Introduction

Lung cancer, a major serious public health problem, is one of the most common malignant tumors and has become the leading cause of cancer-related deaths in the world1, with more than one million deaths from this disease annually2,3. There are two main types of lung cancer: non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC), of which NSCLC accounts for approximately 80%–85% and nearly 70% of patients present with the locally advanced stages (such as stage IIIB or IV) or metastatic disease at the time of diagnosis, losing the opportunity of surgical resection and making curative surgery impossible3,4,5,6,7,8,9. Although intensive effort has been made to improve the efficacy of lung cancer diagnosis and therapy in the last decades, the overall five-year survival rate still remains only about 15% and even lower in China1,2.

The treatment of lung cancer patients is surgery for early stages, whereas chemotherapy regimen is the main conventional therapeutic method for locally advanced stages and metastatic cancers. Currently, the platinum-based chemotherapy is one of the most extensively accepted and used treatments in advanced lung cancer patients, which has been shown to improve overall survival (OS)10,11,12. However, individual in clinical practice, the therapeutic efficacy of chemotherapy varies remarkably among different individuals, with a response rate from 26% to 60% in lung cancer patients13,14. Some individuals response to the chemotherapy, while others confer intrinsic or acquired drug resistance15. It has been well recognized that genetic factors are considered to play an important role in disease development, drug response, treatment effectiveness and survival of lung cancer, thus affect the prognosis of patients8,14. Seeking optimal therapeutic and prognostic biomarkers, which could improve prognostic and predictive assessment accuracy to help early detection, better chemotherapeutic agents, drug sensitivity, developing personalized cancer treatment and patient-tailored chemotherapy and eventually achieving better outcomes for patients7,8,16. The therapeutic efficacy of platinum-based chemotherapy treatment remains a research hotpot in this field7,8,16. However, the reliable biomarkers are still lacking in the clinical practice.

It has been speculated that the single nucleotide polymorphisms (SNPs) or mutations occurring in the DNA repair pathways may alter gene expression and the activity of DNA repair17,18, thus influence the effectiveness of cancer therapy, prognosis and survival of patients19. The DNA repair pathways mainly include the base excision repair (BER), nucleotide excision repair (NER), DNA double strand break repair (DDSBR) and mismatch repair (MMR)20. The X-ray repair cross-complementing group 1 (XRCC1) gene, located on chromosome 19q13.2, has long been recognized as a central role in BER pathway by interacting with other DNA repair proteins, acting as a “scaffold” in both DNA single-strand break repair and BER activities21,22,23,24,25. It has been proposed that the XRCC1 protein is critical for repairing other types of DNA damage induced by the platinum-based anticancer drugs (such as cisplatin (DDP) and carboplatin (CBP))26, also including DNA double-strand breaks27,28. Therefore, genetic variants in XRCC1 gene might modulate DNA repair capacity and hence markedly influence the anticancer effect of platinum drugs, the efficacy of platinum-based chemotherapy treatment and the prognosis and survival of patients. Although a great number of studies have extensively investigated the pharmacogenetics of XRCC1 genetic variants (such as rs1799782, C → T, arginine (Arg)194 tryptophan (Trp), exon 6; rs25489, G → A, Arg280 histidine (His), exon 6; rs25487, G → A, Arg399glutanine (Gln), exon 10; rs3213245, C→T, T-77C, 5′-untranslated region (5′-UTR); and rs3213239 (insertion variation GGCC), http://www.ncbi.nlm.nih.gov/projects/SNP) on clinical outcomes of platinum-based chemotherapy in advanced lung cancer patients, the observed results from these studies remain conflicting rather than conclusive6,7,10,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78. Because a single study may be too underpowered to detect slight effects of the genetic variants on cancer, especially when the sample size is relatively small, the meta-analysis could provide more comprehensive and reliable conclusions through systematically summarizing existed data. There are several meta-analyses reported the inconsistent results for evaluating the associations between the XRCC1 gene polymorphisms and response to platinum-based chemotherapy in lung cancer14,15,16,79,80,81. These meta-analyses may have not enrolled all of available studies which are published before or after their meta-analysis, thus may have biased the conclusions. Therefore, in the present study, we conducted an update system review and meta-analysis to combine all available publications on pharmacogenetic studies and derived more precise and comprehensive assessment on the reliable associations of the commonest XRCC1 genetic polymorphisms on the efficacy and clinical outcomes of advanced lung cancer patients treated with platinum-based chemotherapy.

Results

General characteristics of eligible studies

Overall, 1618 publications were retrieved by systematic literature search using different combinations of key terms. According to our inclusion and exclusion criteria, 53 eligible studies were included for evaluating the data pool for our systematic review (Fig. 1), altogether 7433 advanced lung cancer patients. The general characteristics of the included publications are shown in Table 1. Only two XRCC1 genetic polymorphisms, rs1799782 (Arg194Trp) and rs25487 (Arg399Gln) were finally enrolled in this meta-analyses. Other XRCC1 genetic polymorphisms were not included because of insufficient numbers of relevant publications on each polymorphism, such as three studies with rs25489 (Arg280His), two studies with rs3213245 (T-77C), one study with rs3213239 (insertion variation GGCC). 24 of the included studies were reported in the Arg194Trp genetic polymorphism and 45 were reported in the Arg399Gln genetic polymorphism. Seven of the included studies were conducted on Caucasian patients and 46 were conducted on Asian patients. 22 studies were published in English and 31 studies were published in Chinese. There were 46 studies reported the objective response rate (ORR), 22 studies reported the OS and hazard ratios (HRs), 20 studies reported the median survival time (MST), four studies reported the median time to progression (TTP) and 11 studies reported the median progression-free survival (PFS) (Tables 1 and 2). The sample size ranged from 45 to 460 advanced lung cancer patients. The quality score (QS) of the included studies is summarized in Tables 1 and ranged from 6 to 23. Among these studies, 40 studies were high quality, 13 were low quality.

Table 1 The general characteristics of eligible studies in the meta-analysis.
Table 2 Association between the XRCC1 Arg194Trp and Arg399Gln polymorphisms and median survival time, median time to progression, and median progression-free survival of platinum-based chemotherapy in advanced lung cancer patients.
Figure 1
figure 1

The flow chart of literatures search and selection of included studies.

Objective response rate of XRCC1 genetic polymorphisms

23 eligible studies including 3662 advanced lung cancer patients were qualified for the final analysis for the XRCC1 Arg194Trp polymorphism. The results from the meta-analysis suggested that there were statistically significant associations between the XRCC1 Arg194Trp polymorphism and ORR under all the genetic models (TrpTrp versus (vs.) ArgArg: odds ratio (OR) = 2.07, 95% confidence intervals (CIs), 1.67–2.58, P < 0.001; TrpArg vs. ArgArg: OR = 2.11, 95% CI, 1.68–2.65, P < 0.001; TrpTrp+TrpArg vs. ArgArg: OR = 2.02, 95% CI, 1.66–2.45, P < 0.001, Fig. 2A; TrpTrp vs. TrpArg+ArgArg: OR = 1.56, 95% CI, 1.27–1.91, P < 0.001; Trp vs. Arg: OR = 1.69, 95% CI,1.46–1.95, P < 0.001, Table 3). Subgroup analyses by QS suggested that, for the high quality group, the Arg194Trp polymorphism was significantly associated with ORR of advanced lung cancer patients under all the genetic models (TrpTrp vs. ArgArg: OR = 2.08, 95% CIs, 1.66–2.63, P < 0.001; TrpArg vs. ArgArg: OR = 1.96, 95% CI, 1.51–2.54, P < 0.001; TrpTrp+TrpArg vs. ArgArg: OR = 1.91, 95% CI, 1.53–2.38, P < 0.001; TrpTrp vs. TrpArg+ArgArg: OR = 1.62, 95% CI, 1.30–2.02, P < 0.001; Trp vs. Arg: OR = 1.67, 95% CI,1.39–1.99, P < 0.001, Table 3). For the low quality studies group, the significant associations between Arg194Trp polymorphism and ORR of advanced lung cancer patients were found in the heterozygote genetic model (TrpArg vs. ArgArg: OR = 2.83, 95% CI, 1.90–4.22, P < 0.001), in the dominant genetic model (TrpTrp+TrpArg vs. ArgArg: OR = 2.63, 95% CI, 1.80–3.84, P < 0.001) and in the allele genetic model (Trp vs. Arg: OR = 1.78, 95% CI,1.36–2.33, P < 0.001, Table 3). The lung cancer patients bearing the favorable 194Trp allele, TrpTrp and TrpArg genotypes were more likely to better response rates to platinum-based chemotherapy compared to those with the unfavorable 194Arg allele and ArgArg genotype (Table 3).

Table 3 Meta-analysis of the association between XRCC1 Arg194Trp polymorphism and platinum-based chemotherapy in objective response rate, overall survival and median progression-free survival for advanced lung cancer patients.
Figure 2
figure 2

Forest plots of clinical outcomes in advanced lung cancer patients treated with platinum-based chemotherapy by the XRCC1 Arg194Trp polymorphism.

(A) Odds ratios (ORs) (and its 95% confidence interval (CI)) of objective response rate (ORR) stratified by study quality levels for TrpTrp+TrpArg vs. ArgArg. (B). Hazard ratios (HRs) (and its 95% CI) of overall survival (OS) for TrpTrp vs. ArgArg. (C). HRs (and its 95% CI) of median progression-free survival (PFS) for TrpTrp vs. ArgArg.

There were 38 eligible studies with a total number of 5360 advanced lung cancer patients, qualified for the final analysis for the XRCC1 Arg399Gln polymorphism. In the dominant genetic model, the GlnGln and GlnArg genotypes of XRCC1 Arg399Gln polymorphism were significantly associated with the unfavorable ORR in advanced lung cancer patients treated with platinum-based chemotherapy compared to those with the favorable 399ArgArg genotype (GlnGln+GlnArg vs. ArgArg: OR = 0.68, 95% CI, 0.54–0.86, P = 0.001, Fig. 3A, Table 4). Subgroup analyses by stratified patient populations indicated the Arg399Gln polymorphism was not significantly associated with ORR of advanced lung cancer patients in Caucasians under all genetic models (Fig. 3A, Table 4). For the Asian group, the Arg399Gln polymorphism was also not significantly associated with ORR in four genetic models (P > 0.05, Table 4), except for dominant genetic model (GlnGln+GlnArg vs. ArgArg: OR = 0.65, 95% CI, 0.50–0.86, P = 0.002, Fig. 3A, Table 4). We also performed subgroup analysis by QS. For the high quality group, significant association between the Arg399Gln polymorphism and ORR of advanced lung cancer patients was only found in the dominant genetic model (GlnGln+GlnArg vs. ArgArg: OR = 0.72, 95% CI, 0.56–0.94, P = 0.017, Table 4). For the low quality studies group, the Arg399Gln polymorphism was statistically significantly associated with ORR of advanced lung cancer patients in the dominant genetic model (GlnGln+GlnArg vs. ArgArg: OR = 0.53, 95% CI, 0.32–0.89, P = 0.017, Table 4), in the homozygote genetic model (GlnGln vs. ArgArg: OR = 0.36, 95% CI, 0.14–0.94, P = 0.037 and in the allele genetic model (Gln vs. Arg: OR = 0.62, 95% CI, 0.43–0.90, P = 0.012, Table 4). Results suggested that the 399Gln allele may be indicative of poorer response rate to platinum-based treatment than the 399Arg allele (Table 4).

Table 4 Meta-analysis of the association between XRCC1 Arg399Gln polymorphism and platinum-based chemotherapy in objective response rate, overall survival and median progression-free survival for advanced lung cancer patients.
Figure 3
figure 3

Forest plots of clinical outcomes in advanced lung cancer patients treated with platinum-based chemotherapy by the XRCC1 Arg399Gln polymorphism.

(A) Odds ratios (ORs) (and its 95% confidence interval (CI)) of objective response rate (ORR) stratified by ethnicity for GlnGln+GlnArg vs. ArgArg. (B) Hazard ratios (HRs) (and its 95% CI) of overall survival (OS) stratified by ethnicity for GlnGln vs. ArgArg. (C) HRs (and its 95% CI) of median progression-free survival (PFS) stratified by ethnicity for GlnGln vs. ArgArg.

Overall survival of XRCC1 genetic polymorphisms

Five studies including 1559 advanced lung cancer patients were eligible for the final analysis of the relationship between the XRCC1 Arg194Trp polymorphism and OS. The results from the meta-analysis indicated that no statistically significant relationships between the XRCC1 Arg194Trp polymorphism and OS (TrpTrp vs. ArgArg: HR = 0.84, 95% CI, 0.64–1.11, P = 0.223, Fig. 2B; TrpArg vs. ArgArg: HR = 0.96, 95% CI, 0.79–1.16, P = 0.653; TrpTrp+TrpArg vs. ArgArg: HR = 0.93, 95% CI, 0.72–1.21, P = 0.590, Table 3). No evidence of heterogeneity with respect to predictive value was observed (Table 3).

A total of 19 studies including 3707 advanced lung cancer patients were included in the final analysis of the relationship between the XRCC1 Arg399Gln polymorphism and OS. Variant genotypes of XRCC1 399 genetic polymorphism were more likely to associate with lower rates of OS and higher risks of death for advanced lung cancer patients (GlnGln vs. ArgArg: HR = 1.14, 95% CI, 0.75–1.75, P = 0.533, Fig. 3B; GlnGln vs. GlnArg: HR = 1.42, 95% CI, 1.01–2.00, P = 0.043, Table 4). The results suggested that the 399Gln allele and/or GlnGln genotype might be more closely associated with shorter survival time and higher risks of death for advanced lung cancer patients than the 399Arg allele and/or ArgArg genotype (Table 4).

Median time to progression of XRCC1 genetic polymorphisms

Four studies including 349 advanced lung cancer patients were finally included in this part of analysis for the XRCC1 Arg399Gln polymorphism. Because the data from these enrolled studies on the association of XRCC1 Arg399Gln polymorphism with median TTP was too insufficient to be conducted meta-analysis, we only summarized the general predictive value of the median TTP for XRCC1 Arg399Gln polymorphism in advanced lung cancer patients. The lung cancer patients bearing the unfavorable 399Gln allele, GlnGln and GlnArg genotypes were more likely to lower TTP to platinum-based chemotherapy compared to those with the favorable 399Arg allele and ArgArg genotype (Table 2).

Median progression-free survival of XRCC1 genetic polymorphisms

Three studies with a total number of 721 advanced lung cancer patients were enrolled in the final analysis of the association between XRCC1 Arg194Trp polymorphism and median PFS. The XRCC1 Arg194Trp polymorphism was not significantly associated with the median PFS of advanced lung cancer patients treated with platinum-based chemotherapy (TrpTrp vs. ArgArg: HR = 1.01, 95% CI, 0.69–1.48, P = 0.948, Fig. 2C; TrpArg vs. ArgArg: HR = 1.06, 95% CI, 0.82–1.36, P = 0.662; TrpTrp+TrpArg vs. ArgArg: HR = 1.05, 95% CI, 0.79–1.38, P = 0.753, Table 3).

There were six studies, altogether 1056 advanced lung cancer patients, finally qualified for the analysis of the association between XRCC1 Arg399Gln polymorphism and median PFS. There was no statistically significant evidence for an influence of the XRCC1 Arg399Gln polymorphism on median PFS of advanced lung cancer patients treated with platinum-based chemotherapy (GlnGln vs. ArgArg: HR = 0.80, 95% CI, 0.58–1.11, P = 0.179, Fig. 3C; GlnArg vs. ArgArg: HR = 0.91, 95% CI, 0.71–1.17, P = 0.468; GlnGln+GlnArg vs. ArgArg: HR, = 0.86, 95% CI, 0.70–1.06, P = 1.38, Table 4). Stratified analyses by ethnicity indicated the Arg399Gln polymorphism was not statistically significantly associated with median PFS of advanced lung cancer patients in Asians and Caucasians (Table 4).

No other significant differences were detected with respect to the associations between predictive assessment in pharmacogenetics of XRCC1 genetic polymorphisms and the clinical outcomes of advanced lung cancer patients treated with platinum-based chemotherapy (Tables 3 and 4).

Publication Bias and sensitivity analysis

The publication bias in the enrolled studies on the association of the XRCC1 Arg194Trp and Arg399Gln polymorphisms with clinical outcomes of advanced lung cancer patients treated with platinum-based chemotherapy was assessed by the Begg’s funnel plot and Egger’s test. As shown in Fig. S1, the shapes of the Begg’s funnel plots under the homozygote comparison model of Arg194Trp polymorphism (TrpTrp vs. ArgArg) shown approximately symmetrical and significant evidence of publication bias was not observed by the Egger’s test. As for the Arg399Gln polymorphism, the shapes of the Begg’s funnel plots under the heterozygote comparison model (GlnArg vs. ArgArg) seemed approximately symmetrical, while the Egger’s test show potentially evidence of publication bias (Fig. S2). Sensitivity analysis showed that changing the effect models had no significant effects on the pooled OR and HR and did not change the final strength of the association between the XRCC1 Arg194Trp and Arg399Gln polymorphisms and sensitivity to platinum-based chemotherapy for advanced lung cancer patients. The integrated effects of the exclusion of low quality studies were also evaluated and the results indicated that excluding of low quality studies did not significantly change the final effect, suggesting that the assessment results of this system are reliable.

Cumulative meta-analyses

As shown in Fig. 4, our cumulative meta-analyses based on year of publication indicated that there was a distinct trend toward a better response rate to platinum-based chemotherapy treatment with advanced lung cancer patients for the XRCC1 Arg194Trp polymorphism (TrpTrp+TrpArg vs. ArgArg, Fig. 4). Between 2004 and 2014, a total of 23 studies were published, with a cumulative effect estimate of OR being 2.02 (95% CI, 1.66–2.45). As for the XRCC1 Arg399Gln polymorphism, a cumulative meta-analysis of total 38 studies was conducted to evaluate the cumulative effect estimate over time. We found an apparent poorer response rate in advanced lung cancer patients treated with platinum-based chemotherapy (GlnGln+GlnArg vs. ArgArg), especially in the recent studies published for the past 4 to 5 years (Fig. 5). In 2004, Wang et al. reported an effect estimate of OR being 0.38 (95% CI, 0.16–0.89). Between 2006 and 2009, a total of 12 studies were published, with a cumulative effect estimate of OR being 0.76 (95% CI, 0.58–0.99). Between 2010 and 2015, a total of 25 more publications were added cumulatively, resulting in an overall effect estimate of OR being 0.68 (95% CI, 0.54–0.86, Fig. 5).

Figure 4
figure 4

Forest plot of cumulative meta-analysis to sort out the time-tendency of clinical outcomes in advanced lung cancer patients treated with platinum-based chemotherapy by the XRCC1 Arg194Trp genetic polymorphism (Odds ratios (ORs) and its 95% confidence interval (CI) of objective response rate (ORR) for TrpTrp+TrpArg vs. ArgArg).

Figure 5
figure 5

Forest plot of cumulative meta-analysis to sort out the time-tendency of clinical outcomes in advanced lung cancer patients treated with platinum-based chemotherapy by the XRCC1 Arg399Gln genetic polymorphism (Odds ratios (ORs) and its 95% confidence interval (CI) of objective response rate (ORR) stratified by ethnicity for GlnGln+GlnArg vs. ArgArg).

Discussion

In this study, we found, by an extensively quantitative and systematic review of all available published studies, that patients carrying XRCC1 194Trp allele were more likely to better response rates compared to those carrying the 194Arg allele and patients with XRCC1 399Gln allele carries (GlnGln+GlnArg) may be indicative of poorer response rates, shorter survival time and higher risks of death than the 399ArgArg genotype. Thus, we suggested that the XRCC1 Arg194Trp and Arg399Gln genetic polymorphisms may be predictive factors for treatment response to advanced lung cancer patients treated with platinum-based chemotherapy.

The platinum-based chemotherapy regimens are the standard first-line therapies for advanced lung cancers and commonly used today. The platinum is an effective chemotherapeutic drug for lung cancer patients. The platinum-based compounds could damage DNA by activating intracellularly to form reactive platinum complexes that covalently bind to DNA to induce intrastrand and interstrand DNA cross-links, as well as DNA-protein cross-links, thereby causing the eventual death of cancer cells82. The cancer cells may be more likely to resist against the platinum-based chemotherapy when its ability to repair DNA damage is enhanced by removing those DNA adducts caused by platinum-based compounds14,83. Previous studies revealed that lung cancer patients carrying a lower DNA repair capacity had an increased OS after the first-line platinum-based chemotherapy11,12. However, the anti-cancer drug therapeutic efficacy of platinum-based chemotherapy varies largely among different individuals. Genetic polymorphisms of drug target genes, genes involving in detoxification pathways and DNA repair pathways may influence the anti-cancer therapeutic efficacy of platinum drugs and reveal platinum sensitivity in patients84,85. These genetic polymorphisms could contribute to the variety in phenotypic drug sensitivity through modifying functions of its related genes14.

The XRCC1 protein, a limiting factor in the BER pathway, is considered to play key roles in DNA damage repair and the association of XRCC1 genetic polymorphisms with sensitivity to platinum-based chemotherapy treatment has attracted much interest and became a research hotpot in individual treatment for lung cancer patients. The XRCC1 genetic polymorphisms have been proved to be associated with an altered DNA repair activity11,12. The XRCC1 transcript abundance levels have shown a significant correlation with DDP chemoresistance in NSCLC cell lines86. Published data have indicated that the XRCC1 genetic polymorphisms might be associated with the clinical outcome of platinum-based chemotherapy treatment in lung cancer patients7,10,34,37,43,45,54,59,60,63,64,69,71,75,76,87. These studies imply that the XRCC1 genetic polymorphisms contribute to the repair of DNA damage induced by the platinum agents. Previous studies have also reported that the XRCC1 genetic polymorphisms could be potential risk factors for the pathogenesis of lung cancer88,89,90,91,92 and acting as promising predictive or prognostic biomakers for patients with lung cancer10,14,93,94. Thus, it is speculated that the functional SNPs in XRCC1 gene might be associated with sensitivity to platinum-based anticancer drugs and have predictive or prognostic values in clinical outcomes for patients with lung cancer. The Arg194Trp and Arg399Gln genetic polymorphisms are the most extensively studied SNPs of XRCC1 gene. These polymorphisms, that lead to the encoded amino acid changes (for Arg194Trp, C→T, Arg→Trp, exon 6; for Arg399Gln, G→A, Arg→Gln, exon 10) which might affect the normal function of XRCC1 protein, might alter the efficiency of DNA repair14,79,95. These polymorphisms are located within the functional domain and could have a significant influence on XRCC1 function. Although, the Arg280His genetic polymorphism is also an amino acid variant and leads to the Arg to His substitution (G→A, exon 6), the codon 280 is located outside the known functional domains of XRCC196. In our study, we also found that few studies have reported the efficacy of clinical outcomes of advanced lung cancer patients with platinum-based chemotherapy treatment for the Arg280His genetic polymorphism. Thus, the study for Arg280His genetic polymorphism was too few to be analyzed in this meta-analysis.

With a pooled dataset of 7433 advanced lung cancer patients treated with platinum-based chemotherapy, we performed a comprehensive and systematic evaluation of clinical outcomes by ORR, OS, TTP and PFS. We are delighted to find that the XRCC1 Arg194Trp and Arg399Gln polymorphisms are significantly associated with clinical outcomes in advanced lung cancer patients treated with platinum-based chemotherapy. Our meta-analysis suggested that there were significant associations between the Arg194Trp polymorphism and ORR under all the genetic models; however, there were significant associations between the Arg399Gln polymorphism and ORR only in the dominant genetic model and in the Asian population. It seemed that the Arg194Trp may be a reliable predictive locus to assess the pharmacogenetics of XRCC1 gene on clinical outcomes of advanced lung cancer patients treated with platinum-based chemotherapy compared with the Arg399Gln based on ORR outcomes. We found that all patients in the enrolled studies examining the Arg194Trp genotypes were Asians, likely because this genetic polymorphism is rare in Caucasians (<6%)16,95. Besides, most of patients in the enrolled studies examining the Arg399Gln genotypes were Asians and patients in only seven studies were Caucasians. Lung cancer is a heterogeneous and complicated disease, therefore ethnic difference may affect anti-cancer therapeutic efficacy of platinum drugs. To clarify this concern, we evaluated the relationship of genetic polymorphisms with the anti-cancer therapeutic efficacy of platinum drugs stratified by different ethnic populations. Our results indicated that most of the separately analyzed results were consistent with the overall populations, while the Arg399Gln genetic polymorphism was only significantly associated with ORR in Asian population in the dominant genetic model, not in Caucasian population (Fig. 3A, Table 4). We also noted different clinical outcomes with respect to the Arg399Gln genetic polymorphism (Table 4). These findings show that the clinical outcomes of platinum-based chemotherapy treatment for advanced lung cancer patients differ between Asian and Caucasian populations. Therefore, the conclusions drawn from this meta-analysis about the ethnicity subgroup should be weighed with caution and the ethnic factor should be considered if individual chemotherapy treatment for lung cancer patients is conducted in the future. The consistent quality of studies or trials is an important influence factor and might vary in meta-analyses of genetic epidemiology studies79,97,98. Wu et al. firstly designed the quality assessment system79. Here, based on their quality assessment system, we added more clinical variables, such as MST and median PFS. The consistent study quality could be evaluated using the criteria established in this meta-analysis. These criteria, including phenotypic, genetic epidemiologic and clinical variables, will help standardize study design and experimentation and might influence future pharmacogenomic studies in the field of lung cancer research. Thus, subgroup analyses were also conducted with respect to QS (Tables 3 and 4). Our results showed that most of the separately analyzed results were consistent with the overall QS. It was noted that the correlation between Arg399Gln genetic polymorphism and clinical outcomes appeared in the low-quality studies, but not in the high-quality studies (Table 4). Results from this study suggested that the predictive role of genetic polymorphisms in clinical outcomes might need to be explored more carefully in future pharmacogenomic studies incorporating more credible criteria in the design and experimentation to obtain more accurate and robust conclusion. We observed that the Arg194Trp and Arg399Gln genetic polymorphisms in XRCC1 gene were not statistically influenced the median PFS in all advanced lung cancer patients with platinum-based chemotherapy treatment (Figs. 2C and 3C, Tables 3 and 4). These results have not shown significance because there are few studies provided enough median PFS data and finally involved into our meta-analysis.

To our knowledge, for the first time, we conducted a sequential year-to-year cumulative meta-analysis for clinical outcomes of platinum-based chemotherapy treatment with advanced lung cancer patients. Provided time span of all the available studies was considerable (from 2004 to 2015), our cumulative meta-analysis was encouraged to sort out the cumulative evidence from time-tendency of clinical outcomes by successively adding published studies to the given results. As for the XRCC1 Arg194Trp polymorphism, sequential cumulative meta-analyses consistently and stably showed equivalent effects of better response rates to platinum-based chemotherapy treatment with advanced lung cancer patients since the several initial studies were pooled, which also showed the stable time-dependent trend (Fig. 4). As for the XRCC1 Arg399Gln polymorphism, sequential cumulative meta-analyses results were calculated for each year from 2004. Our cumulative meta-analysis did not show a significant change in trend of reporting response rate to platinum-based chemotherapy treatment with advanced lung cancer patients between 2004 and 2009, which range from 0.38 to 0.76. Nonetheless, from 2010 onwards, it is clear from the sequential pooled cumulative meta-analyses results from studies that there is consistently and stably showed an apparent trend toward a poorer response rate to platinum-based chemotherapy treatment with advanced lung cancer patients, which range from 0.73 to 0.68 (Fig. 5). The trial sequential analysis of cumulative meta-analyses (performing a new meta-analysis each time the results of a new clinical trial are published) would have made the evidence much clearer earlier and lead to sufficient assessment of clinical outcomes. It may also have prevented many redundant trials, redirected trial design and help planning further clinical trials. Therefore, the conclusions drawn from this cumulative meta-analysis about the response rates to platinum-based chemotherapy treatment with advanced lung cancer patients should be weighed with caution.

Several previous published meta-analyses have been carried out to reveal the potential correlation of XRCC1 gene polymorphisms and platinum-based chemotherapy in lung cancer patients, these results remain conflicting rather than conclusive14,15,16,79,80,81. The different studies which enrolled in the meta-analysis could possibly biase the conclusions. In the current meta-analysis, we systematically summarized all available up-to-date studies on the association of XRCC1 gene polymorphisms with platinum-based chemotherapy treatment for lung cancer through conducting to comprehensive literature search in multiple databases without limiting publication date and language. As a result, our updated meta-analysis enrolled 53 available studies including 7433 advanced lung cancer patients, which were significantly more than the previous published meta-analyses. Therefore, findings from this study could provide more precise and reliable comprehensive assessment than those published meta-analyses on the predictive role of XRCC1 genetic polymorphisms in clinical outcomes of platinum-based chemotherapy treatment for advanced lung cancer patients.

Despite our efforts in performing a comprehensive analysis, several limitations of this meta-analysis should be taken into consideration when interpreting the present results. Firstly, the sample sizes of enrolled studies were from 45 to 460 and the small sample size of studies might not generate solid conclusion in some situation. Some of the findings in subgroups may have been undervalued because of there was only one trial available for analyses. Secondly, the significant heterogeneity between studies was found in our pooled analysis, while it is unlikely to alter our main conclusions because our results reflected the most current state of this issue in clinical practice and studies. Stratified analyses by possible confounding factors that could contribute to the ORR and OS of patients, such as gender, age, ethnicities, cancer pathology types, cancer stage, smoking history, specific anti-cancer drugs, chemotherapy regimens, test methods and toxic effects of various platinum-based therapies or other chemotherapies, could provide additional useful information and be helpful to reduce the heterogeneity. However, few of these studies provided sufficient data by subgroups, thus making such subgroup analyses were impossible to implement in the present study. Furthermore, our analyses mostly used unadjusted estimates, because not all published studies calculate adjusted estimates by possible confounding factors. A more precise analysis with the adjustment estimates could be conducted if more detailed individual data were available. Therefore, in order to make the result more accurate and reliable, future studies are necessary to avoid these pitfalls. Thirdly, among all 53 trials utilized in this meta-analysis, most of studies were conducted in Asians (Chinese and Korean populations), only seven studies in Caucasian populations (Greek, Spanish, Swiss, Italian and American populations). Thus, the findings from this meta-analysis might be applicable only to these ethnic populations, while could limit the generalizability to other patient populations. Fourthly, only published studies in English and Chinese were enrolled in this meta-analysis, published studies in other languages, ongoing studies and unpublished data were not collected, which could cause some biases in our estimates.

Conclusions

In conclusion, despite the limitations, our meta-analysis indicates a predictive role for the genetic polymorphisms of XRCC1 gene in clinical outcomes of platinum-based chemotherapy for advanced lung cancer patients. The XRCC1 Arg194Trp and Arg399Gln polymorphisms are likely to be associated with the ORR, OS and sensitivity to platinum-based chemotherapy for advanced lung cancer patients. The use of XRCC1 gene polymorphisms as predictive factor of clinical outcomes in personalized chemotherapy treatment for lung cancer patients might need to be investigated more carefully in well-designed pharmacogenetics and functional studies with large sample sizes in diverse ethnic populations to ensure a more accurate and robust conclusion in the future.

Methods

Identification of eligible studies

We conducted a systematic literature search using the following search terms “lung cancer or tumour or tumor or neoplasm or carcinoma”, “XRCC1 or X-ray repair cross complementing group 1”, “base excision repair” or “BER”, “SNPs or genetic polymorphisms or variations”, “pharmacogenomics”, “platinum or cisplatin or carboplatin or nedaplatin, lbaplatin, oxaliplatin” and “chemotherapy” from PubMed, MEDINE, Excerpta Medica Database (EMBASE), ISI Web of Science, ScienceDirect, Wiley Online Library, Cochrane Central Register of Controlled Trials, Chinese Biomedical Literature Database (CBM), Wangfang Data and Chinese National Knowledge Infrastructure (CNKI) databases up to March 30, 2015. No restriction of publication date was applied. Reference lists were screened manually to further identify additional eligible studies. We retrieved all potentially studies to identify the most eligible literatures.

Inclusion and exclusion criteria

The following inclusion criteria for the current meta-analysis were as follows: (1). Patients had pathologically or histologically confirmed advanced lung cancer. (2). XRCC1 genetic polymorphisms were genotyped; (3). Treatments had platinum-based chemotherapy; (4). Studies had sufficient data to estimate relative risks for prognostic effects of advanced lung cancer (i.e., ORR, OS, median PFS, MST, TTP, OR and HR with corresponding to 95% CIs); (4). only full-text studies were included. Studies were excluded if any of the following exclusion criteria applies: (1). Duplicated publications; (2). Abstracts, comments, letters and review articles; (3). Not reported any clinical outcomes; (4). No sufficient data were reported. (5). The studies with animals or cell lines were reported; (6). The corresponding authors were not provided the valid data or critical information upon our request.

Data extraction

Two investigators (Z.R. Yuan and J. Li) independently extracted data from each eligible publications. Discrepancies between investigators were resolved by discussion from the third investigator, or the team’s decision. The following information was collected from the included studies: the first author’s name, year of publication, country, ethnicities (categorized as Asian and Caucasian), number of patients, median age (year), clinical stage, evaluation criterion, genotyping methods, outcomes (ORR, OS, PFS, MST, TTP, with corresponding to 95% CI) and the number of responders and non-responders in different genotypes.

Quality assessment

The QS of each eligible literature was also evaluated separately by two investigators (Z.R. Yuan and J. Li) using a predefined scale (Table 5). Based on the previous published studies79,81,97,99,100, the scale for quality assessment and quality scoring criteria were determined with the following eight factors: cancer clinical stage, evaluation criteria, platinum combinations, genotyping methods, OS, median PFS, MST and sample size (Table 5). The total QS ranged from 0 (worst) to 24(best). The final QS were assigned to each included literatures after disagreement discussed and resolved by consensus from reviewers. The literature with QS ≤ 12 (or >12) was considered low (or high) quality.

Table 5 The scale for quality assessment.

Statistical analysis

Five genetic models, including homozygote genetic model (mutational homozygote vs. wild homozygote), heterozygote comparison (heterozygote vs. wild homozygote), dominant genetic model (heterozygote+mutational homozygote vs. wild homozygote), recessive genetic model (mutational homozygote vs. heterozygote+wild homozygote) and allele genetic model (mutational vs. wild) were considered in this meta-analysis. We extracted the ORR, OS, ORs, HRs, MST, PFS and TTP from all enrolled studies. We evaluated the ORs and 95% CIs for the ORR and no response after platinum-based chemotherapy (complete response (CR) + partial response (PR) vs. progressive disease (PD) + stable disease (SD), using the World Health Organization (WHO) or the Response Evaluation Criteria in Solid Tumors (RECIST) criteria). We conducted the PRISMA checklist as the protocol of this meta-analysis and followed its guideline (Fig. S3 and Table S1)101. The strength of association between XRCC1 genetic polymorphisms and response rate of platinum-based chemotherapy for advanced lung cancer patients was estimated by calculating pooled ORs with corresponding to 95% CIs. The significance of the pooled ORs was estimated using the Z-test and P-value < 0.05 was considered statistically significant. HRs and 95% CIs were estimated for OS, MST, median TTP, median PFS, directly from the raw data of enrolled articles, or indirectly from the Kaplan–Meier (KM) curve of enrolled articles using published methods102,103. Subgroup analyses were detected by stratified patient populations, clinical outcomes and QS. The between-study heterogeneity was evaluated by the Cochran’s chi-square-based Q-test104,105 and the I2 index106. P-value < 0.10 and/or I2 index >50% for Q-test indicated the existence of heterogeneity between studies107, so the pooled ORs was calculated by the random-effects model using the DerSimonian and Laird method108. Otherwise, the fixed-effects model (the Mantel-Haenszel method) was used109. The sensitivity was conducted by changing the effect models and excluding low quality studies (defined as QS ≤ 12) to estimate confidence. The effects of publication bias of literatures were examined by the Begg’s funnel plots110 and Egger’s linear regression method111. Cumulative meta-analyses were performed to sort out the time-tendency of clinical outcomes of platinum-based chemotherapy for advanced lung cancer patients. All tests were two-sided and P-values < 0.05 were considered representative of statistically significant level. All analyses were conducted using the STATA software (version 11.0; STATA Corporation, College Station, TX, USA).

Additional Information

How to cite this article: Yuan, Z. et al. Predictive assessment in pharmacogenetics of XRCC1 gene on clinical outcomes of advanced lung cancer patients treated with platinum-based chemotherapy. Sci. Rep. 5, 16482; doi: 10.1038/srep16482 (2015).