Design challenges, recommendations, assumptions, and available software for early-phase trials of imunne-oncology agents
| Design Challenge | Design Recommendation | Endpoints | Model Assumptions | Available software |
|---|---|---|---|---|
| Late-onset DLTs | TITE-CRM [21]TITE-BOIN [23] | Binary toxicity | Probability of DLT increases with increasing dose level. | R Package dfcrmSAS macros https://sph.umich.edu/ccb/tite-resources.htmlwww.trialdesign.org |
| Additional endpoints | Wages & Tait [30]Zang et al. [29]Ursino et al. [55] | Binary toxicity and binary efficacy (biologic activity or clinical response)Binary toxicity and continuous PK measures (AUC) | Probability of DLT increases with increasing dose level. Probability of efficacy increases or plateaus with increasing dose level.Toxicity related to the PK measure & PK triggers toxicity when above a certain threshold | https://uvatrapps.shinyapps.io/wtdesign/www.trialdesign.orgR Package dfpk |
| Drug combinations | POCRM [37]BOIN [39]PIPE [38] | Binary toxicity | Probability of DLT increases with increasing dose level of each agent for a fixed dose level of the other agent. | R Package pocrmwww.trialdesign.orgR package pipe.design |
| Dose and schedule | Braun et al. [52]Wages et al. [54] | Binary toxicity | Probability of DLT increases with increasing dose level of each agent for a fixed schedule. | https://biostatistics.mdanderson.org/softwaredownload/SingleSoftware.aspx?Software_Id=75R Package pocrm |