Opens in a new tab

Brussels 

17 Nov |

Book your meeting

Predict how the same patient responds across cancer therapies.

One patient.
Multiple possible outcomes.

Patient digital twins: modelling how the same patient may respond to different treatments, using multimodal data.

Same patient, different therapies.

Therapy A
Tumour persists
6-month prediction
Therapy B
Complete response
6-month prediction

Illustrative example. Not CE marked.

Portraits of a drug developer and a patientPortraits are AI-generatedPause animations
SUPPORTED BY

    STARTING WITH BREAST CANCER

    31 authorised active substances.
    One individual treatment decision.

    ESMO guidelines recommend an individual risk–benefit analysis for each patient, which grows more complex with every new approval.

    31

    Active substances authorised in the EU for breast cancer

    Sources: EMA medicines database, 01.11.2026 (Aileen Health count); ESMO guideline, Loibl et al., Ann Oncol 2024.

    Compare digital-twin predictions across therapies.

    How it works

    Scatter plot of digital-twin predicted response for each patient under standard versus alternative therapy

    Retrospective model output on neoadjuvant clinical trial data (I-SPY2); research use only, not CE marked.

    How to read the chart. Each dot is one patient, predicted under two therapy options: standard of care (→) versus alternative (↑).

    Response score: 0–1; 1 indicates predicted complete response.

    Above the diagonal: Higher predicted response with the alternative therapy.

    Shaded region: Preferential responders — low predicted response under standard of care, high under the alternative.

    SCIENTIFIC VALIDATION

    Trained on randomised trials.
    Built to be validated in real‑world practice.

    A patient receives only one therapy, so individual counterfactuals cannot be observed. Cancer4D therefore groups patients by predicted response pattern.
    01

    Learn treatment effects from randomised trials

    Therapy X

    Standard chemotherapy (taxane)

    Therapy Y

    Chemotherapy + immunotherapy (anti-PD-1)
    02

    Group patients by predicted response pattern

    218

    Lower predicted response with both therapies

    252

    Strongest preferential response to Therapy Y

    120

    Higher predicted response with both therapies
    03

    Check observed outcomes in each group

    ■ Therapy X   ■ Therapy Y   · observed pCR rate
    Lower predicted response with both therapies
    X
    9%n = 53 · CI 5–18
    Y
    13%n = 23 · CI 5–29
    Difference Y − X+4 pp · 90% CI −10 to +17
    Strongest preferential response to Therapy Y
    X
    22%n = 64 · CI 15–31
    Y
    74%n = 23 · CI 57–86
    Difference Y − X+52 pp · 90% CI +35 to +69
    Higher predicted response with both therapies
    X
    35%n = 23 · CI 21–52
    Y
    75%n = 12 · CI 51–90
    Difference Y − X+40 pp · 90% CI +14 to +66
    Whole cohort · Therapy X 19% (27/140) · Therapy Y 50% (29/58)
    Difference Y − X+31 pp · 90% CI +19 to +43
    Patients with a lower predicted response to both therapies rarely reach pCR with either. Patients predicted to respond preferentially to Therapy Y show the largest observed benefit: +52 pp, compared with +31 pp across the whole cohort.
    pCR: pathological complete response. n: patients treated with that therapy in the group. Group sizes include patients from all I-SPY2 arms; observed pCR is shown for patients treated with X or Y. CI: 90% confidence interval. pp: percentage points. Retrospective analysis of randomised trial data (I-SPY2); small subgroups. Real-world validation is under way.

    ABOUT US

    Built by people who know AI and oncology care

    Founder-led and part of ABIGAIL4D, an EU Horizon project with clinical and research partners across Europe.

    Overall Impact Award 2026
    Nucleate Germany
    More about the team
    Anna Tidstam
    Anna Tidstam
    Founder & CEO
    The ABIGAIL4D consortium at the project kick-off at Eurecat in Barcelona
    ABIGAIL4D kick-off at Eurecat, Barcelona, 21–22.10.2026

    Our next stops

    Upcoming events

    Join us at one of our next stops.

    1. 17 Nov 202618 Nov 2026

      Brussels

      WKÖ Global Insights: Executive Mission Brussels

    3. Canal houses in Amsterdam

      Past event

      6 Nov 2026

      Amsterdam

      EMA SME Info day

    4. Linz skyline on the Danube

      Past event

      5 Nov 2026

      Linz

      JKU Entrepreneurship Center Opening

    5. Aerial view of Munich old town

      Past event

      4 Nov 2026

      Munich

      HTGF Digital Health Pitch Day 2026

    Book a meeting
    FEATURED IN

    • Get notified when we have more news

    Subscribe to our updates and follow our research progress.

    Newsletter