OnCo
ideasIdea

Retire the 3+3: model-based dose finding that counts late and chronic side effects

The traditional way of finding a dose looks only at severe side effects in the first month. Modern statistical designs can use all patients' data and count the grumbling, long-lasting problems that make people quit months later.

Phase 1 designs (BOIN, CRM, TITE variants) that incorporate late-onset toxicity, cumulative low-grade toxicity over multiple cycles, patient-reported tolerability and pharmacokinetics into dose selection, with a target of the 'optimal biological dose' rather than the MTD, and with backfill cohorts at lower doses. Regulators expect a model-based design; the 3+3 becomes the exception requiring justification.

Hypothesis
Model-based designs incorporating late toxicity will recommend lower doses than 3+3 for the same agents in a substantial fraction of cases, and those doses will show lower real-world discontinuation.
Rationale
For chronic oral therapies and immunotherapies, the toxicities that matter appear after cycle 1 and are often grade 2 but persistent; 3+3 is statistically inefficient and structurally blind to them.
What would test it
Re-analyse completed phase 1 datasets under TITE-BOIN with cumulative toxicity and compare recommended doses to the original; prospectively adopt in a sponsor's early-phase portfolio and track later dose changes.
Maturity
early clinical
Who has to act
industry
Cost to try
Small (under $1M)
Years to first evidence
2
Bottlenecks it attacks
  • Wrong doses · Most drug doses were chosen as the highest a person can tolerate, which is often more than they need.
  • Trial design, endpoints and cost · A phase 3 trial takes years and hundreds of millions of dollars, and often answers a question that has already moved on.

Connected

6top