Objective To explore the application effectiveness of multiple classic and innovative early-stage clinical trial design and analysis methods in various dose-response scenarios.

Methods Combined with clinical experience, the dosetoxicity curve (monotonously increasing) and dose-efficacy curve (monotonously increasing and non-monotonously increasing) were designed for 4 scenarios at 4 dose levels. The R software was used to perform numerical simulation analysis, and the safety and efficacy measures of EffTox, STEIN, UBOIN, TEPI, PRINTE and Ji3+3 were compared under different scenarios and tested on the actual data set.

Results STEIN had good safety and efficacy, and it was very robust in multiple dose-efficacy scenarios; the efficacy of EffTox measures were less robust, especially for different dose-efficacy scenarios; TEPI required sacrificing more toxic dose patients and greater proportion of optimal biological dose (OBD) patients to proportion of toxic dose selection and OBD accuracy of PRINTE. The performance trend and effect of Ji3+3 were similar to those of PRINTE; UBOIN was slightly less effective than PRINTE and Ji3+3.

Conclusions STEIN is suitable for a variety of dose-efficacy scenarios, with good and stable results; it is not recommended to use EffTox in non-monotonously increasing dose-efficacy scenarios; PRINTE has better safety and efficacy than TEPI; UBOIN may be preferred in a trial expected lower number of patients with toxic doses.