First-line treatment for advanced ovarian cancer: paclitaxel, platinum and the evidence.
Bottom Line: Four main explanations for this heterogeneity have been proposed: differences in the extent and timing of 'crossover' to taxanes in the control groups; differences in the types of patient included; differences in the effectiveness of the research regimens used; differences in the effectiveness of the control regimens used.Heterogeneity 'within' and 'between' groups was formally compared using the F-ratio.Further investigation of this question would be useful.
Affiliation: Department of Public Health and Epidemiology, University of Birmingham, Birmingham, UK.Show MeSH
Related in: MedlinePlus
Mentions: The results of GOG-132 were somewhat unexpected, given the earlier results of GOG-111 and the preliminary data from OV10. The GOG-132 investigators put forward a number of possible explanations for their unexpected results including the play of chance, the different control regimens used in the trials, and the fact that a high proportion of patients on the single agent cisplatin (control) arm in GOG-132 went on to receive some treatment with paclitaxel prior to clinical evidence of disease progression. In relation to this last point, it was reported that around half of the patients on each of the three arms of the trial had gone on to receive some form of additional treatment after protocol chemotherapy but before progression of disease, with a proportion of those on the single agent cisplatin control arm receiving paclitaxel at this time. This early crossover has been widely assumed to account for the failure to detect any benefit for paclitaxel/cisplatin in the trial (Gore et al, 1997; Adams et al, 1998; National Cancer Guidance Steering Group, 1999). However, a number of individuals, including the GOG-132 investigators (Muggia et al, 2000; Torri et al, 2000), have been reluctant to accept that this explanation is necessarily sufficient to explain the difference in the results. The most obvious difficulty is that only a minority (24%) of the single agent cisplatin control group crossed over to paclitaxel before progression. Whilst this is certainly a substantial minority, if the true difference between single agent cisplatin as used in GOG-132 and paclitaxel/cisplatin were as large as the difference observed in GOG-111 and OV10 (both using a control arm of cyclophosphamide/cisplatin) it seems unlikely that crossover to paclitaxel in just one quarter of the control patients could apparently eliminate the difference between the two regimens, i.e. to give a relative risk of death (hazard ratio) of around one for both progression-free survival and overall survival (see Figure 1Figure 1
Affiliation: Department of Public Health and Epidemiology, University of Birmingham, Birmingham, UK.