Biomarker & Patient Population

Use biomarker results
to shape the next study.

Check whether the biomarker identifies patients who benefit before changing the next study.

Biomarker & Patient Population Strategy workflow overview

Find who benefits
before changing the study.

Compare the marker results, explain what remains uncertain, and let scientific leaders decide what to test next.

Start with: Lab tests, samples, and study results
Prepare: What the marker shows and what to check
Team decides: Scientific leaders choose who to include

How it works.

Reconstruct the marker definition across the study record.

Agents trace assay versions, cutoffs, sample availability, and analysis history, including choices made after results were seen.

Original rationale

Selection memo · 2018

70% response above the threshold.
40% below.

The threshold was chosen after seeing the outcomes.

Lab bridge · 2020

The later test scores the same specimens 20 points higher.

Same threshold, different scales

Assay A ≥60 Assay B ≥80

Test the signal against patient and study evidence.

Agents connect marker results to drug and control outcomes, account for missing samples and measurement differences, and surface other explanations for the pattern.

Study B · week 16 responses

Treatment	Marker +	Marker −
Drug	70% 35 / 50	40% 20 / 50
Control	50% 25 / 50	20% 10 / 50
Drug benefit	+20 points	+20 points

Extra benefit for marker + is unproven.

Difference: 0.0 points
95% interval: −25.7 to +25.7 points

Separate study · rule fixed beforehand

Study C: difference −1.4 points
95% interval: −22.8 to +20.0 points

Prepare an independent validation plan.

Agents lay out the assay, cutoff, evidence gaps, and criteria to test. Scientific leaders decide whether the marker should change the study population.

Study drafts

Clinical Development + Biostatistics

Keep both marker groups.

Before: ~~Marker-positive only~~
After: Include both marker groups
Analysis: Test whether drug benefit differs by marker group.

Study drafts reflect the team's decision.

Use biomarker resultsto shape the next study.

Find who benefitsbefore changing the study.