Docetaxel, Cisplatin and Fluorouracil (DCF) Regimen Compared with Non-Taxane-Containing Palliative Chemotherapy for Gastric Carcinoma: A Systematic Review and Meta-Analysis

Background Gastric carcinoma (GC) is one of the highest cancer-mortality diseases with a high incidence rate in Asia. For surgically unfit but medically fit patients, palliative chemotherapy is the main treatment. The chemotherapy regimen of docetaxel, cisplatin and 5-fluorouracil (DCF) has been used to treat the advanced stage or metastatic GC. It is necessary to compare effectiveness and toxicities of DCF regimen with non-taxane-containing palliative chemotherapy for GC. Methods PubMed, EmBase, Cochrane Central Register of Controlled Trials and China National Knowledge Infrastructure databases were searched to select relative randomized controlled trials (RCTs) comparing DCF to non-taxane-containing chemotherapy for patients with palliatively resected, unresectable, recurrent or metastatic GC. Primary outcome measures were 1-year and 2-year overall survival (OS) rates. Secondary outcome measures were median survival time (MST), median time to progression (TTP), response rate and toxicities. Results Twelve RCTs were eligible and 1089 patients were analyzed totally (549 in DCF and 540 in control). DCF regimen increased partial response rate (38.8% vs 27.9%, p = 0.0003) and reduced progressive disease rate (18.9% vs 33.3%, p = 0.0005) compared to control regimen. Significant improvement of 2-year OS rate was found in DCF regimen (RR = 2.03, p = 0.006), but not of 1-year OS rate (RR = 1.22, p = 0.08). MST was significantly prolonged by DCF regimen (p = 0.039), but not median TTP (p = 0.054). Both 1-year OS rate and median TTP had a trend of prolongation by DCF regimen. Chemotherapy-related mortality was comparable (RR = 1.23, p = 0.49) in both regimens. In grade I-IV toxicities, DCF regimen showed a major raise of febrile neutropenia (RR = 2.33, p<0.0001) and minor raises of leucopenia (RR = 1.25, p<0.00001), neutropenia (RR = 1.19, p<0.00001), and diarrhea (RR = 1.59, p<0.00001), while in other toxicities there were no significant differences. Conclusion DCF regimen has better response than non-taxane containing regimen and could potentially improve the survival outcomes. The chemotherapy-related toxicity of DCF regimen is acceptable to some extent.


Introduction
Gastric carcinoma is one of the highest cancer-mortality diseases [1][2][3] with a high incidence rate in Asia [4]. A lot of patients are diagnosed at advanced even end stage carcinoma, indicating poor outcomes [5]. For resectable diseases, surgery is considered as the mainstream treatment [6][7][8][9]. Adjuvant or neoadjuvant chemotherapy has also been proven to benefit the survival rate in some studies and meta-analyses [10][11][12][13][14]. For surgically unfit but medically fit patients, palliative chemotherapy is the main treatment [6,12,[15][16][17][18]. The regimens of cisplatin and 5-fluorouracil (CF) or epirubicin, cisplatin and 5-fluorouracil (ECF) have been used widely [19], and were often considered as the reference regimens for advanced gastric cancer [20]. Among new generation chemotherapy regimens, docetaxel, which is a semisynthetic taxane, promoting the assembly and stabilization of microtubules to inhibit the depolymerization [21], has been used more and more extensively with more potent effects [18,[22][23][24][25]. The chemotherapy based on docetaxel may be effective [26], because docetaxel lacks cross-resistance with other anti-tumor drugs [23]. The chemotherapy regimen of docetaxel, cisplatin and 5-fluorouracil (DCF) has been used to treat the advanced stage or metastatic gastric carcinoma with encouraging survival outcomes [27][28][29][30][31][32][33] and better quality of life [34,35] in several studies. However, it was reported in some researches [36][37][38] that more toxicity, such as hematotoxicity, happened in DCF than in other regimens. Therefore, evaluation of benefits against the chemotherapy-related toxicities was needed. Present systematic review and meta-analysis were done to evaluate the survival outcomes and toxicities of DCF for palliatively resected, unresectable, recurrent or metastatic gastric carcinoma, compared with those of non-taxane-containing regimens.

Methods
No protocol was registered.

Search Strategy
We searched the electronic databases of PubMed

Inclusion and Exclusion Criteria
Only randomized controlled trials (RCTs) were eligible for inclusion. Included patients were diagnosed with palliatively resected, unresectable, recurrent or metastatic gastric carcinoma. Both, patients with previous surgery and without, were acceptable. DCF palliative chemotherapy could be administrated as the firstline regimen. If the control arm was blank or contaminated with taxane, the trials were excluded. Response, survival outcomes or toxicities were mandatory to be reported.

Selection, Assessment and Data Extraction
Two independent reviewers (Chen XL, Yang C) read the title and abstract of every searched citation to select eligible studies for further assessment. Full text of potentially eligible citation was retrieved and determined for inclusion. Jadad scale described by Jadad, et al was used to assess the quality of RCTs [39]. Data was extracted independently by two reviewers mentioned above. Primary outcome measures included 1) 1-year and 2-year overall survival (OS) rates. Secondary outcome measures were 2) median survival time (MST), 3) time to progression (TTP), 4) response rate (WHO Criteria) [40] containing complete response (CR), partial response (PR), stable disease (SD), progressive disease (PD) and overall response rate (ORR), 5) toxicities (grade I-IV) and 6) chemotherapy-related death. ORR meant the combination of CR and PR. Study sample and regimen details were also extracted. Any disagreements in studies assessment and data collection were discussed and resolved by a third party (Hu JK, Chen XZ) as the referees.

Statistical Analysis
The statistical analysis was performed by Reviewer Manager (RevMan) software, version 5.0 offered by The Cochrane Collaboration. The Mantel-Haenszel (M-H) test was used for comparison of dichotomous data and risk ratio (RR) or risk difference (RD) estimate. The 95% confidence interval (CI) of RR or RD was also calculated. A two-sided P value less than 0.05 was considered as a significant difference. Between-trials heterogeneity was evaluated by the Chi-square test; P value less than 0.1 was considered as significant heterogeneity [41]. Providing heterogeneity existed, random effect model was used for meta-analysis; otherwise fixed effect model was used.

Response
Response rate was based on WHO Criteria [40]. The overall response rate (ORR) combined of CR and PR, was 44.4% (244/ 549) vs 30.6% (165/540) in DCF and non-taxane-containing regimens, respectively. Its meta-analysis showed significantly better ORR of DCF regimen (RR = 1.45, 95% CI 1.24-1.69, p,0.00001) ( Figure 2). The subgroup meta-analysis results of CR, PR, SD and PD were listed in Table 2. CR and SD rates were not significantly different between two groups (
MST was reported in 9 RCTs and it ranged from 9.0 to 14.6 months in DCF and 5.0 to 12.0 months in non-taxanecontaining regimens (Table 3). Six RCTs analyzed median TTP which ranged from 4.6 to 6.8 months in DCF and 2.6 to 5.5 months in non-taxane-containing regimens (Table 3). Comparison of MST and median TTP between two groups by one-way ANOVA test (SPSS 13.0) demonstrated significantly prolonged MST in DCF regimen (p = 0.039) ( Figure 4). Median TTP showed a trend of prolongation in DCF regimen without reaching statistically significant difference (p = 0.054) ( Figure 5).
The meta-analysis results of these toxicities were listed in Table 3  In grade I-IV toxicities, DCF regimen showed major raise of febrile neutropenia and minor raise of leucopenia, neutropenia, and diarrhea. Other toxicities such as thrombocytopenia, anemia, nausea/vomiting, stomatitis, anorexia, constipation, alopecia, neurological damage and infection appeared with no significant differences in both arms. Some toxicities including febrile neutropenia, liver damage, anaphylaxis and fluid retention were rare.

Discussion
Although our meta-analysis did not show the significant difference of 1-year OS rate between DCF chemotherapy and non-taxane-containing regimens, we found that DCF arm was relatively better than control arms in cumulative 1-year OS rate, MST, median TTP and ORR. Only one study, which had a large sample size (445 patients), reported their 2-year OS rate (DCF 18% vs CF 9%) [43]. Many studies preferred to show the MST or OS time, TTP and especially ORR which represented the short survival outcomes. There was an interesting phenomenon in one study, in which DCF arm had shorter median TTP but longer MST than ECF arm [38]. According to our results, the difference  of ORR was significant between two arms, with DCF having more benefits than control group in short time.
On the toxicity analysis, we found that DCF showed worse hematological toxicity, diarrhea, and fatigue. Many studies reported that these toxicities in both group could be accepted or controlled by granulocyte colony stimulating factors (G-CSF), antiemetic, vitamin B6 and drug discontinuance [38,46,49,[51][52][53]. Further, the total mortality of treatment, mainly caused by infection, was 6.6% in DCF arm and 5.5% in control arms, and showed no significant difference. Prophylactic antibiotics might be necessary for patients with severe leucopenia. Quality of life (QOL) after chemotherapy was of less focus than survival outcomes. Four RCTs demonstrated that DCF group did not have lower QOL. In fact, it even had obvious improvements in global QOL and Karnofsky performance status as well as prolonged the time to worsening of global health [35,38,43,47]. Although causing more toxicities, DCF was a relatively safe and acceptable chemotherapy.
In every chemotherapy cycle for three or four weeks, the total dosage of docetaxel was from 105 mg/m 2 to 300 mg/m 2 , cisplatin from 90 mg/m 2 to 360 mg/m 2 , and 5-FU from 3000 mg/m 2 to 15000 mg/m 2 in these studies. The single dosage of docetaxel was noticeably lower in four studies from China (25 mg/m 2 to 40 mg/ m 2 ) [42,44,47,49], that was about one half of the dosage of three foreign studies (60 mg/m 2 to 85 mg/m 2 ) and five other Chinese studies (60 mg/m 2 to 75 mg/m 2 ) [35,38,43,45,46,48,50,51] in which single dosage was similar to that recommended in National Comprehensive Cancer Network (NCCN) clinical practice guideline in oncology [54]. Similar phenomenon happened in cisplatin. Studies in China showed a noticeably lower single dosage of cisplatin (6 mg/m 2 to 35 mg/m 2 ) [35,44,[46][47][48][49]51] than foreign RCTs [35,38,43] (60 mg/m 2 to 75 mg/m 2 ) [45,50] and NCCN recommendation [54]. From all these, we could see that there was a big difference in approach to dosage of DCF. The reason for this, we believe, was the individual distinctions of people in China and Western countries. Moreover, among Chinese RCTs, there was also a difference in the dosage of docetaxel and cisplatin. The question about which plan has better balance of survival outcomes and toxicities, a bigger single dosage with lower frequencies or a  A previous systematic review showed that there was significant difference in ORR but no significant difference in overall survival, when docetaxel was compared with non-taxane-containing regimens, which is in accord to our study [55]. However, in former review non-docetaxel-containing regimens might include both single and combined chemotherapy regiments. If docetaxel alone is compared with combined regimens, it might show relatively inferior efficacy of docetaxel. However, in our review,  DCF was compared with other combined chemotherapy in all RCTs, which in our opinion seemed more reasonable.
In retrieved literatures of RCTs on DCF, we found that there were no studies from Japan. On one hand, it can be explained by the fact that we only retrieved the English or Chinese language articles. On the other hand, many studies [56][57][58][59][60][61] in Japan talked about the new generation chemotherapy drug S-1 alone or S-1 combined with other drugs, such as docetaxel, and reported promising effects in treating gastric carcinoma. However, they did not compare S-1 with DCF.
Further, we could see that DCF as palliative chemotherapy was reported in all RCTs. At present, adjuvant single or combined chemotherapy following radical surgeries included 5-FU, cisplatin, FOLFOX, capecitabine and S-1. Docetaxel was not the routine first line chemotherapy regimen. From NCCN introduction, DCF was recommended as first line chemotherapy for metastatic or locally advanced cancer [54]. It might raise the question whether or not DCF regimen could be used in postoperative adjuvant chemotherapy and have better survival outcomes than traditional adjuvant chemotherapy.  In our review, 9 out of 12 RCTs came from China, in which the population was small and follow-up time was short. Their aim was to evaluate the short term outcomes, ORR and toxicities. Only three RCTs including a foreign one reported the survival rate. This might influence the analysis in our review. More RCTs with longer follow-up time are needed. Conclusion DCF regimen had better response than non-taxane containing regimen and could potentially improve the survival outcomes. The chemotherapy-related toxicity of DCF regimen is also acceptable to some extent. At the same time, more high-quality RCTs are needed to decide on the effectiveness of DCF regimen.