Restricted Net Treatment Benefit in oncology

Area of Science:

• Biostatistics
• Clinical Trial Design
• Survival Analysis

Background:

• The restricted Net Treatment Benefit (rNTB) is a clinically meaningful estimand for overall treatment effect in randomized trials.
• It is particularly useful when survival endpoints with time restrictions are employed.
• rNTB interpretation does not require parametric assumptions like proportional hazards and is unbiased with independent right-censoring.

Purpose of the Study:

• To demonstrate that rNTB, representing NTB within a defined time interval, is a meaningful and adaptable measure of treatment effect.
• To evaluate the impact of various factors on rNTB estimation and power.
• To compare rNTB-derived tests with traditional log-rank tests.

Main Methods:

• A simulation study was conducted to assess the influence of delayed treatment effects, minimal clinically relevant difference thresholds, restriction time, and the inclusion of efficacy and toxicity.
• The bias introduced by right-censoring on rNTB was evaluated.
• Power comparisons were made between rNTB-derived statistical tests and log-rank tests.

Main Results:

• rNTB estimates remain unbiased even with right-censoring.
• rNTB facilitates benefit-risk ratio estimation, incorporating survival and toxicity.
• rNTB is more interpretable than NTB with censoring, as it is not dependent on follow-up duration.
• Including toxicity enhances test power when the experimental treatment is less toxic.
• rNTB-derived test power increases with longer survival and lower toxicity, and may improve with a cure rate or delayed treatment effect.
• Case applications from PRODIGE, Checkmate-066, and Checkmate-067 trials are presented.

Conclusions:

• rNTB is a valuable alternative for describing and testing treatment effects clearly, especially with restricted timeframes.
• It is particularly advantageous in scenarios with nonproportional hazards or when balancing benefit and safety.
• rNTB can be adjusted for short- or long-term survival differences and multiple prioritized outcomes.