Find the responder subpopulation, from real-world data, without ever exporting it.
Translational hypotheses and trial designs live or die on who responds. Polnor lets you interrogate real-world clinical data inside your own cloud, on an HDS-certified host, to define a biomarker subpopulation and sharpen your next study, with a pipeline where almost every step runs itself.
Which patients carry the marker that predicts response?
A translational question, answered on real-world data. You segment a condition population by a lab or genomic marker, size the biomarker-positive subpopulation, and read its response signal, without a single row leaving your cloud. Here is the path, and how little of it is manual.
| marker status | n | response rate | enrichment |
|---|---|---|---|
| biomarker-positive | 3,417 | 48.6% | 2.4× |
| biomarker-negative | 18,905 | 20.1% | ref |
Illustrative interface and figures, shown to explain the workflow. Not results from a Polnor customer.
Everything up to the cohort is automated.
By the time you write that query, the hard part is done: the platform has already ingested, de-identified, standardized and quality-checked your data. You start where the science starts.
Real-world data in, at the FHIR standard
EHR extracts, lab feeds and genomic panels are pulled and normalized to FHIR R4 automatically, one typed table per resource. No bespoke ETL per source.
| resource | rows | status |
|---|---|---|
| Observation | 9,884,120 | ✓ auto |
| ConditionOccurrence | 3,760,441 | ✓ auto |
| DiagnosticReport | 1,205,338 | ✓ auto |
GDPR de-identification and OMOP mapping, hands-off
Direct identifiers are masked and dates shifted by preset, then the data is converted to the OMOP CDM with OHDSI concept mapping and clinical quality control. Your biomarker strata are defined on standardized LOINC measurements, not free text.
| source | → OMOP concept | std |
|---|---|---|
| loinc:3037639 | Biomarker assay | S |
| icd10:C50 | Malignant neoplasm | S |
| atc:L01 | Antineoplastic agent | S |
Cohorts, stratification and models, where the science is
This is the part your translational team owns: define the biomarker-positive subpopulation, stratify by response, and, if you need it, train and track a classifier with MLflow. Everything runs on compute next to the data, and nothing is exported.
| model | AUC | endpoint |
|---|---|---|
| responder-classifier | 0.81 | /v1/respond |
| marker-imputation | 0.86 | /v1/marker |
The platform does the plumbing. Your team does the science.
The line is deliberate: the repetitive, error-prone, compliance-heavy work runs itself, so your scientists spend their time on judgment, not preparation.
Findings you can carry into a trial design, with a trail behind them.
Every access to patient data is logged, PHI is classified, and export follows EHDS-aligned manifests. Because the data never leaves your HDS-certified cloud and we never train our models on it, the provenance of your subpopulation is unambiguous.
From raw RWD to a biomarker subpopulation, in your cloud.
Book a 30-minute demo. We run the automated pipeline on a case that looks like yours, and show you exactly where the automation ends and the science begins.