventive Medicine, Third Military Medical University, NO.30 Gaotanyan Street, Chongqing 400038, China. Tel.: +86 23 68752293; E-mail: [email protected]; Li Bai, Department of Respiratory Disease, Xinqiao Hospital, Third Military Medical University, Xinqiao Main Street, Chongqing 400037, China. Tel.: +86 23 68755281; E-mail: [email protected]. ∗ Corresponding

Lung cancer is still the leading cause of cancer deaths in the world [1]. Non-small cell lung cancer (NSCLC) accounts for about 80% of all lung cancer cases [2]. Despite of improved treatment in lung cancer, the prognosis of NSCLC is still poor. For advanced stage NSCLC, 5-year survival varies

c 2015 – IOS Press and the authors. All rights reserved ISSN 1574-0153/15/$27.50 

66

Q. Liu et al. / Thymidylate synthase and prognosis of NSCLC

widely (3%–50%) depending on the number of lymph nodes involved, resectability, and tumor histology [3]. Chemotherapy remains the major component of the standard care in conjunction with radiation therapy and supportive care for patients with advanced-stage lung cancer. For a given treatment, usually prescribed by a standard dosing protocol, survival time differs significantly even after stratification by tumor stage, histology, and other clinical information, highlighting the need for improved predictive markers. In addition to clinical and biological behaviors of the tumor, it has been reported that some molecular biomarkers may also predict outcome of NSCLC [4,5]. Thymidylate synthase (TS) catalyzes the methylation of deoxyuridine monophosphate (dUMP) to deoxythymidine monophosphate (dTMP) [6,7], which is critical for DNA replication/repair and cancer cell proliferation [6,8]. Considering its biological role, researchers utilized TS as a target of some antitumor drugs, such as pemetrexed (also known as Alimta, LY-231514), 5-fluorouracil (5-FU) and its prodrugs (UFT, Capecitabine, S-1) [9]. These drugs are TS inhibitors [10–13]. Several studies reported that TS expression was associated with prognosis of NSCLC, but inconsistent conclusions existed in different studies [7, 14–17]. For example, some studies favored low TS expression [7,14,15]; while others showed high expression as beneficial [16,18]. It was also reported that TS expression level was a predictive factor for efficacy of pemetrexed [19–21] and 5-FU [14,22,23] therapies. To clarify the prognostic and predictive significance of TS protein expression of NSCLC, we performed this systematic review and meta-analysis. This analysis used all available data to reconcile the consistency of the findings, overcoming drawbacks of small sample sizes of individual studies and increasing the statistical power.

2. Materials and methods 2.1. Search strategy We performed this meta-analysis in accordance with the guideline of the Meta-analysis of Observational Studies in Epidemiology Group (MOOSE) [24]. Electronic search was performed in PubMed, EMBASE, Web of Science, and Chinese Biomedical Literature Database (CBM) for relevant studies published up to April 2014 using web-based search engine. Search terms were “lung cancer/lung neoplasm/non-small cell

lung cancer/NSCLC”, “thymidylate synthase/TS”, and “prognosis/prognostic/predictive”. The reference lists of all retrieved articles and those of relevant review articles were also cross-referenced. The literature retrieval was performed in duplication by two independent reviewers (Q.L. and Z.Y.). 2.2. Selection criteria Studies published in English or Chinese language met the following criteria were eligible for inclusion regardless of publication year and study design: (1) the study subjects were patients diagnosed as NSCLC by pathology; (2) TS expression in lung tumor tissue was measured at protein level; (3) evaluate the correlation between TS expression level and progression-free survival (PFS) and/or overall survival (OS). When multiple publications reported on the same or overlapping data, we used the most recent or largest population as recommended by Little et al. [25]. Hazard ratio (HR) and its 95% confidence interval (CI) of TS expression level (high/low, or positive/negative) for PFS and/or OS should be provided, or can be converted from summary data or Kaplan-Meier curve according to Tierney et al. [26]. For the studies without adequate data or full text, we contacted the authors to get the unpublished data or full text; if the author did not provide the necessary material, these studies were excluded. 2.3. Quality assessment Quality of the enrolled studies was assessed using the European Lung Cancer Working Party quality scale for biological prognostic factors for lung cancer [27]. This is a specific scale for biological prognostic factors which was designed based on the opinion of experts and prior experience in the field [27]. The scale mainly includes 4 categories: scientific design, laboratory methodology, generalizability, and result analysis. Each category has several items, which are assessed using an ordinal scale (possible values: 2, 1, and 0). Here, we used absolute value of score to describe quality of study. The score ranges of 4 categories in the scale are as follow: 0–10 for scientific design, 0–14 for laboratory methodology, 0–12 for generalizability, and 0–8 for result analysis. Theoretically, maximum total score for a study is 44. Two reviewers (Z.Y. and Y.X.) independently assessed the quality of study, and the consensus was reached after discussion to resolve any disagreements.

Q. Liu et al. / Thymidylate synthase and prognosis of NSCLC

67

Fig. 1. Flow diagram of literature search and study selection.

2.4. Data extraction Using standard data extraction forms, two reviewers (Q.L. and Y.X.) independently extracted data from relevant studies. We extracted publication information (first author’s name, publication year, country), characteristics of participants (age, gender, sample size, histology, and stage), outcome information (HRs and its 95% CI for PFS or OS, and follow-up time) and TS expression information (TS measuring methods, and numbers of TS high (positive) /low (negative) expression cases, cutoff value of TS expression levels). Disagreement was resolved by consensus with a third reviewer (Y. L.). 2.5. Statistic methods We used pooled HR to describe prognostic value of TS expression. We took high/positive TS expression as reference. If a study provided HRs of both multivariate analysis and univariate analysis, we chose the former. For studies not providing HR directly and exact data

cannot be obtained after trying to contact with authors, we obtained the HR value and its 95% CI from available summary data or by reading Kaplan-Meier survival cure in the original studies using the spreadsheet developed by Tierney et al. [26]. We used program Engauge Digitizer Version 4.1 (M. Mitchell, Engauge Digitizer, http://digitizer.sourceforge.net) to read exact survival rates of each time points from the figures of Kaplan-Meier survival curve. This program can read exact values by digitizing data points from an image file after manually setting coordinate axis. Natural logarithm of HR (lnHR) for each eligible study and standard error (SE) of lnHR were used to calculate a pooled HR. SE of lnHR was calculated from 95% confident of HR according to the method of Cochrane Handbook for Systematic Reviews of Interventions (version 5.2). Statistical heterogeneity across studies was assessed using I 2 statistic, which describes the proportion of total variation across studies that is due to heterogeneity rather than chance [28]. I 2  50% suggests there is significant heterogeneity between studies. HR was pooled using a random effects model to manage heter-

Retrospective Retrospective Retrospective Retrospective Retrospective Perspective Retrospective Retrospective Retrospective Perspective Retrospective Perspective Perspective Perspective Retrospective Retrospective Perspective Retrospective Perspective Retrospective Retrospective Retrospective Perspective Perspective

Study design

Follow-up Region of Number of Male Median age (months) study patients (%) − Asia 42 57% 61.5 (26–79) 16.1 (7,30.8) Asia 55 51% 59 (24–84) − Asia 41 61% 68 (32–79) 19 Asia 30 53% − 14.9 Asia 54 50% 65 (43,82) 41 America 16 68% 60 − Asia 193 63% − − Asia 24 71% 661 (38–85) − Europe 193 54% 59 (23–90) − Asia 106 59% − 77.0 ± 28.9 Asia 173 67% 67 (35–76) 65.7 Asia 151 69% 64 (35–76) 63.6(13.2–82.8) Europe 168 78% 63 33 (9–42) Asia 70 67% 65.51 − Asia 54 80% 61.71 (38–80) − Asia 104 61% − 34.2 Asia 183 56% 68.51 (23–88) 73 (7–102) Asia 160 61% 67 (43–83) − America 160 54% − − Asia 45 84% 63.8 (46–76) − Europe 13 − − 34.4 ± 21.2 Asia 68 69% − − Europe 60 58% 65.1 (41–78) − Asia 117 78% 65.7 AD SCC NSCC SCC AD AD LCC AD LCC other AD SCC AD LCC other AD SCC other ADLCCSCCother AD AD SCC LCC AD SCC LCC ADSCCLCCother AD SCC ADSCCLCCother AD AD AD SCC LCC ADSCCLCCBAC AD SCC − AD SCC LCC AD other AD SCC LCC

Histology IIIB IV − IIIB IV IIIB IV IIIB IV IIIAIIIB IIIB IV IIIB IV I-IV IIIB-IV I-III I-III I-IV I-II I-IV p-stage I I-III I-III I III − I, III IIIB-IV I-IV

Stage

TS high/ TS targeted low drug IHC 20/22 PEM IHC 11/44 PEM IHC 21/20 PEM IHC 10/20 PEM IHC 23/31 PEM IHC 7/9 PEM IHC 92/101 PEM IHC 12/12 PEM IHC 83/110 PEM IHC 67/39 PEM IHC 93/80 no IHC 82/69 UFT IHC 79/89 no PCR/ IHC 33/37 no IHC 16/38 UFT IHC 75/29 no IHC 39/144 no IHC 74/86 no IHC/AQUA 120/40 no IHC 20/25 no IHC 8/5 no IHC 40/28 5-FU/UFT IHC 12/48 PEM IHC 63/54 no

Method

HR estimate Curve Curve Original Original Curve Original Original Curve Original Original Original Original Curve Curve Curve Original Original Curve Original Original Curve Original Original curve

age; 2 HRs for stage I and stage II-III were available for analyses. Method, method of TS protein detection; PFS, progression-free survival; OS, overall survival; HR, hazard ratio; Original, HR was available from original full text; Curve, HR was calculated from Kaplan-Meier curve; AD, adenocarcinoma; SCC, squamous cell carcinoma; LCC, large cell carcinoma; NSCC, nonsquamous non-small cell lung cancer; IHC, immunohistochemical method; AQUA, automated in situ protein quantification; PEM, pemetrexed; 5-FU, 5-Fluorouracil, UFT, tegafur and uracil (prodrug of 5-FU), no, not use TS-targeted drugs; “−”, data not available.

1 mean

Publication year Chen [20] 2011 Chang [29] 2010 Lee [30] 2013 Duan [31] 2012 Igawa [32] 2012 Gadgeel [42] 2011 Sun [19] 2011 Takezawa [33] 2011 Christoph [21] 2013 Wang [18] 2013 Huang [34] 2 2005 Nakano [14] 2006 Wynes [17] 2012 Shintani [35] 2003 Miyoshi [23] 2007 Nakagawa [36] 2002 Shimokawa [37] 2011 Kaira [7] 2012 Zheng [16] 2008 Yokomise [38] 2013 VOLM [40] 1992 Huang [22] 2000 Nicolson [41] 2013 NAKAGAWA [39] 2013

First author

Table 1 Characteristic of included studies

68 Q. Liu et al. / Thymidylate synthase and prognosis of NSCLC

Q. Liu et al. / Thymidylate synthase and prognosis of NSCLC

ogeneity on the condition that heterogeneity was significant (I 2  50%); otherwise, a fixed effects model was applied. To explore heterogeneity, and clarify TS predictive value for efficacy of the two TS-targeted antitumor drugs pemetrexed and 5-FU, we conducted subgroup analyses according to TS-targeted drug use status (no TS-targeted drug, pemetrexed and 5-FU). Subgroup analyses were further conducted according to stage, region of study (Asia and non-Asia), and HR estimation [original HR and KM-HR (HR was calculated from Kaplan-Meier survival curve)]. Sensitivity analysis was performed to assess robustness of results by removing one study each time. To evaluate publication bias, Begg’s test was conducted and funnel plot was generated according to lnHRs and their respective SEs. Statistical analysis was performed using Review Manager (RevMan) Version 5.2 software (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2012) and Stata 11.0 softwares (Stata Corp., College Station, TX, USA).

3. Results 3.1. Study characteristics Initial search identified 652 citations. We firstly removed 180 duplicate studies; 414 studies were then excluded after screening their titles or abstracts. A total of 58 potentially relevant full text articles were assessed. Finally, a total of 24 studies met the inclusion criteria (Fig. 1). These studies were published between 1992 and 2013. A total of 2280 patients analyzed with TS protein expression were included in the meta-analysis. Eighteen studies were performed in Asia [7,14,18–20, 23,29–39], 4 in Europe [17,21,40,41] and 2 in America [16,42]. OS or PFS were compared according to tumor TS protein expression levels (high/low or positive/negative) (Table 1). Overall, the total quality score of each included study ranged from 24 to 33, with a median score of 29.5 (Table 2). 3.2. Meta-analysis A total of 13 HRs for PFS [17,19–21,29–33,35,37, 41,42] and 20 HRs for OS [7,14,16–18,21–23,29–32, 34,36,38–42] were available for pooled HR calculation. Meta-analysis by a random effect model showed that patients with low TS protein expression had significantly longer OS than those with high TS expres-

69

sion (pooled HR, 0.51; 95% CI, 0.36–0.73) (Fig. 2A). The pooled HR for PFS was 0.49 (95% CI, 0.37–0.67) (Fig. 2B), indicating that reduced TS expression in lung tumor tissue was significantly associated with a good PFS in NSCLC patients. 3.3. Subgroup analysis We conducted subgroup analyses according to TStargeted drug use status. For OS, pooled HRs were statistically significant for pemetrexed subgroup (HR, 0.42; 95% CI, 0.21–0.83) and 5-FU subgroup (HR, 0.34; 95% CI, 0.22–0.53), but not for no TS-targeted drug subgroup (patients receiving neither pemetrexed nor 5-FU; HR, 0.68; 95% CI, 0.43–1.09) (Fig. 2A). For PFS, pooled HR was also significant for pemetrexed subgroup (HR, 0.45; 95% CI, 0.33–0.60), but not for no TS-targeted drug subgroup (HR, 0.80; 95% CI, 0.40–1.61) (Fig. 2B). We further conducted subgroup analyses in the no TS-targeted drug and pemetrexed subgroups (subsubgroup analysis) according to region of studies (Asia and non-Asia) and HR estimation (original HR and KM-HR) (Fig. 3). For patients who did not receive TStargeted drugs, pooled HR for OS was significant for Asia subgroup (HR, 0.50; 95% CI, 0.37–0.66), but not significant for non-Asia subgroups (HR, 1.41; 95% CI, 0.74–2.69). The pooled results were also not significant in subgroups of original HR and KM-HR (HR, 0.66; 95% CI, 0.29–1.50, and HR, 0.73; 95% CI, 0.45– 1.16, respectively) (Fig. 3A). For patients treated with pemetrexed, TS protein expression was significantly associated with OS for non-Asia subgroup (HR, 0.26; 95% CI, 0.09–0.75), but not for Asia subgroup (HR, 0.55; 95% CI, 0.21–1.45). When pooling HRs according to HR estimation, both results were significant for OS (HR, 0.39; 95% CI, 0.16–0.94 for original HR; HR, 0.44; 95% CI, 0.20–0.97 for KM-HR) (Fig. 3B). When analyzing PFS for patients who received pemetrexed treatment, results were significant for all subgroups (Fig. 3C). Because of limited studies, we did not conduct subgroup analyses in 5-FU subgroups and no TS-targeted drug subgroup for PFS. In order to explore clinical heterogeneity, subgroup analyses were executed according to clinical stage of NSCLC for all available cases. For both OS and PFS analysis, pooled HRs were statistically significant for advanced stage (III-IV) subgroups (HR, 0.39; 95% CI, 0.16–0.95, and HR, 0.48; 95% CI, 0.35–0.66, respectively), but not for early stage (I-II) subgroups (HR, 0.93; 95% CI, 0.15–5.68, and HR 0.82; 95% CI, 0.17– 3.96, respectively) (data not shown).

70

Q. Liu et al. / Thymidylate synthase and prognosis of NSCLC Table 2 Scores of quality assessment of each included studies Studies Chen 2011 [20] Chang 2010 [29] Lee 2013 [30] Duan 2012 [31] Igawa 2012 [32] Gadgeel 2011 [42] Sun 2011 [19] Takezawa 2011 [33] Christoph 2013 [21] Wang 2013 [18] Huang 2005 [34] Nakano 2006 [14] Wynes 2012 [17] SHINTANI 2003 [35] MIYOSHI 2007 [23] Nakagawa 2002 [36] Shimokawa 2011 [37] Kaira 2012 [7] Zheng 2008 [16] YOKOMISE 2013 [38] VOLM 1992 [40] Huang 2000 [22] Nicolson 2013 [41] NAKAGAWA 2013 [39] Median

Scientific 7 6 7 6 7 7 6 6 7 8 6 7 7 6 6 5 5 7 6 6 6 5 10 7 6

Laboratory 9 8 11 6 13 12 11 12 12 8 10 12 8 10 11 11 12 8 12 10 11 12 9 6 11

3.4. Sensitivity analysis To assess robustness of results, we conducted sensitivity analyses for OS and PFS by removing one study and then pooling the results each time. We found that individual studies did not remarkably modify the estimates. Range of pooled HR for OS and PFS were 0.48–0.55 and 0.46–0.54, respectively. When individual studies were omitted each time, pooled HR for OS and PFS remained significant in all tests (data not shown). Since no significant heterogeneity was found in some subgroups (P > 0.05, I 2 < 50%), we also applied a fixed effects model to pool results, and found that the fixed and random effects models generated the consistent results (data not shown). 3.5. Publication bias Publication bias was evaluated with Begg’s test and funnel plot. P-value of Begg’s test for OS and PFS was 0.153 and 0.113, respectively, indicating no publication bias (Fig. 4).

4. Discussion In this systematic review and meta-analysis, we enrolled 24 eligible studies comparing OS and/or PFS

Generalizability 9 6 9 8 9 8 7 10 7 11 8 8 7 6 9 8 7 7 8 9 3 8 8 8 8

Results 3 7 5 7 4 6 5 3 5 5 6 6 6 6 3 5 8 5 7 4 4 6 5 3 5

Total score 28 27 32 27 33 33 29 31 31 32 30 33 28 28 29 29 32 27 33 29 24 31 32 24 29.5

according to tumor TS protein expression of NSCLC. The results suggest that low tumor TS protein expression is significantly associated with better outcome for NSCLC patients. However, subgroup analyses indicate that this significant association is limited to the patients with TS-targeted treatment (pemetrexed or 5-FU). TS is an essential enzyme for de novo DNA synthesis. It was reported that TS expression was related to proliferative activity of cancer cells [43–46]. If TS expression is low in tumor tissue or its catalytic activity is inhibited, DNA replication and cell proliferation may be inhibited. High TS expression level is related to poor outcome in several types of cancer [47–53]. Both pemetrexed and 5-FU are TS inhibitors. Pemetrexed is a folate analog metabolic inhibitor that exerts its action by inhibiting three enzymes involved in the de novo biosynthesis of thymidine and purine nucleotides: TS, dihydrofolate reductase (DHFR) and glycinamide ribonucleotide formyltransferase (GARFT) [54,55]. TS is the primary target of pemetrexed [54,56]. The association between TS overexpression and pemetrexedresistant of tumor cell lines was observed previously in vitro studies [57–60]. A phase II trial suggested that lower TS expression was associated with enhanced clinical activity of pemetrexed for breast cancer [61]. Besides, it was reported that TYMS polymorphisms, which may influence TS expression, mRNA stability and/or TS levels, were also related to pemetrexed ef-

Q. Liu et al. / Thymidylate synthase and prognosis of NSCLC

71

A

B Fig. 2. Association of TS expression and survival, with subgroup analyses. (A) Association between TS expression and overall survival (OS), with subgroup analysis according to TS-targeted drug use; (B) Association between TS expression and progression-free survival (PFS). Squares indicate the HR; and horizontal lines represent 95% CI. The pooled results are indicated by the black diamond. TS, thymidylate synthase; SE, standard error, CI, confident interval. For study of Huang 2005, ∗ indicates HR of stage I lung cancer while ∗∗ indicates HR of stage II–III.

72

Q. Liu et al. / Thymidylate synthase and prognosis of NSCLC

A Fig. 3. Subgroup analyses of TS-targeted subgroups according to region of study and HR estimation. (A) Subgroup analyses for OS of patients who did not receive TS-targeted therapies; (B) Subgroup analyses for OS of patients treated with pemetrexed; (C) Subgroup analyses for PFS of patients treated with pemetrexed. Squares indicate the HR; and horizontal lines represent 95% CI. The pooled results are indicated by the black diamond. OS, overall survival; PFS, progression-free survival; HR, hazard ratio; TS, thymidylate synthase; SE, standard error; CI, confident interval; Original HR, hazard ratio (HR) was available from original full text; KM-HR, HR was calculated from Kaplan-Meier survival curve. For study of Huang 2005, ∗ indicates HR of stage I lung cancer while ∗∗ indicates HR of stage II–III.

ficacy [18,62–64]. 5-FU is an example of a rationally designed anticancer agent, which is a component of the standard therapy for a variety of malignancies, including gastrointestinal cancers, head and neck cancer, breast cancer, and NSCLC [9]. 5-FdUMP, a main active metabolite of 5-FU, inhibits TS activity and then inhibits methylation of carbon 5 of dUMP to form

dTMP, which is essential for DNA synthesis. Previous studies reported that TS expression was associated with clinical efficacy of 5-FU and outcome for some types of cancer [65,66]. The reason why high TS expression is associated with poor response and outcome of pemetrexed and 5-FU may be the limited ability of TS inhibition of the two drugs. But the exact mech-

Q. Liu et al. / Thymidylate synthase and prognosis of NSCLC

73

B Fig. 3. continued.

anism needs further exploration. Our subgroup analysis also indicates that TS protein expression is significantly associated with OS and PFS in NSCLC patients with pemetrexed or 5-FU treatment, suggesting the clinical predictive value of TS expression to identify patients more likely to benefit from TS-targeted treatment. In this study, we also assessed the potential influence of regions of studies and HR estimation methods. For patients who did not receive TS-targeted drugs, TS expression was not significantly associated with OS and PFS. In subgroup analysis, we found that TS expression was significantly associated with OS in Asia sub-

group, but not in non-Asia subgroup (Fig. 3A). This result indicates that the prognostic value of TS may vary with ethnic differences, and needs to be further validated. When we conducted subgroup analyses according to HR estimation methods, pooled HRs of each subgroup kept consistent with the overall pooled results. This indicated that the method we used in this article to calculate HRs from Kaplan-Meier curve did not bring obvious bias. When subgroup analyses were conducted according to region of study in no TS-targeted drug subgroup and pemetrexed subgroup, the results were not consistent for OS, which may be due to the limited number of

74

Q. Liu et al. / Thymidylate synthase and prognosis of NSCLC

C Fig. 3. continued.

non-Asia studies. However, we found that TS expression could be a robust predictive marker of PFS for pemetrexed therapy, because the pooled HRs for PFS were significant for all subgroup patients treated with pemetrexed. Recently three meta-analyses reported relevant results with ours [67–69]. However, these three studies did not include the studies of 5-FU and no TS-

targeted treatment; all the three meta-analyses aimed to evaluate predictive value of TS expression for pemetrexed efficacy for NSCLC. For patients not treated with pemetrexed, they did not assess prognostic significance of TS expression, which was analyzed in our study. Moreover, unlike our analysis only focusing on protein level, these three meta-analyses pooled results derived from different TS expression level

Q. Liu et al. / Thymidylate synthase and prognosis of NSCLC

2

75

Begg's funnel plot with pseudo 95% confidence limits

lnhr

0

-2

-4 .5 s.e. of: lnhr

0

1

A 1

Begg's funnel plot with pseudo 95% confidence limits

lnhr

0

-1

-2 0

.2

.4 s.e. of: lnhr

.6

.8

B Fig. 4. Begg’s funnel plots for OS (Fig. 4A) and PFS (Fig. 4B).

(mRNA and protein), which probably led to heterogeneity. All the three meta-analyses reported accordant results about the association between TS expression and survival/response rate for pemetrexed, except one reported that the OS/PFS differences between -/low and +/high TS expression was not statistically significant [69], which may be explained by the index they used (median ratio). Several limitations should be considered when interpreting our results. Significant heterogeneity was

found in this meta-analysis. Thus, the random effect model was used to pool the results. Subgroup analyses were also conducted to explore the potential source of heterogeneity. Different drug use, study region, HR estimation and clinical stage may explain part of heterogeneity. However, some important clinical characteristics were not available for heterogeneity analysis in our meta-analysis. The heterogeneity and unknown confounders may still influence the true association. In addition, there is no standard method to evaluate

76

Q. Liu et al. / Thymidylate synthase and prognosis of NSCLC

TS protein level; cutoff values of TS (high/low or positive/negative) expression were not consistent among studies. Sensibility and accuracy may vary among different studies, which possibly also contribute to heterogeneity. In conclusion, this meta-analysis suggests that low TS protein expression is associated with better outcome for NSCLC patients treated with TS-targeted drugs. TS may be a useful predictive marker in personalized therapy for NSCLC. Future randomized clinical trials are warranted to elucidate the predictive and prognostic value of TS in the well-defined subgroups of lung cancer patients.

Acknowledgements This work was supported by National Natural Science Foundation of China (No. 81171903 to Y.L.) and Chongqing Natural Science Foundation of China (No. CSTC 2011BB5024 to Y.L.).

Conflicts of interest None declared.

References [1]

[2]

[3]

[4]

[5]

J. Ferlay, H.R. Shin, F. Bray, D. Forman, C. Mathers and D.M. Parkin, Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008, Int J Cancer 127 (2010), 2893-917. N. Hanna, F.A. Shepherd, F.V. Fossella, J.R. Pereira, F. De Marinis, J. von Pawel, U. Gatzemeier, T.C. Tsao, M. Pless, T. Muller, H.L. Lim, C. Desch, K. Szondy, R. Gervais, Shaharyar, C. Manegold, S. Paul, P. Paoletti, L. Einhorn and P.A. Bunn, Jr., Randomized phase III trial of pemetrexed versus docetaxel in patients with non-small-cell lung cancer previously treated with chemotherapy, J Clin Oncol 22 (2004), 1589-97. A. Spira and D.S. Ettinger, Multidisciplinary management of lung cancer, The New England journal of medicine 350 (2004), 379-92. C. Mascaux, N. Iannino, B. Martin, M. Paesmans, T. Berghmans, M. Dusart, A. Haller, P. Lothaire, A.P. Meert, S. Noel, J.J. Lafitte and J.P. Sculier, The role of RAS oncogene in survival of patients with lung cancer: A systematic review of the literature with meta-analysis, Br J Cancer 92 (2005), 131-9. L.X. Liu, X.Y. Shao, W. Gao, J.L. Bai, R.S. Wang, P.W. Huang, Y.M. Yin, P. Liu and Y.Q. Shu, The Role of Human Epidermal Growth Factor Receptor 2 as a Prognostic Factor in Lung Cancer A Meta-Analysis of Published Data, Journal of Thoracic Oncology 5 (2010), 1922-1932.

[6]

C. Chung and M. Christianson, Predictive and prognostic biomarkers with therapeutic targets in breast, colorectal, and non-small cell lung cancers: A systemic review of current development, evidence, and recommendation, J Oncol Pharm Pract (2013). [7] K. Kaira, Y. Ohde, K. Nakagawa, T. Okumura, H. Murakami, T. Takahashi, H. Kondo, T. Nakajima, M. Endo and N. Yamamoto, Thymidylate synthase expression is closely associated with outcome in patients with pulmonary adenocarcinoma, Medical oncology (Northwood, London, England) 29 (2012), 1663-72. [8] E. Galvani, G.J. Peters and E. Giovannetti, Thymidylate synthase inhibitors for non-small cell lung cancer, Expert opinion on investigational drugs 20 (2011), 1343-1356. [9] M. Malet-Martino and R. Martino, Clinical studies of three oral prodrugs of 5-fluorouracil (capecitabine, UFT, S-1): a review, Oncologist 7 (2002), 288-323. [10] H. Hochster, The role of pemetrexed in the treatment of colorectal cancer, Semin Oncol 29 (2002), 54-6. [11] H.S. Hochster, The role of pemetrexed in the treatment of gastrointestinal malignancy, Clin Colorectal Cancer 4 (2004), 190-5. [12] E. Chu, M.A. Callender, M.P. Farrell and J.C. Schmitz, Thymidylate synthase inhibitors as anticancer agents: from bench to bedside, Cancer Chemotherapy and Pharmacology 52(Suppl 1) (2003), S80-9. [13] M.G. Rose, M.P. Farrell and J.C. Schmitz, Thymidylate synthase: A critical target for cancer chemotherapy, Clin Colorectal Cancer 1 (2002), 220-9. [14] J. Nakano, C. Huang, D. Liu, D. Masuya, T. Nakashima, H. Yokomise, M. Ueno, H. Wada and M. Fukushima, Evaluations of biomarkers associated with 5-FU sensitivity for non-smallcell lung cancer patients postoperatively treated with UFT, Br J Cancer 95 (2006), 607-15. [15] P.P. Grimminger, P.M. Schneider, R. Metzger, D. Vallbohmer, A.H. Holscher, P.V. Danenberg and J. Brabender, Low thymidylate synthase, thymidine phosphorylase, and dihydropyrimidine dehydrogenase mRNA expression correlate with prolonged survival in resected non-small-cell lung cancer, Clin Lung Cancer 11 (2010), 328-34. [16] Z. Zheng, X. Li, M.J. Schell, T. Chen, D. Boulware, L. Robinson, E. Sommers and G. Bepler, Thymidylate synthase in situ protein expression and survival in stage I nonsmall-cell lung cancer, Cancer 112 (2008), 2765-73. [17] M.W. Wynes, K. Konopa, S. Singh, B. Reyna-Asuncion, J. Ranger-Moore, A. Sternau, D.C. Christoph, R. Dziadziuszko, J. Jassem and F.R. Hirsch, Thymidylate synthase protein expression by IHC and gene copy number by SISH correlate and show great variability in non-small cell lung cancer, Journal of Thoracic Oncology: Official Publication of the International Association for the Study of Lung Cancer 7 (2012), 982-92. [18] X. Wang, Y. Wang, J. Cheng and M. Ha, Association of thymidylate synthase gene 3’-untranslated region polymorphism with sensitivity of non-small cell lung cancer to pemetrexed treatment: TS gene polymorphism and pemetrexed sensitivity in NSCLC, J Biomed Sci 20 (2013), 5. [19] J.M. Sun, J. Han, J.S. Ahn, K. Park and M.J. Ahn, Significance of thymidylate synthase and thyroid transcription factor 1 expression in patients with nonsquamous non-small cell lung cancer treated with pemetrexed-based chemotherapy, J Thorac Oncol 6 (2011), 1392-9. [20] C.Y. Chen, Y.L. Chang, J.Y. Shih, J.W. Lin, K.Y. Chen, C.H. Yang, C.J. Yu and P.C. Yang, Thymidylate synthase and di-

Q. Liu et al. / Thymidylate synthase and prognosis of NSCLC hydrofolate reductase expression in non-small cell lung carcinoma: The association with treatment efficacy of pemetrexed, Lung Cancer 74 (2011), 132-8. [21] D.C. Christoph, B.R. Asuncion, B. Hassan, C. Tran, J.D. Maltzman, D.J. O’Shannessy, M.W. Wynes, T.C. Gauler, J. Wohlschlaeger, M. Hoiczyk, M. Schuler, W.E. Eberhardt and F.R. Hirsch, Significance of folate receptor alpha and thymidylate synthase protein expression in patients with nonsmall-cell lung cancer treated with pemetrexed, J Thorac Oncol 8 (2013), 19-30. [22] C.L. Huang, H. Yokomise, S. Kobayashi, M. Fukushima, S. Hitomi and H. Wada, Intratumoral expression of thymidylate synthase and dihydropyrimidine dehydrogenase in non-small cell lung cancer patients treated with 5-FU-based chemotherapy, Int J Oncol 17 (2000), 47-54. [23] T. Miyoshi, K. Kondo, H. Toba, M. Yoshida, H. Fujino, K. Kenzaki, S. Sakiyama, M. Takehisa and A. Tangoku, Predictive value of thymidylate synthase and dihydropyrimidine dehydrogenase expression in tumor tissue, regarding the efficacy of postoperatively administered UFT (tegafur+uracil) in patients with non-small cell lung cancer, Anticancer Research 27 (2007), 2641-8. [24] D.F. Stroup, J.A. Berlin, S.C. Morton, I. Olkin, G.D. Williamson, D. Rennie, D. Moher, B.J. Becker, T.A. Sipe and S.B. Thacker, Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group, JAMA 283 (2000), 2008-12. [25] J. Little, L. Bradley, M.S. Bray, M. Clyne, J. Dorman, D.L. Ellsworth, J. Hanson, M. Khoury, J. Lau, T.R. O’Brien, N. Rothman, D. Stroup, E. Taioli, D. Thomas, H. Vainio, S. Wacholder and C. Weinberg, Reporting, appraising, and integrating data on genotype prevalence and gene-disease associations, American journal of epidemiology 156 (2002), 300-10. [26] J.F. Tierney, L.A. Stewart, D. Ghersi, S. Burdett and M.R. Sydes, Practical methods for incorporating summary time-toevent data into meta-analysis, Trials 8 (2007), 16. [27] E. Steels, M. Paesmans, T. Berghmans, F. Branle, F. Lemaitre, C. Mascaux, A.P. Meert, F. Vallot, J.J. Lafitte and J.P. Sculier, Role of p53 as a prognostic factor for survival in lung cancer: a systematic review of the literature with a meta-analysis, European Respiratory Journal 18 (2001), 705-19. [28] J.P. Higgins, S.G. Thompson, J.J. Deeks and D.G. Altman, Measuring inconsistency in meta-analyses, BMJ 327 (2003), 557-60. [29] M.H. Chang, J.S. Ahn, J. Lee, K.H. Kim, Y.H. Park, J. Han, M.J. Ahn and K. Park, The efficacy of pemetrexed as a thirdor fourth-line therapy and the significance of thymidylate synthase expression in patients with advanced non-small cell lung cancer, Lung Cancer 69 (2010), 323-9. [30] S.H. Lee, K.B. Noh, J.S. Lee, E.J. Lee, K.H. Min, G.Y. Hur, S.Y. Lee, J.H. Kim, C. Shin, J.J. Shim, C.H. Kim, K.H. Kang and K.H. In, Thymidylate synthase and ERCC1 as predictive markers in patients with pulmonary adenocarcinoma treated with pemetrexed and cisplatin, Lung Cancer (2013). [31] J.C. Duan, M.N. Wu, J. Zhao, T.T. An, L. Yang, H. Bai, Z.J. Wang, M.L. Zhuo, S.H. Wang, Y.Y. Wang, X.H. Liu and J. Wang, [Effect of pemetrexed plus platinum for chemotherapynaive advanced non-small cell lung cancer], Zhonghua Jie He He Hu Xi Za Zhi 35 (2012), 97-101. [32] S. Igawa, S. Ryuge, M. Wada, S. Otani, S. Maki, A. Takakura, K. Katono, J. Sasaki, Y. Sato and N. Masuda, Pemetrexed for previously treated patients with non-small cell lung cancer

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

77

and differences in efficacy according to thymidylate synthase expression, Chemotherapy 58 (2012), 313-20. K. Takezawa, I. Okamoto, W. Okamoto, M. Takeda, K. Sakai, S. Tsukioka, K. Kuwata, H. Yamaguchi, K. Nishio and K. Nakagawa, Thymidylate synthase as a determinant of pemetrexed sensitivity in non-small cell lung cancer, Br J Cancer 104 (2011), 1594-601. C. Huang, D. Liu, D. Masuya, T. Nakashima, K. Kameyama, S. Ishikawa, M. Ueno, R. Haba and H. Yokomise, Clinical application of biological markers for treatments of resectable non-small-cell lung cancers, Br J Cancer 92 (2005), 1231-9. Y. Shintani, M. Ohta, H. Hirabayashi, H. Tanaka, K. Iuchi, K. Nakagawa, H. Maeda, T. Kido, S. Miyoshi and H. Matsuda, New prognostic indicator for non-small-cell lung cancer, quantitation of thymidylate synthase by real-time reverse transcription polymerase chain reaction, Int J Cancer 104 (2003), 790-5. T. Nakagawa, F. Tanaka, Y. Otake, K. Yanagihara, R. Miyahara, K. Matsuoka, T. Takata, T. Yamada, M. Fukushima and H. Wada, Prognostic value of thymidylate synthase expression in patients with p-stage I adenocarcinoma of the lung, Lung Cancer 35 (2002), 165-70. H. Shimokawa, H. Uramoto, T. Onitsuka, T. Iwata, M. Nakagawa, K. Ono and T. Hanagiri, TS expression predicts postoperative recurrence in adenocarcinoma of the lung, Lung cancer (Amsterdam, Netherlands) 72 (2011), 360-4. H. Yokomise, D. Liu, S. Chang, T. Go, S. Ishikawa, N. Misaki and N. Nakashima, Biomarkers as prognostic factors for cN2 or 3 non-small cell lung cancer treated by induction chemoradiotherapy and surgery, Anticancer Research 33 (2013), 110715. Y. Nakagawa, T. Shimizu, N. Takahashi and S. Hashimoto, Impact of thymidylate synthase protein expression on efficacy of chemotherapy in advanced lung cancer patients, Mol Clin Oncol 1 (2013), 411-417. M. Volm and J. Mattern, Elevated expression of thymidylate synthase in doxorubicin resistant human non small cell lung carcinomas, Anticancer Research 12 (1992), 2293-6. M.C. Nicolson, D.A. Fennell, D. Ferry, K. O’Byrne, R. Shah, V. Potter, G. Skailes, S. Upadhyay, P. Taylor, V. Andre, T.S. Nguyen, S.P. Myrand, C. Visseren-Grul, M. Das and K.M. Kerr, Thymidylate synthase expression and outcome of patients receiving pemetrexed for advanced nonsquamous nonsmall-cell lung cancer in a prospective blinded assessment phase II clinical trial, J Thorac Oncol 8 (2013), 930-9. S.M. Gadgeel, J.C. Ruckdeschel, B.B. Patel, A. Wozniak, A. Konski, M. Valdivieso, D. Hackstock, W. Chen, K. Belzer, A.M. Burger, L. Marquette and A. Turrisi, Phase II study of pemetrexed and cisplatin, with chest radiotherapy followed by docetaxel in patients with stage III non-small cell lung cancer, J Thorac Oncol 6 (2011), 927-33. L.G. Navalgund, C. Rossana, A.J. Muench and L.F. Johnson, Cell cycle regulation of thymidylate synthetase gene expression in cultured mouse fibroblasts, J Biol Chem 255 (1980), 7386-90. B.C. Pestalozzi, C.J. McGinn, T.J. Kinsella, J.C. Drake, M.C. Glennon, C.J. Allegra and P.G. Johnston, Increased thymidylate synthase protein levels are principally associated with proliferation but not cell cycle phase in asynchronous human cancer cells, British Journal of Cancer 71 (1995), 1151-7. M. Derenzini, L. Montanaro, D. Trere, A. Chilla, P.L. Tazzari, F. Dall’Olio and D. Ofner, Thymidylate synthase protein expression and activity are related to the cell proliferation rate in human cancer cell lines, Mol Pathol 55 (2002), 310-4.

78

Q. Liu et al. / Thymidylate synthase and prognosis of NSCLC

[46]

T. Nakagawa, Y. Otake, K. Yanagihara, R. Miyahara, S. Ishikawa, M. Fukushima, H. Wada and F. Tanaka, Expression of thymidylate synthase is correlated with proliferative activity in non-small cell lung cancer (NSCLC), Lung Cancer 43 (2004), 145-149. [47] H.B. Hu, L. Kuang, X.M. Zeng, B. Li, E.Y. Liu and M.Z. Zhong, Predictive value of thymidylate synthase expression in gastric cancer: A systematic review with meta-analysis, Asian Pacific Journal of Cancer Prevention: APJCP 13 (2012), 2617. [48] T. Yamachika, H. Nakanishi, K. Inada, T. Tsukamoto, T. Kato, M. Fukushima, M. Inoue and M. Tatematsu, A new prognostic factor for colorectal carcinoma, thymidylate synthase, and its therapeutic significance, Cancer 82 (1998), 70-7. [49] D. Edler, M. Hallström, P.G. Johnston, I. Magnusson, P. Ragnhammar and H. Blomgren, Thymidylate synthase expression: an independent prognostic factor for local recurrence, distant metastasis, disease-free and overall survival in rectal cancer, Clinical Cancer Research 6 (2000), 1378-1384. [50] M. Suzuki, S. Tsukagoshi, Y. Saga, M. Ohwada and I. Sato, Enhanced expression of thymidylate synthase may be of prognostic importance in advanced cervical cancer, Oncology 57 (1999), 50-54. [51] P.G. Johnston, E.R. Fisher, H.E. Rockette, B. Fisher, N. Wolmark, J.C. Drake, B.A. Chabner and C.J. Allegra, The role of thymidylate synthase expression in prognosis and outcome of adjuvant chemotherapy in patients with rectal cancer, J Clin Oncol 12 (1994), 2640-7. [52] H.J. Lenz, C.G. Leichman, K.D. Danenberg, P.V. Danenberg, S. Groshen, H. Cohen, L. Laine, P. Crookes, H. Silberman, J. Baranda, Y. Garcia, J. Li and L. Leichman, Thymidylate synthase mRNA level in adenocarcinoma of the stomach: A predictor for primary tumor response and overall survival, Journal of Clinical Oncology: Ifficial Journal of the American Society of Clinical Oncology 14 (1996), 176-82. [53] D.H. Harpole, Jr., M.B. Moore, J.E. Herndon, 2nd, T. Aloia, T.A. D’Amico, T. Sporn, A. Parr, I. Linoila and C. Allegra, The prognostic value of molecular marker analysis in patients treated with trimodality therapy for esophageal cancer, Clin Cancer Res 7 (2001), 562-9. [54] C. Shih, V.J. Chen, L.S. Gossett, S.B. Gates, W.C. MacKellar, L.L. Habeck, K.A. Shackelford, L.G. Mendelsohn, D.J. Soose and V.F. Patel, LY231514, a pyrrolo [2, 3-d] pyrimidinebased antifolate that inhibits multiple folate-requiring enzymes, Cancer Research 57 (1997), 1116-1123. [55] A.R. Hanauske, V. Chen, P. Paoletti and C. Niyikiza, Pemetrexed disodium: A novel antifolate clinically active against multiple solid tumors, Oncologist 6 (2001), 363-73. [56] G. Grindey, C. Shih, C. Barnett, H. Pearce, J. Engelhardt, G. Todd, S. Rinzel, J. Worzalla, L. Gossett and T. Everson, LY231514, a novel pyrrolopyrimidine antifolate that inhibits thymidylate synthase (TS), in: Proc Am Assoc Cancer Res, 1992, pp. 411. [57] A.-R. Hanauske, U. Eismann, O. Oberschmidt, H. Pospisil, S. Hoffmann, H. Hanauske-Abel, D. Ma, V. Chen, P. Paoletti and C. Niyikiza, In vitro chemosensitivity of freshly explanted tumor cells to pemetrexed is correlated with target gene expression, Investigational New Drugs 25 (2007), 417-23. [58] J. Sigmond, H.H.J. Backus, D. Wouters, O.H. Temmink, G.

[59]

[60]

[61]

[62]

[63]

[64]

[65]

[66]

[67]

[68]

[69]

Jansen and G.J. Peters, Induction of resistance to the multitargeted antifolate Pemetrexed (ALIMTA) in WiDr human colon cancer cells is associated with thymidylate synthase overexpression, Biochemical Pharmacology 66 (2003), 431-8. E. Giovannetti, V. Mey, S. Nannizzi, G. Pasqualetti, L. Marini, M. Del Tacca and R. Danesi, Cellular and pharmacogenetics foundation of synergistic interaction of pemetrexed and gemcitabine in human non-small-cell lung cancer cells, Mol Pharmacol 68 (2005), 110-8. H. Ozasa, T. Oguri, T. Uemura, M. Miyazaki, K. Maeno, S. Sato and R. Ueda, Significance of thymidylate synthase for resistance to pemetrexed in lung cancer, Cancer Sci 101 (2010), 161-6. H.L. Gomez, S.L. Santillana, C.S. Vallejos, R. Velarde, J. Sanchez, X. Wang, N.L. Bauer, R.D. Hockett, V.J. Chen and C. Niyikiza, A phase II trial of pemetrexed in advanced breast cancer: clinical response and association with molecular target expression, Clinical Cancer Research 12 (2006), 832-838. W.J. Li, H. Jiang, X.J. Fang, H.L. Ye, M.H. Liu, Y.W. Liu, Q. Chen, L. Zhang, J.Y. Zhang, C.L. Yuan and Q.Y. Zhang, Polymorphisms in thymidylate synthase and reduced folate carrier (SLC19A1) genes predict survival outcome in advanced nonsmall cell lung cancer patients treated with pemetrexed-based chemotherapy, Oncology Letters 5 (2013), 1165-1170. D. Kasai, H. Ozasa, T. Oguri, M. Miyazaki, T. Uemura, O. Takakuwa, E. Kunii, H. Ohkubo, K. Maeno and A. Niimi, Thymidylate Synthase Gene Copy Number as a Predictive Marker for Response to Pemetrexed Treatment of Lung Adenocarcinoma, Anticancer Research 33 (2013), 1935-1940. A. Lima, R. Azevedo, H. Sousa, V. Seabra and R. Medeiros, Current approaches for TYMS polymorphisms and their importance in molecular epidemiology and pharmacogenetics, Pharmacogenomics 14 (2013), 1337-51. H.J. Lenz, C.G. Leichman, K.D. Danenberg, P.V. Danenberg, S. Groshen, H. Cohen, L. Laine, P. Crookes, H. Silberman, J. Baranda, Y. Garcia, J. Li and L. Leichman, Thymidylate synthase mRNA level in adenocarcinoma of the stomach: a predictor for primary tumor response and overall survival, Journal of Clinical Oncology 14 (1996), 176-82. P.G. Johnston, H.J. Lenz, C.G. Leichman, K.D. Danenberg, C.J. Allegra, P.V. Danenberg and L. Leichman, Thymidylate synthase gene and protein expression correlate and are associated with response to 5-fluorouracil in human colorectal and gastric tumors, Cancer Res 55 (1995), 1407-12. Y. Liu, T.-J. Yin, R. Zhou, S. Zhou, L. Fan and R.-G. Zhang, Expression of thymidylate synthase predicts clinical outcomes of pemetrexed-containing chemotherapy for nonsmall-cell lung cancer: a systemic review and meta-analysis, Cancer Chemotherapy and Pharmacology (2013), 1-8. T. Wang, C.C. Pan, J.R. Yu, Y. Long, X.H. Cai, X. De Yin, L.Q. Hao and L.L. Luo, Association between TYMS Expression and Efficacy of Pemetrexed–Based Chemotherapy in Advanced Non-Small Cell Lung Cancer: A Meta-Analysis, PLoS One 8 (2013), e74284. L. Wang, R. Wang, Y. Pan, Y. Sun, J. Zhang and H. Chen, The pemetrexed-containing treatments in the non-small cell lung cancer, is -/low thymidylate synthase expression better than +/high thymidylate synthase expression: A metaanalysis, BMC Cancer 14 (2014), 205.