Research – Innovate · ESPA 2021–2027

Digital twins for precision medicine

SynthesizeCare is developing a platform that optimizes diagnostic and therapeutic protocols using 3D synthetic patients: an anatomical digital twin of each individual, generated from clinical imaging.

Explore the project The consortium
Start
28 Nov 2025
Consortium
2 partnersBIOEMTECH · IKNOWHOW
Public grant
€ 799,970EU & Greek national funds

01 — Aim

An anatomical map of every patient

From clinical data to personalized care.

SynthesizeCare is a fully novel approach to building a platform that creates digital twins of patients: a digital anatomical map of each individual. These twins are designed to be used directly in clinical practice and, indirectly, as the foundation for more specialized digital twins for targeted therapies, each one built on the anatomical map of the patient concerned.

In doing so, SynthesizeCare becomes the vehicle that opens the way for digital twins with immediate application in operational clinical systems.

Read the background

02 — Methodology

From a CT series to a synthetic patient

Anonymized CT and MR examinations are segmented organ by organ, a generative model learns to complete the anatomy that a scan does not show, and the result is delivered to clinicians as a whole-body 3D model inside their own workflow.

Illustration in three stages. Left: a wall of clinical CT and MR slices of the head, chest, abdomen and knee. Centre: streams of data from the slices assemble a translucent human body with organs and skeleton. Right: the completed synthetic patient model, a glowing whole-body 3D anatomical twin standing on a platform.
Fig. 1 — Concept: clinical imaging in, anatomical digital twin out.Artist’s impression

  1. Data processing

    • Image preprocessing and standardization
    • Organ segmentation with TotalSegmentator
    • 20 anatomical structures
    • Body outline extraction

  2. Generative AI model

    • Vox2Vox volumetric model
    • 3D reconstruction of missing parts
    • Training and evaluation with MSE, PSNR and SSIM

  3. Model validation

    • Quantitative metrics
    • Visual inspection
    • Clinical evaluation
    • Continuous improvement

  4. Platform & integration

    • User-friendly graphical interface
    • 3D anatomical model
    • Integration in clinical workflows
    • Decision-support capabilities

Methodology in detail

03 — Outcomes

What the project will deliver

  • TRL3 → 6

    SynthesizeCare software platform

    A platform that loads a DICOM study and returns the patient’s anatomical digital twin, demonstrated in a relevant clinical environment.

  • 2datasets

    Clinical and synthetic database

    A curated database of paired clinical images and segmentations, and the digital population of synthetic patients generated from it.

  • 3use cases

    Beta pilot use cases

    Three pilot applications that put the digital twins to work in diagnostic and therapeutic protocols.

  • >100people

    Training and education

    Training activities that introduce more than one hundred healthcare and imaging professionals to digital twins.

  • IP

    Protection of results

    Patent and trademark filings, a business plan and a market analysis that prepare the platform for commercial exploitation.

  • Open

    Publications and conferences

    Peer-reviewed publications and presentations at scientific conferences, listed on the dissemination page as they appear.

  • Personalized medicine
  • Improved clinical decision-making
  • Optimized therapeutic protocols
  • Broader access to advanced healthcare

04 — Work plan

Five work packages

Two partners share the work, from the database of clinical digital twins and the learning algorithms through to clinical evaluation and exploitation.

  1. WP1Database of digital twins built from clinical dataLead: IKNOWHOW
  2. WP2ML/DL algorithms for generating digital patientsLead: BIOEMTECH
  3. WP3Software for creating digital twins for clinical applicationLead: IKNOWHOW
  4. WP4Clinical evaluation of the digital twins and their use in diagnostic and therapeutic protocolsLead: BIOEMTECH
  5. WP5Commercial exploitation of results, training activities and clinical impactLead: BIOEMTECH

The work plan in detail

05 — Consortium

Two Greek companies, one platform

  • BIOEMTECH logo

    Coordinator · leads WP2, WP4, WP5

    BIOEMTECH

    Imaging technologies and research services for preclinical and translational medicine. In SynthesizeCare, BIOEMTECH develops the machine- and deep-learning models, leads the clinical evaluation and drives exploitation.

    bioemtech.com
  • IKNOWHOW logo — Crafting Empowering Technologies

    Partner · leads WP1, WP3

    IKNOWHOW

    Software engineering and automation for the public and private sector. In SynthesizeCare, IKNOWHOW builds the database of clinical digital twins and the software that brings them into clinical use.

    iknowhow.com

Get in touch

Working on digital twins, imaging or clinical AI?

Contact the consortium News & events