The case for new approach methodologies in place of an animal study, built from cited evidence and ready for FDA discussion
The case for a shorter or replaced animal study, weighed against FDA expectations and ready to sign in days.
See one case, screen by screen ↓FDA’s 2025 roadmap opens a door for monoclonal antibodies: when the 1-month study and NAM tests show no concerning signal, the routine 6-month primate study can come down to three months. Walking through that door takes a package — and the package has to answer every question a reviewer will ask: is the monkey a relevant species, do the exposure margins hold, what do human whole blood, tissue cross-reactivity and organ chips say, and can the in silico models be trusted for the job they are doing.
The evidence is all there, but it is spread across GLP reports, toxicokinetic sheets, in vitro lab reports, CRO deliverables, papers and FDA and ICH text. The toxicologist reads it by hand, builds the matrix in a spreadsheet, checks donors and positive controls, writes the justification and the meeting questions, and then someone checks every number again. A one-donor liver chip run or a sentence that says “proves” is easy to miss in that last read.
Estimated split, in days, for one package by hand and with the solution. The prototype’s dashboard shows about 24 days by hand against a median of 5.
Seven specialist agents read, map, weigh, draft and check the evidence; a toxicologist and a regulatory lead sign.
Veltrimab, a humanized IgG1 against a cytokine receptor in plaque psoriasis, finished a clean 4-week GLP study in cynomolgus monkeys. The ask: shorten the chronic study from 6 to 3 months. Here is PKG-0412, screen by screen, in the working solution.
Lena opens the study-reduction matrix: 9 programmes against the 1-month GLP, chronic, embryo-fetal / ePPND, transgenic or surrogate and second-species studies in their plans. Each cell says Shorten, Replace with NAMs, Weight of evidence, Keep as planned or Completed. Across the portfolio, 112 monkeys could be spared — 40 agreed with FDA, 72 proposed.
“Where can a monkey or mouse study be shortened or replaced? Every proposal is built from cited evidence and signed by a person.”
She clicks the Veltrimab chronic-study cell. The panel explains the route and shows PKG-0412 at a score of 81, In review, one gap, 120 NHP dosing-months — with each of the eight FDA expectations marked Met, Partly or Gap.
“The 1-month GLP study and human-based data show no concerning signal, so the chronic study can be proposed at 3 months instead of 6 — the route FDA’s April 2025 roadmap sets out for monoclonal antibodies.”
The package opens on a balance. Supports: a clean 4-week GLP study with the NOAEL at the top dose, exposure margins of 38× (AUC) and 52× (Cmax), no cytokine release in 12 human donors, only expected binding in 36 human tissues. Open concerns: the liver chip run used one donor, and ADA appeared in 2 of 10 high-dose animals. 7 of 8 expectations met, 46 citations checked, model risk Medium.
The evidence matrix grades each FDA expectation against five lines of evidence — animal studies, human in vitro, in silico, literature and class, clinical. Lena clicks cytokine release in human in vitro: Strong, “Whole blood, 12 donors: no cytokine above isotype”, with the passage of report CRA-0311 highlighted beside it, positive control and all.
The monkey cytokine data on the same row are graded Weak — low predictivity for people.
LC-0131 arrived from the in vitro lab: the liver chip across three hepatocyte donors. Lena presses “Add the 3-donor results”. The NAM assay reader reads 42 endpoints and confirms the reference hepatotoxin was detected in all three chips; the mapper moves organ toxicity in human systems from Gap to Strong; the critic checks the two new citations; the writer updates section 4 and the position under question 3.
“Gap closed — organ toxicity is now Met by two lines of evidence.” Score 81 → 88, 8 of 8 expectations met.
ADA appeared in 2 of 10 high-dose animals, but in the one with lower exposure the AUC margin is still 25-fold, and monkey ADA does not predict human ADA. Lena accepts the mitigation; her reason goes on the audit trail and the score moves to 91.
“Exposure maintained with ≥ 25-fold margin; ADA risk in people addressed by the in silico model, DC–T-cell assay and Phase 1 (3%).”
Two computational models stand in the package. The in silico immunogenicity model v3.2 is taken through question of interest, context of use, model risk, credibility plan, execution, documentation and adequacy: influence medium × consequence medium, checked against 217 antibodies with known clinical ADA, with 84% correctly classified. Adequate for a supporting role — not as the only immunogenicity evidence.
Eight sections — purpose and request, programme and relevance of the species, findings to date in animals, human-based evidence, weight of evidence, proposed study plan, model credibility, limitations and commitments. The proposed plan: a 3-month GLP study at 0, 10, 30 and 100 mg/kg/week with an 8-week recovery group, toxicokinetics and ADA, NAM data submitted in parallel.
The critic removed the word “proves” from section 5.
Each question is written as “Does the Agency agree…”, with a cited sponsor position beneath it — from whether the 4-week study plus human in vitro data show no concerning signal, to the credibility assessment for the immunogenicity model. Type C is set: request Oct 20, 2026, FDA meeting within 75 days. Omar has already given the regulatory sign-off.
Before she signs, the checks are shown: every expectation has evidence, every sentence cited — 48 citations checked by the critic, no overstated wording, model credibility assessed for 2 models, regulatory sign-off by Omar Haddad. Lena signs with her password; the meaning, “Approved for submission to FDA”, is recorded with it, and Priya Raman, the 3Rs officer, is told.
Signing with a gap open is allowed — but the gap and the reason go on the record.
The same package, seen by the five people who build it, sign it and report on it — what the work looked like, and what it looks like now.
Seven specialist agents read, map, weigh, draft and check the evidence; a toxicologist and a regulatory lead sign.
Reads GLP study reports, SEND datasets and toxicokinetic tables; extracts design, findings, NOAEL, exposure and ADA with page citations.
Reads cytokine-release, tissue cross-reactivity, membrane-protein screens and organ-chip reports; checks controls, donors and acceptance criteria.
Finds target biology, class experience, NAM predictivity studies and FDA or ICH statements; watches for new ones weekly.
Builds the evidence matrix — FDA expectation × evidence line — grades each cell, finds gaps and weighs the case.
Runs FDA’s 7-step risk-based credibility framework for each computational model whose output stands in for an animal endpoint.
Drafts the justification and the FDA meeting questions with a sponsor position under each; every sentence cited.
Checks every number and quote against its source and every claim against the matrix before a person sees the draft.
Both sign the package. The agents propose; a named person decides.
Toxicologists, regulatory leads and the 3Rs officer ask in plain words — or tell it what to change. Every answer cites the passage behind it.
Why is Veltrimab ready to ask for a 3-month study?
The 4-week GLP study found no adverse effects up to 100 mg/kg, with margins of 38× (AUC) and 52× (Cmax). Human-based data agree: no cytokine above isotype in 12 donors, expected staining only in 36 human tissues, and no injury signal in the three-donor liver chip. That fits FDA’s roadmap for antibodies. All eight expectations are now met.
What is still weak in the Teriquamab package?
PKG-0409 has 3 gaps out of 8 expectations: organ toxicity (liver chip run not yet done), immunogenicity (DC–T-cell assay) and target biology (human genetics review). Because the target is human-specific, FDA’s roadmap names this case — a human in vitro battery plus a PBPK model instead of a transgenic mouse. The PBPK model is high risk under the credibility framework because it carries the conclusion alone.
Add a rule: every cytokine-release assay needs a positive control
Done. New evidence rule: a cytokine-release assay without a responsive positive control is graded as a gap. I re-checked the open packages — all 3 cytokine assays have one (anti-CD3 responded in 12 of 12 donors), so no grade changes. The rule is now in Settings.
How many animals could we spare this year?
Over the last 12 months: 196 monkeys that could be spared if FDA agrees, of which 124 are already agreed, plus 120 mice and 360 monkey dosing-months from three chronic studies cut from 6 to 3 months. Packages reached a draft ready to sign in a median of 6 days, against about 24 by hand.
13 screens from the working solution, on its sample data. Pick one to see it large.
Who signs, the rules the agents follow and where the evidence comes from are settings, changed in plain words.
| Setting | Default | Choose from |
|---|---|---|
| Scientific sign-off | Dr. Lena Ortiz | Dr. Lena Ortiz · Dr. Maya Chen |
| Regulatory sign-off | Omar Haddad | Omar Haddad · Sam Patel |
| Tell when signed | Priya Raman · 3Rs officer | Priya Raman · Nobody |
| Meaning of signature | Approved for submission to FDA | Approved for submission to FDA · Reviewed, not yet for submission |
| Two independent lines of evidence per expectation | On | On · Off |
| Flag human in vitro results from fewer than 3 donors | On | On · Off |
| Credibility assessment for every model standing in for an animal endpoint | On | On · Off |
| Re-check open packages when FDA or ICH publish something new | On | On · Off |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample programme data · “target” = the design goal, measured in the live solution · “estimated” = our estimate · “typical” = published figures (FDA, Roadmap to Reducing Animal Testing in Preclinical Safety Studies, 2025 — about 144 non-human primates in a typical antibody programme, at up to $50,000 per animal). Sources also referenced: FDA draft guidance on AI to support regulatory decision-making (January 2025), ICH S6(R1). People named on this page are characters in the working solution, and the programmes are its sample data.
It is the case for using new approach methodologies — human whole blood, tissue cross-reactivity, organ chips, in silico models — in place of or alongside an animal study, written for FDA discussion. In this solution it holds the evidence matrix, the weight of evidence, the justification, the model credibility assessments and the meeting questions, signed before it goes out.
The evidence mapper grades eight expectations — from relevant species and exposure margins to cytokine release, off-target binding, organ toxicity in human systems, immunogenicity and target biology — against five lines of evidence. An expectation is Met only when two independent lines agree, and gaps are shown rather than hidden.
That is the hero case in the working solution: a monoclonal antibody with a clean 4-week GLP study, following the roadmap item FDA published in April 2025. FDA decides whether it agrees. The solution builds and checks the package, and its counts of animals that could be spared stay targets until FDA agrees.
Each computational model standing in for an animal endpoint is taken through FDA’s seven-step risk-based credibility framework. Human in vitro results from fewer than three donors are flagged, and single-donor organ-chip or cytokine data are graded as a gap. A missing or failed positive control is flagged.
Yes. The package writer drafts questions in the “Does the Agency agree…” form with a cited sponsor position under each. The regulatory lead edits, includes or leaves out each question and chooses Type C, Type D or pre-IND.
Yes. Agents read, grade and draft; a toxicologist gives the scientific sign-off and a regulatory lead the regulatory sign-off, each with password and the recorded meaning of the signature. Concerns are resolved by a person with a reason, and every step is on the audit trail.
The GLP study archive, SEND datasets, the in vitro lab system, a CRO portal, a weekly literature search and an FDA and ICH watch. Sources are read in place, and new results are mapped to open packages and their owners are told.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your expectations checklist, house style for briefing packages and a sample of your study and assay reports — and it goes live once every quality gate has passed.
Ask the literature in plain words and get a sourced answer in seconds.
1-min filmPipeline WatchEvery rival trial, readout and approval lands in a sourced brief for each programme, every Monday.
1-min filmPreclinical ReportsRaw study data becomes a draft report in hours, with every number traced to its source.
1-min filmWe’ll build a NAMs evidence package from one of your own programme’s study reports and assay data.