Comparative survival with diverse chemotherapy regimens for cancer of unknown primary site: Multiple-treatments meta-analysis

Objectives: To synthesize the evidence from randomized controlled trials concerning systemic treatment regimens for patients with cancer of unknown primary site (CUP). Data sources: PubMed and the Cochrane Library Central Registry of Controlled Trials. Review methods: We retrieved all randomized controlled trials comparing at least two arms of different systemic treatment regimens or a systemic regimen to no treatment in patients with CUP, excluding data on favorable subset CUP, whenever these could be separated. Treatments were categorized according to whether they involved platinum, taxane, both, or neither; non-platinum/non-taxane regimens were also categorized in monotherapy and combination regimens. We extracted or estimated the logarithm of the hazard ratio and its variance for death for each randomized comparison. Multiple-treatments meta-analysis with a hierarchical Bayesian model obtained summary hazard ratios with 95% credibility intervals. Results: Ten articles were eligible for the meta-analysis. No trials compared systemic treatment to best supportive care and all arms referred to chemotherapy regimens. Overall 683 subjects were randomly assigned and eight randomized comparisons were used for the multiple-treatments meta-analysis of survival (543 patients). Multiple-treatments meta-analysis showed no significant benefit for any treatment group over others, with wide credibility intervals. Point estimates of hazard ratios favored platinum, taxane, or both (hazard ratios 0.69, 0.66, and 0.81, respectively, as compared with monotherapy with an agent other than platinum or taxane). Conclusion: No type of chemotherapy has been solidly proven to prolong survival in patients with CUP. Regimens using either platinum or taxanes or both need further testing. (C) 2009 Elsevier Ltd. All rights reserved.