Signal Intelligence · a worked example from free public data
A competitor wrote its acquisition into a public database field.
Not a leak. Not a subscription. A field on ClinicalTrials.gov, filled in by the trial site, readable by anyone who thought to look. We read trial registries, deal disclosures, regulatory actions and patents together, and tell you what changed on your indication before it reaches a journal or a press release.
NCT06691685
ESO-T01 in relapsed/refractory multiple myeloma — Union Hospital, Tongji Medical College
- Asset
- ESO-T01 · BCMA
- Overall status
- SUSPENDED
- Why stopped
- “The collaborator EsoBiotec was acquired by AstraZeneca.”
- Status verified
- 2026-07
Registry record as published, read on 28 September 2026. The commercial cause of the suspension is stated in plain language, in a free public database.
What that one field explains
Every step below can be checked separately against a public source. None of it required privileged access.
- 18 Nov 2024Trial starts. Planned enrolment: about 24 patients.
- Mar–May 2025AstraZeneca announces, then completes, the acquisition of EsoBiotec: $425m upfront, up to $1bn.
- 25 Mar 2026Nature Medicine publishes the dataset: enrolment stopped at five patients; four responses, grade 3+ adverse events in every patient, one death.
- 20 Jul 2026The registry is updated: status SUSPENDED, and the “why stopped” field names the acquisition. Until then the record still read RECRUITING, sixteen months after the deal was announced. The field explains the small cohort; its timing shows why we never rely on one source.
Flagship case: in vivo CAR-T
Three FDA clearances in six weeks, each reported as one company’s news. Read together with the registry, deal and patent layers, they point to a shift in how early offshore data opens US trials.
3
FDA IND clearances between 9 Aug and 16 Sep 2026
$14.4bn
Disclosed deal value (up to, milestones included), six deals since Mar 2025
~40
Patients dosed across all published cohorts
6
Dated forecasts, scored in public
Four public layers. The value is in the joins.
Any one of these is a news alert. Read against each other, on a dated timeline, they describe intent before it becomes a public fact.
Pipeline
Who is dosing patients, where, and which programmes quietly stop.
Deals
What well-informed buyers believe, with a price on it.
Regulatory
Whether the path is open. This is what actually sets the timetable.
Patents
Whether the position is defensible once the science works.
How the work is done
Six steps, published so you can hold us to them. The scan is a commodity; the judgement is the service.
One client, one decision, one horizon. We do not sell open-ended monitoring.
Six to ten sources, spanning at least three kinds: science, money, rules, behaviour.
Keep only what is verifiable, dated, and would surprise an expert in the field.
Two independent sources minimum, or the item is labelled single-source and unconfirmed.
What would have to be true for this to matter to you by a given date, and what would invalidate it.
Timestamp, confidence level, review date. We publish our own hit rate.
Four ways to buy it
Priced per deliverable, never per day. Scope is fixed before work starts.
Specimen
Free · 72 hours
- One indication, signals chosen by us
- Every claim carries its NCT or DOI
- Yours to circulate internally
Free
Pipeline Radar
Registry layer · 48–72 hours
- New entrants, stopped and late-stage trials
- Focused scan or full landscape
- Underlying data as CSV
$500–$1,200 · full landscape $1,500–$3,000
Signal Landscape
All four layers · 15–20 working days
- Full signal report: registry, deals, regulatory, patents
- Complete dataset as CSV
- Counter-signals and dated forecasts with falsification conditions
From $4,000
Signal Watch
Recurring · monthly or quarterly
- Scheduled re-runs on your indication
- Only what changed since last time
- Ad-hoc questions answered in 72 hours
From $900 / month
What this cannot do
A registry status reports what a sponsor last submitted, not what is happening. Absence of an update is not evidence of activity. In the example above, a second record for the same asset still showed RECRUITING with a status-verified date about eighteen months old. We flag that rather than quietly average it away.
Every signal we deliver is a candidate, never a conclusion. A person with domain expertise reviews each one before it reaches you. That review is the billable step, and it is the reason the output is worth reading.
We publish our forecasts with dates, confidence levels and the conditions that would prove them wrong. Then we publish whether we were right, on our track-record page.
Tell us your indication. We will run it and send you the signals, at no cost, within 72 hours.
If what comes back is useful, we will talk about scope. If it is not, you have lost nothing and you keep the analysis.
The button opens your email with a short template: just add your indication. You can also write directly to admin@research.transitionconsultings.com.