OnCo
ideasIdea

Comparability by design: a digital twin and sentinel panel for cell process changes

Improving how a cell therapy is made currently risks having to repeat clinical trials. A validated computer model plus a fixed set of product measurements would let changes be approved on data alone.

Because potency assays are imperfect, regulators often require clinical bridging when a cell-therapy process changes (new device, new site, shorter culture), freezing suboptimal processes. The proposal is a pre-agreed comparability framework: a mechanistic and statistical model of the process (a digital twin) validated against historical runs, a sentinel panel of product attributes (phenotype, transcriptomic signature, cytotoxicity, cytokine profile, vector copy number) with pre-specified equivalence margins, and a regulatory commitment to accept changes that fall within the margins without clinical data.

Hypothesis
Under the framework, the median time to implement a major cell-therapy process change falls from more than two years to under six months, with no detectable change in clinical outcomes in the outcomes registry.
Rationale
Biosimilars established that analytical similarity within margins can replace clinical trials; cell therapies need richer attribute panels, but single-cell and transcriptomic characterisation now make such panels feasible.
What would test it
Retrospectively apply the sentinel panel to products that underwent process changes with known clinical outcomes to calibrate margins; then pilot the framework prospectively with two manufacturers and regulators.
Maturity
speculative
Who has to act
regulator
Cost to try
Medium ($1M to $50M)
Years to first evidence
4
Bottlenecks it attacks

Connected

4top

Pages like this

not linked directly; found by shared links