Biomarker Strategy Planning: a cited biomarker strategy in days, with every assay timed to the study that needs it
A cited biomarker strategy for every programme in days, each assay timed to its study.
See one case, screen by screen ↓A plan for which biomarkers a programme uses to show target engagement, pharmacodynamic change, select patients and see harm early — and, for each, the assay, whether it is validated for that context of use and when it is ready for the study that needs it. Biomarker Strategy Planning builds it with six agents and shows it as a strategy map across the phases, with every claim cited.
How to choose AI solutions for life sciences →
Updated
Every programme that reaches a stage gate needs a biomarker strategy: which markers show target engagement, which show the biology moving downstream, which select the patients and which show harm early — and for each, whether an assay is validated for that context of use. Today a translational medicine lead reads the literature, the Phase 1 biomarker and laboratory reports, vendor specifications, registry records and guidance, then writes it by hand in the house template over four to eight weeks. A biomarker scientist re-checks the numbers, and the board gets a static document that ages the day it is signed.
What it rarely shows is when each assay will actually be ready for the study that needs it. A patient-selection test that is late for its study is found late — often after rivals in the same indication are already enrolling on it — and the trial pays in scans, screen failures and months. In the sample case, a late CLIA transfer means study 201 screens every candidate by amyloid PET: about 420 scans.
Estimated split for one programme’s biomarker strategy before a stage gate, by hand and with the solution; the totals follow the solution’s own figures of about 38 days by hand and a 6.5-day target.
Four readers gather evidence and the field in parallel, a writer drafts, a reviewer times every assay against its study; a scientist checks the values and the board signs.
Northwind Bio’s NWB-3140, an anti-amyloid antibody in early Alzheimer’s disease, is moving from Phase 1b to Phase 2. Study 201 (n = 180) starts in May 2027, study 202 (n = 600) in Q2 2028, the pivotal study in 2029. Strategy v3 is at scientist review and the translational board meets on Oct 21. Here is what happens, screen by screen, in the working solution.
NWB-3140 opens on its strategy map: target engagement, pharmacodynamic, patient selection and safety as lanes, Phase 1b to Phase 3 as columns, each card with its assay, readiness and the decision it supports. The header reads 9 biomarkers, 3 with an open gap, 1 not ready when needed. In patient selection, the Phase 2a card is red: plasma p-tau217, “Pre-screen not ready: CLIA transfer ends Jun 2027”.
“Show me which test is ready for which study — not a list of markers.” — Dr. Maya Chen
Below the lanes, the field lane places rival trials in the same indication by phase, read from registries, posters and papers. HLV-210 (Halvor Therapeutics, Phase 3, n = 1,600) pre-screens with plasma p-tau217 and confirms with PET, enrolling since Mar 2025. SLW-4 (Selwyn Bio, Phase 3, n = 1,150) enriches with p-tau217 plus Aβ42/40, enrolling since Nov 2024. Both carry the flag “same marker as ours”.
“Two rivals already enrol on plasma p-tau217 … our plan starts it only at Phase 2b, so study 201 screens ≈ 420 candidates by PET.”
The science is not the problem — AUC 0.93 against amyloid PET, 91 % sensitivity, 89 % specificity — the laboratory is. Maya asks the agents. The trial design analyst re-reads the HLV-210 and SLW-4 registry records, the assay and precedent reader checks 6 CLIA laboratories for the same digital immunoassay platform, and the gap and risk reviewer compares timelines with first participant in.
“Arden Clinical Laboratory: platform in place, validation slot from Jan 11.”
Option A: a second CLIA laboratory. Arden Clinical Laboratory runs the same platform and has a January slot — transfer in 14 weeks, ready March 2027, two months before first participant in. Cost $165,000, about 190 PET scans avoided, about $1.0 M and 9 weeks of enrolment saved. Option B: keep the plan — PET for every candidate in study 201, about 420 scans, no transfer risk.
The estimates are on screen: PET ≈ $5,800 a scan, pre-screen ≈ $240 a test, 55 % of candidates go on to PET.
Maya clicks “Propose option A”. The patient-selection card now shows the pre-screen as proposed for study 201, the gap records the option, and the decision goes to Omar Haddad, the diagnostics lead, due Nov 14. Nothing is booked: an option stays a proposal until he decides.
“Option added to the plan. CLIA transfer at Arden Clinical Laboratory by Mar 2027; Omar Haddad decides on Nov 14.”
The assay development plan lays out each activity with its owner against the study calendar: the Kestrel CLIA transfer, the proposed Arden transfer, two-platform bridging in 120 Phase 1b samples, GFAP re-validation with stabiliser tubes, APOE genotyping set-up, the central imaging charter — and the milestones for protocol final, the locked sample-collection plan and first participant in.
Critical path: “GFAP re-validation ends in mid-March, just before the sample-collection plan is locked.”
Every extracted value sits beside the passage it came from. Plasma GFAP is waiting for him at 0.82 confidence: sample stability study SS-118 reports 84 % recovery after 9 months at −80 °C, below the 85 % acceptance criterion. Values under the 0.90 threshold wait for a biomarker scientist; he confirms them or corrects a value.
“A number I cannot trace back to the page does not go in the strategy.” — Sam Patel
Strategy v3 follows the house template — context of use, target engagement, pharmacodynamics, patient selection, safety, assay development plan, regulatory route, gap analysis — with 112 cited claims from 47 sources and 0 uncited claims. Each section shows who checked or co-signed it, and each citation opens the source at its passage. A section can be regenerated; the whole strategy exports as .docx with the evidence table as .xlsx.
The reviewer reads the draft like a board member and lists the gaps: GFAP stability (High, Sam Patel, Nov 30), the p-tau217 pre-screen not ready for study 201 (High, Omar Haddad, Nov 14), tau PET tracer at only 9 of 40 sites (Medium, Kenji Mori, Dec 15), no two-platform bridging (Medium, Sam Patel, Jan 15) and the cost of the intermediate zone (Low, Maya Chen, Dec 1).
The gap and risk reviewer never closes a gap — only a person can assign it or accept the risk.
Maya signed as author, Sam Patel checked the values, Lena Ortiz co-signed target engagement and Omar Haddad co-signed patient selection. The strategy goes to the translational board for Oct 21. Dana Okafor decides — approve, approve with conditions, or return for rework — with an electronic signature recording the date, time and reason.
“All co-signs are in. Send the strategy to the translational board — the pack closes Oct 14.”
The same strategy, seen by the five people who carry it to the board — what their weeks looked like, and what they look like now.
Four readers gather evidence and the field in parallel, a writer drafts, a reviewer times every assay against its study; a scientist checks the values and the board signs.
Finds and reads papers and preprints for each candidate biomarker in full text and extracts sensitivity, specificity, AUC, cut-offs and cohort details, with the cohort size, reference standard and page.
Reads our own Phase 1 biomarker reports, laboratory validation reports and stability studies, and extracts effect sizes, confidence intervals, LLOQ, precision, parallelism and stability against the report’s acceptance criteria.
Reads rival trial registry records and congress posters and records how each biomarker is used — selection, enrichment, stratification, endpoint, safety — with cut-offs and timing, compared with our design.
Reads vendor specifications, validation reports and regulatory records; sets each biomarker’s validation status and regulatory precedent for its context of use, and finds CLIA laboratories and slots when an assay is late.
Drafts the strategy in the 8-section house template — context of use, target engagement, pharmacodynamics, patient selection, safety, assay development plan, regulatory route, gap analysis — with every factual sentence cited.
Reviews the draft like a board member: unsupported or single-source claims, missing validation, lead times and site feasibility. Times each assay against first participant in, lists every gap with severity, owner and date, and costs the options.
Signs, with conditions on record. The agents propose; a named person decides.
Ask in plain words about a programme, a marker, a rival design or a gap. Every answer cites its source and can open the screen it refers to.
When can we pre-screen with p-tau217, and who already does?
Not before Phase 2b as planned: the CLIA transfer at Kestrel Assay Labs ends in late June 2027, after study 201 starts in May. Two rivals already enrol on it — HLV-210 (Phase 3, plasma p-tau217 → PET, since Mar 2025) and SLW-4 (Phase 3, p-tau217 + Aβ42/40, since Nov 2024) — while NWB-3140-201 uses PET for every candidate, about 420 scans. The evidence is strong, AUC 0.93 against amyloid PET, so the limit is the laboratory, not the science.
Which programmes lack a validated patient-selection assay?
Three: Veltrimab (c-Met IHC — trial assay, companion test not contracted), NWB-8840 (B7-H3 IHC — assay not validated) and NWB-9152 (interferon signature — exploratory only). Veltrimab is the urgent one: co-development needs about 24 months and Phase 3 starts Q4 2027.
Why is GFAP a gap?
Plasma GFAP recovered only 84 % after 9 months at −80 °C, below the 85 % acceptance criterion, and Phase 2 samples will be stored up to 18 months. The leading competitor Phase 3 measures GFAP at weeks 26, 52 and 78. Proposed: re-validate with stabiliser tubes (Tamsin, Dec–Mar) — owner Sam Patel, due Nov 30.
How do we manage ARIA risk?
Two measures: APOE ε4 genotype before randomisation — ARIA-E 34 % in homozygotes against 11 % in non-carriers — and MRI before doses 3, 5, 7 and 14; Phase 1b ARIA-E was 2 of 22 (9 %), all asymptomatic. Unlike one competitor, homozygotes are included, with extra MRI.
13 screens from the working solution, on its sample data. Pick one to see it large.
The house template, the evidence rules the agents follow on every run, and who signs what are settings.
| Setting | Default | Choose from |
|---|---|---|
| House template | Biomarker strategy v4 — 8 sections | Biomarker strategy v4 — 8 sections · Biomarker strategy v3 — 6 sections · Early-development biomarker plan — 5 sections |
| Citation style | Numbered | Numbered · Author–year |
| Independent sources per key claim | 2 | Step up or down; claims with fewer sources are flagged in the draft |
| Flag single-source claims | Off | On · Off |
| Scientist check below this confidence | 0.90 | Step up or down; values under it wait for a biomarker scientist |
| Recency window | 5 years | 3 years · 5 years · 10 years — older papers only when nothing newer exists |
| Preprints | Allowed, always labelled as not peer reviewed | On · Off |
| Approvals | Values: Sam Patel · target engagement: Lena Ortiz · patient selection: Omar Haddad · board chair: Dana Okafor | An approver chosen for each section and for the board |
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 (BIO/Biomedtracker/Amplion, Clinical Development Success Rates 2006–2015 — 25.9 % vs 8.4 % likelihood of approval from Phase I with and without selection biomarkers). Context of use as in FDA, Bioanalytical Method Validation for Biomarkers, guidance for industry, 2025. People, companies, programmes, laboratories and trials named on this page are fictional — characters and sample data in the working solution.
A plan for which biomarkers a programme uses to show target engagement, pharmacodynamic change, select patients and see harm early — and, for each, the assay, whether it is validated for that context of use and when it is ready for the study that needs it. Biomarker Strategy Planning builds it with six agents and shows it as a strategy map across the phases, with every claim cited.
The gap and risk reviewer times each assay against first participant in for the study that needs it, using the assay plan and the study calendar. A card that will not be ready when its phase needs it turns red, and the gap is listed with an owner and a date. In the sample case it found a CLIA transfer ending in June 2027 for a study starting in May 2027.
The trial design analyst reads rival registry records and congress posters, from public sources only, and records how each marker is used — selection, enrichment, stratification, endpoint or safety — with cut-offs and since when. The field lane places those trials under our lanes and flags the ones using the same marker as ours.
The assay and precedent reader sets each marker’s validation status — from candidate to regulatory precedent — for its context of use, and never infers qualification without a record. Each card names the decision the marker supports, so the depth of validation can be read against it.
No. Agents report numbers exactly as written, values under the confidence threshold wait for a biomarker scientist, and the writer uses only checked values. The gap and risk reviewer never closes a gap or chooses an option; a person does.
Yes. Agents can move a strategy no further than scientist review. Section owners co-sign their sections, the diagnostics lead decides laboratory and diagnostic routes, and the translational board chair approves, approves with conditions or returns the strategy with an electronic signature and a reason.
Yes. The house template, citation style, number of independent sources per key claim, single-source flag, confidence threshold for a scientist check, recency window, preprints and the approvers for each section are all settings.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your house template, a past strategy, your Phase 1 reports and the sources you license — and it goes live once every quality gate has passed. We will walk you through it on one of your own programmes first.
Sourced answers from papers, patents and reports in seconds, our own disagreements first.
1-min demoCompetitive Pipeline TrackingEvery rival trial, readout and approval in a sourced brief each Monday
1-min demoPreclinical Study Report AutomationA draft GLP report in hours, every number traced to raw data.
1-min demoWe'll run Biomarker Strategy Planning on one of your own programmes before its next stage gate.