The Immune Twin Platform

Decode immunity.
Predict response.

We expand living B, T and NK cells from a single ~3 mL blood draw and keep them functional for 21 days — long enough to test dose, combination and time on a faithful model of one patient's immune system.

01The problem

Primary immune cells don't survive long enough to be tested.

Immunotherapy response is decided by the patient's own immune system — yet the moment those cells leave the body they begin to fail. Within 24 to 48 hours a standard preparation is no longer functional, and NK cells go first. So the industry tests on models that contain no immune compartment at all, and finds out what actually happens in Phase II.

<24h
Functional window of a standard PBMC preparation
>90%
of oncology drugs fail in Phase II or III
$2.6B
Average cost per approved oncology drug
~50%
of patients do not respond to first-line immunotherapy

FDA Oncology Center of Excellence · DiMasi et al., J Health Econ · PBMC viability literature

02The platform

EBViCells™

Our platform takes a routine ~3 mL clinical blood draw and turns it into an Immune Twin — a living, queryable model of one person's immune system that stays functional for at least three weeks.

It is built on three proprietary layers. Each one removes a constraint that has kept patient-specific immune testing out of reach.

Request the technical brief
Layer 01 — Capture

Ultra-low-volume isolation

B, T and NK cells are isolated simultaneously from about 3 mL of blood — a volume compatible with routine care and with immunocompromised patients, where conventional methods need 10 to 30 mL.

Layer 02 — Preservation

Selective 3D immortalisation

Proprietary 3D-matrix niches preferentially expand immortalised immune cells while progressively eliminating the rest, holding multi-lineage function for 21 days or more without losing the patient's genetic and phenotypic signature.

Layer 03 — Readout

Explainable immune analytics

A 30+ gene panel, cytometry and secretome data feed a model that ranks drug response, toxicity and combination synergy per patient — documented for traceability under the EU AI Act.

03The evidence

Three weeks of function, where the field gets one day.

Measured functional viability of multi-lineage immune cells after sampling. Everything to the right of the red band is testing that simply cannot be done on a standard preparation.

EBViCells™ Immune Twin Standard PBMC preparation
0%25%50%75%100% EBViCells™ Standard PBMC 24 h Day 0Day 7Day 14Day 21
View as table
DayEBViCells™Standard PBMC
0100%100%
199%88%
398%9%
797%1%
1495%<1%
2191%<1%

20 samples · patients and healthy donors · clinical-grade material · full readouts on request

04Why EBViome is different

Five capabilities that enable functional testing of patient-derived immune cells.

01

All three lineages, one draw

B, T and NK cells captured and kept functional together — NK biology is the hardest to preserve and the most predictive for cell therapy.

PBMC assays lose NK within hours. Organoids and PDX models contain no immune compartment at all.
02

A clinical blood volume

About 3 mL — a routine draw that fits inside standard care with no additional invasive procedure.

Conventional assays require 10–30 mL. Tissue models require a tumour biopsy.
03

A three-week window

21 days or more of measured multi-lineage function, from the same original sample.

Primary cells fail at 24–48 h. Organoids reach 7–14 days — without immune cells.
04

Longitudinal by design

Dose, schedule, combination and exhaustion studied over time on one patient's own cells.

Single-timepoint assays cannot show kinetics, and kinetics is where immunotherapy fails.
05

An explainable AI layer

A 30+ gene signature paired with functional readouts, built to be auditable rather than a black box.

Purely computational platforms infer response; they never observe it.
Partners & ecosystem

From guesswork to guidance.

Talk to us.

We use your details only to reply. Privacy