ABSTRACT
In the bortezomib trial, each patient would receive up to six 21-day cycles of chemotherapy, and would thus be at risk of toxicity throughout the entire treatment period. If each patient was to be followed for 126 days before a dose decision was made for the next patient, the trial duration would be impractically long. On the other hand, the common phase I practice would usually count toxicity occurring only in the first cycle, and would underestimate the incidence of adverse events as toxicity might occur at a later cycle. To deal with the prospect of late toxicity, the bortezomib trial adopted a variant of the CRM, called the time-to-event CRM (TITE-CRM) that would allow continual accrual throughout the trial while using the 126-day toxicity endpoint as basis of dose escalation.
