Chemotherapy for advanced gastric cancer
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
To evaluate, in terms of efficacy and toxicity, the effect of chemotherapy in patients with advanced adenocarcinoma of the stomach.
Comparisons will be as follows:
-
1.) Best supportive care versus chemotherapy plus best supportive care
-
2.) Single agent versus combination chemotherapy
-
3.) Different combination chemotherapy regimens,
classified into the following categories, against each other
a) Regimens containing 5‐FU (results for protocols in which 5‐FU was given as bolus will be separated from those in which 5‐FU was given as a continuous infusion)
b) Regimens not containing 5‐FU
Both categories will be further subdivided into whether they contain a platinum analogue (usually cisplatin) or not, and whether they contain anthracyclines or not. As far as randomised studies with newer substances (e.g. taxanes and oral antimetabolites) are or become available, the non‐5‐FU‐containing regimens will be further classified. Studies with a documented second‐line therapy will be considered separately.
Background
EPIDEMIOLOGY: Adenocarcinoma of the stomach has been a main cause of cancer death through most of the twentieth century. Stomach cancer incidence rates show substantial variation internationally, with highest rates in Japan and eastern Asia (Miller 1996). In the United States, cancer of the distal half of the stomach has decreased since the 1930s. This decline in the incidence of gastric cancer has occurred in the endemic or intestinal form of the disease, which seems to be predisposed by a combination of achlorhydria, super infection with Helicobacter pylori, chronic gastritis and intestinal metaplasia (Boeing 1991; Macdonald 2001).
However, in the last two decades, the incidence of cancer of the cardia and gastroesophageal junction has been rapidly rising, with a dramatic increase of this cancer especially in white males under 40 years of age. A proportion of these cases seems to be associated with Barrett´s epithelium (intestinal metaplasia of the distal esophagus), developing from chronic esophageal reflux disease (Macdonald 1992; Wu‐Williams 1990). Although it is difficult to determine whether these cancers are gastroesophageal junction stomach tumors or distal esophageal malignancies, in clinical trials they are usually treated in the same manner.
PROGNOSIS AND MANAGEMENT OPTIONS: Complete resection is the only potentially curative therapy for gastric cancer. Stage I‐ IV M0‐tumors are principally resectable (Macdonald 2001). But although surgery carries a high cure rate for stage IA and IB cancers, the results for stage IIIA and IIIB cancers are poor. Many patients with advanced disease, especially stage IIIA/B, are technically inoperable. Unfortunately, even after a "curative" gastrectomy, relapse rates in prospective studies are in the range of 40 ‐ 60 % (Macdonald 2001a); Bonenkamp 1999). In unresectable stage III patients, neo adjuvant chemotherapy and/or radiation therapy are under clinical evaluation. In the western world, most patients are diagnosed at an advanced stage when the tumor is inoperable. Patients with inoperable, recurrent or metastatic tumors have a poor prognosis with a median survival time of 3 ‐5 months. Several small randomised trials have provided evidence that systemic chemotherapy improves survival in these patients (Pyrhönen; Scheithauer; Glimelius), but benefit has to be weighed up against treatment‐related toxicities.
SYSTEMIC CHEMOTHERAPY: 5‐FU is not only the most important and extensively studied single agent in this disease, but it is part of most combination chemotherapy regimens as well. Its single‐agent response rate is about 20 % (Coccioni). Differences in effect and toxicity profile are the reasons for its application as continuous infusion in many of the modern protocols. Other single agents with relevant activities are Cisplatin (Leichman 1991), Anthracyclines (Preusser 1988), Etoposide (Macdonald 1992 a)) and Mitomycin C (Schnall 1993), which form the basis for most regimens that have been applied in the last years. Furthermore, oral anti metabolites, taxanes and camptothecans have shown interesting single‐agent activity in phase II‐studies (Karpeh 2001). Their clinical value is currently being evaluated.
Independent of the strength of evidence for combination versus single‐agent therapy, the use of combination chemotherapy regimens has become internationally accepted as palliative treatment in gastric cancer. But although a large number of different chemotherapy regimens has been tested in randomised studies, no single approach can be considered standard of care (Wils 1998; Vanhoefer; Kulke 2000; Karpeh 2001) and uncertainty remains regarding the choice of the regimen.
Objectives
To evaluate, in terms of efficacy and toxicity, the effect of chemotherapy in patients with advanced adenocarcinoma of the stomach.
Comparisons will be as follows:
-
1.) Best supportive care versus chemotherapy plus best supportive care
-
2.) Single agent versus combination chemotherapy
-
3.) Different combination chemotherapy regimens,
classified into the following categories, against each other
a) Regimens containing 5‐FU (results for protocols in which 5‐FU was given as bolus will be separated from those in which 5‐FU was given as a continuous infusion)
b) Regimens not containing 5‐FU
Both categories will be further subdivided into whether they contain a platinum analogue (usually cisplatin) or not, and whether they contain anthracyclines or not. As far as randomised studies with newer substances (e.g. taxanes and oral antimetabolites) are or become available, the non‐5‐FU‐containing regimens will be further classified. Studies with a documented second‐line therapy will be considered separately.
Methods
Criteria for considering studies for this review
Types of studies
-
INCLUSION CRITERIA: Randomised controlled trials with or without blinding. Abstracts or unpublished data will be included, if sufficient information on study design, characteristics of participants, interventions and outcomes is available and if full information as well as final results can be confirmed by contact to the first author
-
EXCLUSION CRITERIA: Cross‐over studies will be excluded in order to assess the overall treatment effect on survival. Quasi‐randomised studies, e.g. treatment allocation alternate or by date of birth, will be excluded as we consider this study design to be of insufficient quality.
Types of participants
-
Patients with histologically confirmed, advanced (T3‐T4NxM0 if technically inoperable; all TxNxM1), recurrent or metastatic adenocarcinoma of the stomach or gastroesophageal junction without any prior chemo‐ or radiotherapy. Studies which include patients with adenocarcinoma of the distal esophagus will also be included. Most studies include patients with locally advanced, relapsed and/or metastasised tumors, with the greater number of patients having already metastasised disease. However, in some studies, only patients with locally advanced cancer of the stomach are included in order to assess secondary resectability. These studies will not be considered in this review. The proportion of patients with locally advanced versus metastasised tumors will be given for each study.
Types of interventions
-
Systemic intravenous chemotherapy and/or best supportive care. Chemotherapy encompasses all cytotoxic or anti neoplastic drug treatment but excluding hormonal or biological therapies. Single agent as well as combination chemotherapy studies in all doses and schedules will be included. Combined radio chemotherapy will not be considered.
Types of outcome measures
-
Primary outcome measure: Overall survival on intention to treat analysis. Median, 1‐, 2‐, and 3‐year as well as 5‐year survival in patients with locally advanced, secondary resectable tumors will be assessed.
-
Secondary outcome measures:
a) Tumor response
b) Time to progression
c) Secondary resectability in patients with locally advanced gastric cancer
d) Toxicity, classified according to WHO‐ or National‐Cancer‐Institute Common‐Toxicity‐Criteria (NCI‐CTC)
Quality of life is difficult to measure and has been assessed with various instruments. As soon as a sufficient number of studies with adequate quality‐of‐life assessment are available, quality of life will be reviewed as well.
Search methods for identification of studies
Searches will be conducted to identify published and unpublished randomised controlled trials:
1) Electronic searches
The Cochrane Controlled Trials Register, Issue 2, 2001, Medline (1966 to 06/2002), EMBASE (1974 to 06/2002), Cancer Lit. (1975 to 06/2002)
Databases of ongoing trials:
http://www.controlled‐trials.com
http://www.clinicaltrials.nci.nih.gov
http://www.eortc.be
http://www.update‐software.com/National/nrr‐frame.html
http://www.CenterWatch.com/
2) Hand searching
Reference lists from trials selected by electronic searching will be hand searched to identify further relevant trials. Published abstracts from conference proceedings (1966 to 06/2002) from the United European Gastroenterology Week (published in Gut) and Digestive Disease Week (published in Gastroenterology), the European Society for Medical Oncology (published in the Annals of Oncology), the European Council of Clinical Oncology, published in the European Journal of Cancer, as well as the American Society for Clinical Oncology, will be hand searched.
The search strategy for the review will be constructed by using a combination of subject headings and text words relating to chemotherapy in gastric cancer.
In addition, members of the Cochrane UGPD Group as well as experts in the field and manufacturers of relevant drugs will be contacted and asked to provide details of outstanding clinical trials and any relevant unpublished material.
Data collection and analysis
TRIAL SELECTION: In order to elect studies for further assessment, two independent reviewers will scan the title, abstract section and keywords of every record retrieved. Full articles will be taken into account for further assessment if the information given suggests that the study:
-
Includes patients with histologically confirmed, inoperable adenocarcinoma of the stomach or gastroesophageal junction
-
Compares either
1.) Best supportive care versus chemotherapy plus best supportive care
2.) Single agent versus combination chemotherapy
3.) Different combination chemotherapy regimens
-
Uses random allocation to the comparison groups
If there is any doubt regarding these criteria from the information given in the title and abstract, the full article will be retrieved for clarification. If differences in opinion exist, they will be resolved by discussion. If no clarification is provided, the review group editorial base will be consulted.
QUALITY ASSESSMENT OF TRIALS: The quality of the eligible studies will be assessed unblinded by two reviewers independently, with disagreements resolved by a third reviewer until consensus is obtained. Data will be extracted using an assessment form designed for the topic of this review (sources used: Schulz 1994; Jadad 1996; Cochrane Reviewers´ Handbook)
-
The following criteria will be considered:
1. Was the allocation truly random ?
2. Was the treatment allocation concealed ?
3. Were the groups similar at baseline regarding the most important prognostic factors ?
4. Were the number of withdrawals, dropouts and losses to follow‐up in each group completely described ? (The drop‐out‐rate will be recorded)
5. Was the analysis done by intention‐to‐treat ?
6. Were type and schedule of the follow up similar in the comparison group ?
7. Have CT/NMR scans been reviewed by an independent radiologist ?
Answers to these questions will be categorized as
‐ yes (a.)
‐ no (b.)
‐ unclear (c.)
Based on these criteria, the studies will be broadly subdivided into the following three categories of quality
A ‐ all of the criteria met: low risk of bias
B ‐ all criteria of question 1., 2., and 3. met; one or more of criteria 4 ‐ 7 not met
C ‐ question one met, one or more of questions 2 ‐ 7 not met
D ‐ Unclear: Insufficient information given. In these cases the first author of the study will be contacted.
This classification will be used as the basis of a sensitivity analysis.
Definition of baseline comparability: Tumor stage (advanced versus metastatic disease) and activity index (ECOG 0‐1 versus 2‐3), as well as the number of organs involved in metastatic disease (one versus more than one) are considered as the most important prognostic factors. A difference of more than 15 % between study arms is considered as an important difference. For age as a further important factor baseline differences of 5 years in median resp. mean are considered as important.
Definition of intention‐to‐treat‐analysis: As randomized analysis, with the analysis restricted to patients which received at least one cycle of chemotherapy and for which survival data is available.
DATA EXTRACTION: Details of study population, interventions and outcomes will be extracted independently by two reviewers using a standardized data extraction form. This form will include at least the following items:
-
General information: title, authors, source, contact address, country, published/unpublished, language and year of publication, sponsoring of trial
-
Trial characteristics including design, duration/follow‐up and quality assessment criteria as specified above
-
Patients: in‐ and exclusion criteria, sample size, baseline characteristics, similarity of groups at baseline, withdrawals and losses to follow up
-
Interventions: dose, route, timing of chemotherapy and comparison intervention
-
Outcomes: Hazard Ratios and their 95% Confidence Intervals, log rank chi square, log rank p‐values, number of events, number of patients per group, median‐, 1‐, 2‐, and 3‐year survival rates, 5‐year survival rates in patients with locally advanced, secondary resectable tumors
A template data extraction form will be developed and tested in a pilot study. Data extraction and data entry will be performed independently in duplicate by two evaluators. Differences in data extraction will be resolved by consensus with a third reviewer, referring back to the original article. If data are missing in a published report, the reviewers will contact the first author.
DATA ANALYSIS: Hazard ratios (HR) and 95 % confidence intervals (CI) as relevant effect measures will be estimated directly or indirectly from the given data (Altman 2001).
Two design situations have to be considered:
a.) Direct comparisons of corresponding therapy arms in relevant trials (comparison 1: studies comparing best supportive care versus chemotherapy plus best supportive care and comparison 2: single agent versus combination chemotherapy):
For each individual trial HR´s and their variances have to be extracted. If the figures are not given directly, methods of indirect determination have to be used. HR´s can be estimated (under some assumptions) from log rank chi‐square, from log rank p‐values, from observed to expected event ratios, from ratios of median survival times resp. time point survival rates (Machin 1997; Parmar 1998).
b.) Indirect comparisons of corresponding therapy arms in relevant trials (comparison 3: studies comparing different categories of combination chemotherapy regimens):
As a standard therapy does not exist, comparing different chemotherapy regimens is more difficult. Nevertheless, because combination chemotherapy has become the internationally accepted standard, it is the clinically most relevant question to find out if some regimens are superior to others.
Currently used regimens can be classified by their most important components, e.g. 5‐FU, anthracyclines and cisplatin. As studies with head‐to‐head comparisons of chemotherapy regimens from all these different categories do not (or very rarely) exist, indirect comparisons of corresponding therapy arms in relevant trials have to be made, and corresponding statistical measures (median, 1 and 2‐year survival rates) from single arms will be condensed from various relevant trials for each category. Afterwards the condensed rates or medians for one category of chemotherapy regimens will be used for further comparisons with other categories of chemotherapy regimens to estimate a HR (and its CI).
In these indirect comparisons the balance in prognostic factors from randomisation is no longer valid. They depend on the case‐mix of relevant prognostic factors (tumor stage, activity index and number of metastatic sites) and therefore have to be interpreted with caution. For this reason, they will be presented in the discussion section of the review.
STATISTICAL METHODS FOR DATA ANALYSIS: The fixed effect model will be used for meta analysis, with overall survival as primary outcome measure. Programs will be RevMan, and for more sophisticated analysis: STATA (Sterne 2002). All outcomes concerning overall survival will be recalculated (or at least approximated) by the effect measure hazard ratio for which the practical meta analysis is performed.
INVESTIGATING HETEROGENEITY: Seeking for statistical heterogeneity between studies Cochrane Q‐test will be performed (with a significance threshold of alpha =0.1). Additionally, some quantitative measure of heterogeneity will be calculated (Thompson 2002) The following factors are considered as possible sources of heterogeneity:
‐ Differences in prognostic factors
‐ Quality of studies
‐ Second‐line therapy permitted versus no second line therapy
‐ Asian versus non‐Asian studies
