Meta-analysis of Excision Repair Cross-complementation Group 1 (ERCC1) Association with Response to Platinum-based Chemotherapy in Ovarian Cancer

Recent indicated Abstract Recent studies suggested that the ovarian cancers with negative excision repair cross-complementation group 1 enzyme (ERCC1) expression have a better response to platinum-based chemotherapy than those with positive ERCC1 expression. The objective of this study was to evaluate whether ERCC1 expression is associated with response to platinum-based chemotherapy in ovarian cancers. MEDLINE, PubMed, Web of Science and CNKI databases were used for searching studies relating to ERCC1 protein expression and response to platinum-based chemotherapy in ovarian cancers. Statistical analysis was based on the method for a fixed effects meta-analysis. Pooled odds ratios (ORs) with 95% confidence intervals for ERCC1 protein expression and response to platinum-based chemotherapy were generated. Publication bias was investigated with Begg’s test. Five studies involving 306 patients with ovarian cancer were included. Compared to patients with positive ERCC1 expression, those with negative ERCC1 expression had a better response to platinum-based chemotherapy. The pooled OR was 5.264 (95% CI: 2.928 – 9.464, P < 0.001) and publication bias was not found ( P = 0.904). The result was similar in both in Asians and Caucasians ( P < 0.001 and P = 0.028, respectively). ERCC1 protein expression status is significantly associated with response to platinum-based chemotherapy in ovarian cancers. chemotherapy in ovarian cancers. However, these studies were of small sample size and have shown inconsistent results. In this study, we performed a meta-analysis of the published literature to assess an association between ERCC1 protein expression and response to platinum-based chemotherapy in patients with ovarian cancer.


Introduction
Ovarian cancer is the major cause of death among the gynecological malignancies (Siegel et al., 2012). One of the standards of care for the optimal treatments of ovarian cancer is a systemic chemotherapy regimen that bases on a platinum-based drug (Lebwohl & Canetta, 1998). However, a large proportion of patients treated with platinum-based chemotherapy fail to benefit from it. Predictive biomarkers are needed for personalized medicine which identify subpopulations of patients who most likely respond to a given therapy and those who not.
The nucleotide excision repair (NER) system plays a key role in mediating resistance or sensitivity to platinum chemotherapeutic agents (Rabik & Dolan, 2007). It repairs platinum-induced DNA damage by altering the helical structure of the DNA molecule and interfering with DNA replication and transcription. ERCC1 is a key gene involved in NER, and ERCC1 status (ERCC1 RNA expression, ERCC1 protein expression, and ERCC1 polymorphisms) are associated with platinum-based therapy efficacy in some kinds of cancers (Vilmar & Sorensen, 2009;Bohanes et al., 2011;Langer, 2012).

Meta-analysis of Excision Repair Cross-complementation Group 1 (ERCC1) Association with Response to Platinumbased Chemotherapy in Ovarian Cancer
Feng-Ying Li 1 , Xiao-Bin Ren 2 , Xin-You Xie 1 , Jun Zhang 1 * chemotherapy in ovarian cancers. However, these studies were of small sample size and have shown inconsistent results. In this study, we performed a meta-analysis of the published literature to assess an association between ERCC1 protein expression and response to platinumbased chemotherapy in patients with ovarian cancer.

Study selection
We carried out a search in MEDLINE, PubMed, Web of Science and CNKI databases covering all papers published till July 2012. No language restrictions were applied. The following keywords were used: "Ovarian Cancer", "Ovarian Neoplasms" and "ERCC1".
Studies included in the meta-analysis had to meet the following criteria: (1) patients had to have pathologically confirmed ovarian cancer under platinumbased chemotherapy regimens; (2) ERCC1 expression were measured using immunohistochemistry (IHC), or western blot (WB); (3) the study should provided data for objective response rate according to ERCC1 expression.
Reviews, abstracts, letters and papers reported only at academic meetings were excluded for this meta-analysis. When overlapping data of the same patient population were included in more than one publication, only the most recent or complete one was included. Study eligibility was determined by Jun Zhang and Fengying Li, any disagreement were resolved by consensus.

Data extraction
All the data were independently abstracted by two investigators, using a standardized data extraction form. The information was collected including first author's name, year of publication, ethnicity, sample size, disease stage, method used to detect expression of ERCC1, treatment regimen, the numbers of the patients with or without response to chemotherapy in ERCC1 positive or negative expression group. Responders included complete response and partial response patients; non-responders included stable disease and progressive disease patients.

Statistical analysis
Pooled odds ratios (Pooled ORs) with 95 % confidence interval (95 % CI) were calculated to assess the strength of association between ERCC1 expression and response to platinum-based chemotherapy. To investigate the heterogeneity between the included studies, I-squared was performed (I-squared > 50% was considered significant heterogeneity). If heterogeneity was found, Pooled OR was calculated using a random-effects model; if not, pooled OR was calculated using a fixed-effect model. To evaluate the publication bias, the Begg's test was carried out. All the statistical analyses were performed by Stata version 11 (Stata Corporation, College Station, TX, USA). P-value < 0.05 was considered as significant.

Eligible studies
In total, 227 potentially relevant studies were identified. Of these, 167 studies were excluded after title review, leaving 60 studies for abstract review. 50 studies were excluded for review or for no relevance to chemotherapy or for meeting abstract. After full-text review, 5 studies were excluded from the remaining 10 studies (1 studies were lack of sufficient data (Lin et al., 2008) ; 1 study was letter (Stadlmann et al., 2008) ; 3 studies were found to overlap other study population (Wang & Qu, 2006;Steffensen et al., 2008) ). Finally, 5 studies were included in our meta-analysis Steffensen et al., 2009;Tang et al., 2009;Bosmuller et al., 2011;Xie et al., 2011). The flow chart of study selection was shown in Figure 1. Characteristics of the 5 eligible studies are presented in Table 1. A total of 306 participants with ovarian cancer were included in these studies. No publication bias was detected based on the Begg's test (P = 0.904).
Subgroup analysis was conducted based on ethnicity. The patients in three studies were Asian, and those in the other two studies were Caucasian. The correlation between negative expression of ERCC1 and better response to platinum-based chemotherapy remained significant, both in Asian (OR = 5.477, 95% CI: 2.847-10.536, P < 0.001, Figure 2) and in Caucasian (OR = 4.464, 95% CI: 1.178-16.910, P = 0.028, Figure 2) subgroups, and no heterogeneity was found for both of them (I-squared < 0.1%, P = 0.638 and I-squared < 0.1%, P = 0.823, respectively. Figure 2).

Discussion
In the current meta-analysis, we evaluated the association between ERCC1 expression and response to platinum-based chemotherapy in patients with ovarian cancer. We included 5 studies involving 306 patients treated with platinum-based chemotherapy. The results indicated that the patients with negative ERCC1 expression had a better response to platinum-based chemotherapy than those with positive ERCC1 expression. Subgroup analyses by ethnicity yielded the same results.
In 1999, Köberle et al. found that ERCC1 protein expression levels correlated with cisplatin resistance in the well-defined 833K and GCT27 human testi tumour cell lines by immunoblotting, their study indicated high ERCC1 expression were associated with an reduced sensitivity to cisplatin (Koberle et al., 1999). Five years later, another study by the same group indicated that six of thirty five human cancer cell lines were significantly lower mean levels of ERCC1 proteins compared to others and suggested low levels of ERCC1 protein may explain their extreme chemosensitivity (Welsh et al., 2004). Li et al. (2000). performed a vitro trial suggested ERCC1 protein may be a marker to monitor the repair of platinum-DNA damage in tumor cells and inhibiting ERCC1 expression may increase cellular sensitivity to cisplatin. In fact, much clinical studies also showed that negative/low expression of ERCC1 was associated with good response to chemotherapy. Among these clinical studies, a great deal of them have been conducted and focused on ERCC1 expression and response to chemotherapy in lung cancer patients (Ren et al., 2010;Wang et al., 2010;Hayes et al., 2011;Joerger et al., 2011;Kawashima et al., 2011;Su et al., 2011;Zhang et al., 2011;Sereno et al., 2012;Simon et al., 2012;Yu et al., 2012;Zhang et al., 2012). In 2011, Hubner et al. performed a meta-analysis to investigate the correlation mentioned above and concluded that high ERCC1 may adversely influence response in platinumtreated NSCLC patients (Hubner et al., 2011). The similar mete-analysis was performed by Chen and the same conclusion was gained (Chen et al., 2010). In addition, many papers evaluated the relation between ERCC1 status and response to chemotherapy have been reported in various human cancers including Colorectal Cancer, bladder cancer, cervical cancer, Head and Neck Cancer, and most of studies showed ERCC1 expression was a predictor of response to chemotherapy (Handra-Luca et al., 2007;Braun et al., 2008;Kim et al., 2009;Ishibashi et al., 2010;Kim et al., 2010;Hayes et al., 2011;Kawashima et al., 2011;Park et al., 2011;Sun et al., 2012). In 2008, Lin et al. performed a cohort study to evaluate the difference of ERCC1 protein expression levels between 36 sensitive cases and 27 resistant cases of serous ovarian cancer receiving neoadjuvant chemotherapy and demonstrated that ERCC1 protein expression levels in resistance group of serous ovarian cancer were significantly higher than those of sensitivity group, and Lin et al's study was in accordance with our meta-analysis (Lin et al., 2008).
However, our study has several limitations. First, the number of patients included in study was limited. The pooled OR was gained from 5 studies with a small size of 306 patients, only. Because of limited studies, we did not perform subgroup analysis by regimen. Second, many factors that may influence the response to chemotherapy were not analyzed in this study, including age, pathologic classification, tumor grade and surgical resection. Third, this is a meta-analysis based on the data from the published literature, and individual patient data were not available.
In conclusion, our analysis shows that negative expression of ERCC1 was significantly associated with good response to platinum-based chemotherapy in patients with ovarian cancer, and ERCC1 status could be a potential biomarker to predict response to platinum-based chemotherapy of ovarian cancers. However, large scale and well designed studies are needed to investigate the factors that may influence the response to platinum-based chemotherapy.