In this week’s blog post, we look at an extract from an article written by CROS NT’s expert biostatistician, Thomas Zwingers, on adaptive trial design for oncology studies. How can biostatisticians use adaptive trial designs to deal with the advantages and disadvantages of endpoints in oncology studies?
Oncology is significantly different than other therapeutic areas. One contributing factor is the long timelines to reach clinical endpoints. The ultimate endpoint for registration of a new drug is still the “Overall Survival Time”. Surrogate endpoints like “response rates” or “time to progression” are often used but mostly as secondary endpoints or in earlier phases of the development process.
Let’s look at the advantages and disadvantages of these endpoints:
- Overall Survival: the advantages are that it is universally accepted as a direct measure of benefit and easily measured, however its disadvantages are that it usually involves larger studies, may be affected by crossover therapy and sequential therapy and includes non-cancer deaths.
- Disease-Free Survival or Progression-Free Survival: The advantage to this endpoint is that is usually requires a smaller sample size and shorter follow-up compared to survival studies. However, this endpoint is not statistically validated as a surrogate for survival in all settings, and definitions may vary among studies.
- Time to Progression: The advantages to TTP endpoints are that they usually require a smaller sample size and shorter follow-up and they are not affected by crossover or subsequent therapies. Like the above-mentioned endpoint, TTP is not statistically validated as a surrogate for survival. Additionally, it is not precisely measured and is subject to assessment bias particularly in open-label studies. Definitions may also vary among studies.
The long time period before any conclusion from the study can be drawn is the main reason for the use of adaptive designs in this setting. Adaptive designs enable researchers to make early decisions on interim analyses on one or more of the following issues:
- Early stopping
- Adjustment of sample size
- Treatment selection
- Change of observation time
- Change of test statistic
In Phase III trials, or even Phase IIb trials, the main accepted endpoints usually need a long time before they can be evaluated. Implementing a Seamless Phase II/III adaptive design can minimize time and save costs in terms of patient enrollment.