Research-stage platform · Patent pending

Could the drug itself reveal who is more likely to respond—before trial enrollment?

We test how immune cells from a patient's blood respond to a selected drug outside the body. FTIR captures a broad molecular response pattern. Our AI analyzes the drug-versus-control difference and tests whether specific response patterns are associated with real treatment outcomes. The goal is to identify patient groups that may be more likely to benefit.

01

The drug becomes the probe

02

Broad FTIR response measurement

03

Outcome-trained AI analysis

01The problem

Trials can fail on the population—not the molecule.

Patients can respond very differently to the same treatment. When likely responders are hidden inside a mixed trial population, a meaningful treatment effect can be difficult to detect.

Drug developers need better ways to form and test responder hypotheses before or during clinical development.

Mixed trial populations

A benefit in a subgroup can be diluted by patients who do not respond.

Late signal

Response is often only visible after months of treatment.

Single-marker testing

One predefined marker is not a direct test of how a patient's sample reacts to the drug.

02What Tartei adds

A new measurement and analysis layer on an established approach.

Testing patient-derived immune cells with drugs is already an established research approach. Tartei adds a new measurement and analysis layer designed to reveal more of the response.

Instead of selecting one biomarker in advance, we measure the broad molecular change caused by the drug and use AI to analyze how many spectral changes behave together.

Layer

The drug as the probe

The patient's immune cells are exposed to the selected therapy and compared with matched control conditions. Using matched samples from the same patient helps control for baseline differences between donors.

Layer

Broad FTIR response measurement

FTIR measures a broad molecular response pattern in the cells and culture-derived material. This may reveal changes that could be missed when an assay examines only one predefined marker.

Core of the platform

Outcome-trained AI analysis

Instead of examining one predefined biomarker, our AI analyzes many spectral changes together. It tests which drug-induced response patterns are associated with better clinical outcomes.

03 — Technical workflow

From a patient sample to a responder hypothesis

Patient sample → Drug versus control → FTIR response → AI + clinical outcomes → Potential responder groups

  1. 01

    Patient blood sample

    Immune cells are isolated from the patient's blood.

    Peripheral blood mononuclear cells, or PBMCs, are immune cells isolated from blood.

  2. 02

    Drug and control

    The sample is divided into matched conditions. One receives the selected drug and the other serves as the control.

  3. 03

    FTIR measurement

    FTIR measures the broad molecular response pattern produced under each condition.

  4. 04

    AI analysis

    Our AI analyzes the complete drug-versus-control spectral difference and compares the response patterns with patients' actual treatment outcomes.

    Clinical outcomes enter here as an independent input.

  5. 05

    Potential responder groups

    The system looks for shared response patterns among patients who benefited from the treatment, creating a responder hypothesis for further testing.

What is measured

The difference between the drug-treated and control parts of the same patient sample. Each patient acts as their own reference, which reduces differences between donors.

Intended output

A responder hypothesis: which response patterns occur in patients who benefited, and whether they could be tested in a future study.

04Potential value

Potential value for drug development

The intended use is decision support before and during clinical development. These potential applications would need to be evaluated in blinded, outcome-linked studies.

Trial enrichment

Test whether the response signal can help identify patients more likely to benefit.

Earlier development decisions

Add patient-derived response evidence to decisions about a therapy or development program.

Patient-group selection

Compare response patterns across different patient groups and disease subtypes.

Biomarker research

Generate responder hypotheses that can be tested in future studies.

05Current status

An established patient-sample foundation. Promising early drug-response results.

  1. 01

    More than 100 patient samples analyzed

    At Sheba Medical Center, the team established the FTIR workflow and analyzed samples from more than 100 patients with rheumatoid arthritis and related conditions as part of diagnostic and differential-diagnosis research.

  2. 02

    Initial drug-response work

    Building on this foundation, the team has begun ex vivo drug-response experiments using matched drug and control samples. The initial results are promising.

  3. 03

    The next study

    The next step is a blinded, outcome-linked study designed to test whether drug-induced spectral patterns can support responder stratification.

06 — Partnership

Let's test one therapy together.

We are looking for a pharmaceutical partner for a focused study involving one selected therapy and one well-defined patient group.

In a 30-minute meeting, we can discuss the therapy, available patient samples and clinical data, and whether the approach could fit your development program.

One selected therapy

One well-defined patient group

Matched drug and control testing

A blinded analysis and a written assessment

What the partner brings

A drug of interest and access to patient samples with documented treatment outcomes, with the appropriate consent and approvals.

What we bring

The drug-exposure protocol, the FTIR measurement and the AI analysis linking response patterns to outcomes.

What we produce jointly

A blinded analysis and a written assessment of whether a responder-group hypothesis is supported.

Interested in testing the approach with one of your therapies?

Request a 30-minute meeting

07Contact

Request a meeting

Interested in exploring the approach for one of your therapies? Send us a short message to request a 30-minute discussion. A general inquiry is also welcome if you do not yet have a specific therapy or patient group in mind.

Tell us briefly about your therapy, disease area or development question.

Please do not submit confidential information or patient-identifiable data.