Using non-inferiority test of proportions in design of randomized non-inferiority trials with time-to-event endpoint with a focus on low-event-rate setting

Ji L, Alonzo TA. Using non-inferiority test of proportions in design of randomized non-inferiority trials with time-to-event endpoint with a focus on low-event-rate setting. Clin Trials. 2025 Apr;22(2):131-141. doi: 10.1177/17407745241284786. Epub 2024 Oct 12. PMID: 39394852; PMCID: PMC11991896.

For cancers with low incidence, low event rates, and a time-to-event endpoint, a randomized non-inferiority trial designed based on the logrank test can require a large sample size with significantly prolonged enrollment duration, making such a non-inferiority trial not feasible. This paper evaluates a design based on a non-inferiority test of proportions, compares its required sample size to the non-inferiority logrank test, assesses whether there are scenarios for which a non-inferiority test of proportions can be more efficient, and provides guidelines in usage of a non-inferiority test of proportions.