Is the addition of Bevacizumab to ovarian cancer treatment chemotherapy a settled question? It seems to have pushed through the gate with increases in progression free survival (PFS), but not overall survival (OS), particularly in platinum resistant disease.
A paper was published in the New England Journal of Medicine on February 25, 2016. It compared progression free survival treatment with weekly Paclitaxel versus every three weeks in combination with Carboplatinum. During the time this study was enrolling patients, another study, GOG trial 218 showed improved progression free survival (14.1 vs 10.3 months) by adding Bevacizumab to standard Carboplatinum + Paclitaxel every 3 weeks. Therefore study participants were given the option of adding or not adding Bevacizumab to their assigned treatment. The mission of this study was to see if weekly Paclitaxel was better than 3 weekly treatment. They concluded that weekly Paclitaxel /Carboplatinum with Bevacizumab was no better than 3 weekly treatment.
Median Progression free survival With Bevacizumab Without Bevacizumab
Weekly paclitaxel/ carboplatinum 14,9 months 14.2 months
every 3 weeks
Paclitaxel and carboplatinum 14.7 months 10.3 months
every 3 weeks
And that was the only conclusion the authors came to. So I made a table of the results in a way that I could understand, and found the only inferior result was 3 weekly treatment without Bevacizumab. I went back to the GOG 218 study and found that the benefit of Bevacizumab was increasing PFS to 14.1 months. Granted, one cannot compare across trials (different patient groups, etc), but weekly Paclitaxel with 3 weekly Carboplatinum without Bevacizumab is at least in the same ballpark PFS. So why is Bevacizumab (expensive, has toxicity) now a standard recommendation? Dose dense, metronomic use of paclitaxel seems to have an equivalent antiangiogenic effect.
I wish the study authors had further discussed this issue. However they do add “in addition, comparative effectiveness studies are needed with consideration of economic costs. Although paclitaxel and carboplatin administered every 3 weeks and combined with bevacizumab may be more convenient than weekly paclitaxel and carboplatin without bevacizumab, the every-3-week regimen is also associated with higher costs, with an incremental cost-effectiveness ratio as calculated by others of $401,088 versus $5,809 per progression-free life–year saved.”
When clinical trials are conducted, enormous amounts of complicated statistics are involved. The question to be answered, for example drug X in schedule P, or schedule Q, OR drug X vs Drug X+Y in schedule P vs Q is proposed. The statisticians determine how many patients need to be enrolled to be able to answer the question. The enrolled patients need to be balanced between all the groups regarding their clinical characteristics. Analyses of the results are done periodically as the trial progresses, to make sure that if one group is doing far better, or worse than the others, it is ethical for the trial to continue. At the end of the trial, the statisticians go to work on the numbers, and produce graphs and tables. These are buried in the body of the paper. The study authors then provide their conclusion in their discussion.
A practicing physician relies on the analysis and discussion.
Somewhere in the raw data there can be buried an alternate conclusion.