CTD Module 2 authoring with a citation on every statement and every fact checked across modules
Module 2 summaries drafted in weeks, every statement cited to its source.
See one case, screen by screen ↓Module 2 is the part of the dossier a reviewer reads first: the Quality Overall Summary in 2.3, the nonclinical and clinical overviews in 2.4 and 2.5, and the written summaries in 2.6 and 2.7. It condenses thousands of pages of CSRs, integrated summaries and Module 3 documents, and it is written by several authors — a clinical lead, a clinical scientist, a CMC regulatory lead — each from the report version they had at the time.
The same fact is stated in more than one place. Patients randomized appear in 2.5.4, 2.7.3.2 and 2.7.6; patients exposed in 2.5.5 and 2.7.4.1; a specification limit in 2.3.P.5 and 3.2.P.5.1. When a data cut moves or a draft is superseded, one place is left behind. Today a QC reviewer re-reads every number against its source page at the end of the timeline — or the agency finds it.
Estimated split of a first Module 2 draft, by hand (about 22 weeks on the sponsor's last submission, as the working solution shows) and with the solution (the 5-week target).
Six specialist agents read the sources, draft each section, open every cited page and check each fact across modules; the section owner applies fixes and signs.
Veltrimab BLA — a humanized anti-IL-23p19 antibody for moderate to severe plaque psoriasis — is due for submission on December 14, 2026, as eCTD sequence 0000. It is October 6, 69 days out. Here is what happens next, screen by screen, in the working solution.
Maya owns 2.7.3 Summary of Clinical Efficacy. The binder shows the ICH M4 tree — 14 sections, 4,812 statements each cited to a page, 312 facts across modules, 6 of 14 signed. Every node shows the studies it cites, a citation-coverage ring and who signs. Her own section already carries a conflict: 2.7.3.2 says 818 patients randomized in VLT-302; 2.5.4 says 824.
“2.7.3.2 uses the full-analysis-set count (818) but calls it ‘randomized’. The final CSR and 2.5.4 both give 824 randomized.”
Sam Patel, the clinical scientist, has just updated 2.7.4.1 Exposure to the drug. Maya opens 2.7.4 and the consistency agent checks its 29 facts against 2.5, 2.7 and Module 5, live. 2.7.4.1 states 1,947 patients exposed, from the final Integrated Summary of Safety (Sep 9, page 22, Table 1.2). 2.5.5 says 1,935 — it still cites the June data cut, now superseded. Twelve patients apart, both pages side by side, and the owner named: Dr. Omar Haddad owns 2.5.
“Patients exposed to veltrimab: 2.5.5 says 1,935, 2.7.4.1 says 1,947.”
Maya opens the source 2.5.5 cites: Integrated Summary of Safety — June data cut, draft v0.9, page 22 of 296, marked superseded. The passage is highlighted: 1,935 subjects received at least one dose, for 2,207 patient-years of exposure, data cut June 30, 2026.
The agent opens both versions of Table 1.2, finds that 12 patients enrolled in VLT-303 after the June 30 cut, and proposes a tracked change to 2.5.5 cited to the final report, page 22: “Across the program, 1,947 patients received at least one dose of veltrimab, for 2,316 patient-years of exposure.”
The consistency agent proposes; it never applies a fix or closes a conflict.
Maya does not own 2.5, so she cannot apply the change — she sends it to Dr. Omar Haddad, the clinical lead who owns the Clinical Overview. The conflict stays open on the node until he applies it.
“Fix sent to Dr. Omar Haddad. 2.5.5 will read 1,947 once Omar Haddad applies it; the conflict closes then.”
In the 2.7.3 draft, 213 of 214 statements are verified against their cited pages. One is red: “In VLT-302, 818 patients were randomized 2:1 to veltrimab (n=546) or placebo (n=272).” The citation verifier opened CSR VLT-302, page 112, Table 14.1.1.1 Subject disposition: 824 randomized, 818 in the full analysis set. Until it is fixed, Approve section will not let her sign: “2.7.3 cannot be signed yet.”
“Disagrees with the source. Says 818 randomized. The cited table shows 824 randomized and 818 in the full analysis set — and 2.5.4 states 824.”
On Facts across modules, the agent compares populations and proposes: “In VLT-302, 824 patients were randomized 2:1 to veltrimab (n=549) or placebo (n=275); 818 received at least one dose and formed the full analysis set.” It re-checks 2.5.4 and 2.7.6 — both already say 824. Maya owns 2.7.3, so she applies it herself.
“You decide — nothing changes until it is applied.”
Now 214 of 214 statements are verified, all figures match 2.5.4, 2.7.6 and the sources, and the one tracked change is listed. Maya signs as Dr. Maya Chen · Clinical efficacy lead, meaning “Approved as section owner”, with her password.
“Your signature is recorded with date, time and meaning. The section is locked as version 1.0 and goes to the eCTD build.”
2.7.3 is approved and locked as v1.0. The corrected statement shows who did what: edited by the consistency agent, applied by Dr. Maya Chen, re-verified against the cited page. The history lists each agent step and each person's action with its time — from the clinical summary writer's first draft to her signature.
The same dossier, seen by the five people who carry it — what their filing months looked like, and what they look like now.
Six specialist agents read the sources, draft each section, open every cited page and check each fact across modules; the section owner applies fixes and signs.
Reads CSRs, integrated summaries and Module 3 documents page by page, and records every key fact — sample sizes, endpoint results, exposure, doses, specification limits, shelf life — with its population and its exact page, table or figure.
Drafts 2.5 Clinical Overview and the 2.7 Clinical Summary in ICH M4E(R2) order and the company template, with a citation to the source page on every statement that states a fact.
Drafts 2.4 Nonclinical Overview and the 2.6 written and tabulated summaries in ICH M4S(R2) order from the Module 4 study reports.
Drafts the 2.3 Quality Overall Summary — drug substance, drug product, appendices — summarising each 3.2.S and 3.2.P section under the matching 2.3 heading.
Opens every cited page and confirms every number, the population and the direction of the claim; marks each statement verified, to check (wording overstates or understates) or disagrees.
Tracks each key fact across Modules 2 to 5. When two sections disagree, or a section disagrees with the source of truth, it opens a conflict with both values, both pages and a proposed tracked change for the section owner.
Applies fixes and signs. The agents propose; a named person decides.
Ask in plain words — or say what to change. Every answer cites the page it came from.
Where else do we state the VLT-302 sample size?
Three places. 2.5.4 and 2.7.6 say 824 randomized, which matches the final CSR. 2.7.3.2 says 818 randomized — that is the full analysis set, not the randomized count. The conflict is open on Facts across modules.
How many patients were exposed to veltrimab?
1,947 patients received at least one dose, for 2,316 patient-years. 2.7.4.1 states this correctly. 2.5.5 says 1,935 — it cites the superseded June data cut — 12 patients apart.
Does response differ by body weight?
Yes. PASI 75 response was seen in every subgroup but was lower above 100 kg: 61.0% versus 75.9%. Population PK explains it — steady-state exposure is 32% lower above 100 kg. 2.5.4 calls the response “consistent” across subgroups; the verifier suggests “seen in all subgroups, with a lower rate above 100 kg”.
What is the protein concentration limit?
The registered release limit is 95 to 105 mg/mL (UV A280), specification v2.1. 2.3.P.5 quotes 100 ± 10 mg/mL — the development limit — so it is flagged for Lena Ortiz with corrected wording ready.
13 screens from the working solution, on its sample data. Pick one to see it large.
Who signs each section, the citation rules every statement is checked against and the house style are settings.
| Setting | Default | Choose from |
|---|---|---|
| Owner of 2.7.3 Summary of Clinical Efficacy | Dr. Maya Chen · signs the section, no extra reviewer | Any section owner in People |
| Every statement needs a source page | On | On · Off — a statement without a citation cannot leave drafting |
| Numbers must match the key facts register | On | On · Off — checked before a section goes to its owner |
| Flag wording that overstates a result | On | On · Off — the citation verifier compares the claim with the page |
| Re-check when a source gets a new version | On | On · Off — affected statements are proposed as tracked changes |
| How small p-values are written | p<0.001 | p<0.001 · p<.001 · exact |
| Decimal places in results text | 1 | 0 · 1 · 2 |
| Module 2 headings and order | ICH M4 + company style guide | Template |
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 dossier · “target” = the design goal, measured in the live solution · “estimated” = our estimate. Module numbers follow the ICH Common Technical Document (M4, M4E(R2), M4Q(R1), M4S(R2)). The product, studies, data and people named on this page are fictional — characters and sample data in the working solution.
Software that helps regulatory and medical writers draft the Module 2 summaries of a Common Technical Document — the 2.3 Quality Overall Summary, the 2.4 and 2.5 overviews and the 2.6 and 2.7 summaries — from Modules 3 to 5. CTD Module 2 Authoring adds six agents that draft each section with a citation on every statement, open every cited page and check each fact across modules, for the section owner to apply and sign.
The consistency agent tracks every fact stated in more than one place — sample sizes, results, doses, limits — against the same fact elsewhere in Module 2 and its source of truth. When a section changes it re-checks that section; a disagreement opens a conflict on the node with both sources linked and a proposed fix.
Yes, by rule: a statement without a citation cannot leave drafting, and the writers take numbers only from the key facts register, each fact recorded with its page. The citation verifier then opens every cited page and checks the number, the population and the wording.
The new version is read and indexed page by page, the old one is marked superseded, and the statements that cite affected facts are proposed as tracked changes for their owners. In the working solution, the interim CSR for the VLT-303 extension added this way flagged 14 statements in 2.7.3.5 and 2.5.4 that may need an update.
No. The citation verifier cannot change text, and the consistency agent's corrections are proposals until the section owner applies them. The clinical summary writer never approves its own draft, and only the named section owner signs.
Yes. Each section has a named owner who applies or rejects every proposed change and signs with an e-signature with meaning. A section cannot be signed while a statement disagrees with its source, and a signed section is locked as a version; a later change opens a new version with a reason.
Yes. Section owners, citation rules, how p-values are written, decimal places in results text and the Module 2 template — ICH M4 plus your company style guide — are settings, and you can add your own citation rules.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your Module 2 template, style guide and a sample of past study reports and summaries — and it goes live once every quality gate has passed. We will walk you through it on your own dossier first.
Every agency change reaches the right team within hours, with what changed marked.
1-min demoHealth Authority Query ResponseCited first drafts within hours, consistent with every past answer to every agency.
1-min demoCCDS Label Change ManagementEvery local label a CCDS change touches found, with every market on a clock
1-min demoWe'll run CTD Module 2 Authoring on a sample of your own study reports and Module 2 drafts.