StemPredict

IG20140

Predictive AI imaging for stem cell and organoids

Main Description

Human Pluripotent Stem Cells (hPSCs), organoids, and advanced bioimaging are transforming human in vitro model systems in biomedical research. These technologies are anticipated to form the cornerstone of effective and safe preclinical trials and future cellular therapies. However, their full potential remains constrained by technical limitations.

STEMPREDICT is a groundbreaking initiative addressing a critical challenge in biomedical research: efficiently and non-invasively assessing the phenotype, health, and quality of pluripotent stem cells, differentiated cells, and organoid systems. These cell-based models are essential for advancing therapies and deepening our understanding of diseases. However, traditional monitoring techniques are often slow, labour-intensive, and unsuitable for large-scale experimental setups.

By integrating imaging technologies with artificial intelligence (AI), STEMPREDICT aims to develop tools capable of real-time cell analysis. These tools will extract valuable phenotypic information by analysing label free imaging data eliminating the need for destructive testing methods and enhancing the potential to further automate cell production.

This innovative approach will enhance the optimization of stem cell production and improve predictions about their suitability as disease or therapeutic models, delivering unprecedented efficiency.
By benefiting from high-level experts and data from EU Core Facilities, the platform developed within STEMPREDICT ensures wide adoption, fast implementation, and broad accessibility across Europe and beyond. It strengthens European research infrastructures and establishes global standards for quality control in stem cell-based research.

Action keywords

  • AI-powered imaging
  • Label-free phase-contrast microscopy
  • Pluripotent stem cells
  • European core facilities

StemPredict

STEMPREDICT Objectives

Aim – Primary objective

  • Develop affordable, label-free phase-contrast imaging with on-board machine-learning algorithms to deliver real-time, non-invasive predictions of the state and evolution of pluripotent stem-cell, differentiated-cell and organoid cultures. By standardising quality-control workflows across Europe’s core facilities, the tool will boost reproducibility, cut costs and accelerate translation into therapeutics

Secondary objectives

  • Build and validate a proof-of-concept AI pipeline, establishing reference image baselines for pluripotency, monolayer and organoid differentiation; train and test the algorithms across multiple labs to demonstrate usability, accuracy and scalability of the predictive workflow.
  • Draft a preliminary business and exploitation plan that defines the value proposition, business and financial models and marketing strategy, paving the way for commercialisation and broad adoption of STEMPREDICT.
  • Assemble a multidisciplinary team of stem-cell biologists, imaging technologists and AI/data-science experts to design FAIR data-collection protocols, interoperable analysis pipelines and benchmark datasets that underpin predictive imaging of hundreds of cell lines.

StemPredict

CIG Objectives

Assemble a multidisciplinary expert team covering stem cell biology, imaging technologies and artificial intelligence to address the challenges of stem cell research and predictive imaging by:

  • Define optimal strategies for data collection and protocols management: ensuring maximum reusability and transparency (FAIR principles).
  • Evaluate and optimize data analysis workflows and AI pipelines to ensure their interoperability and applicability across different systems and contexts.
  • Platform usability and proof-of-concept: validate applicability across different systems and contexts.

Create an innovation task force for the generation of a preliminary business plan. Draft a plan to outline the:

  • Business concept, target market and voice-of-customers to align our device with the needs of the field.
  • A business model and financial plan.
  • A marketing plan.

StemPredict

Work Plan, tasks and deliverables

Working Group
Tasks
WG1: Business and exploitation
Define a strong value proposition
Analyse the market and competitors’ market strategies
Define tech stack and community building by engaging initial beta-testers and “catapulters”
Evaluate business and financial plan models
Develop a marketing plan with monetization-ready features and strategy
Build an ecosystem and pre-commercialization strategy (brand, early adopters plan)
Set a commercialization strategy and governance model
Define a final go-to-market strategy aligned with technology development milestones
WG2: AI powered tool for stem cell field PoC SmartCam and next smart protocols
Establish reference image baseline parameters for pluripotent state and differentiation protocols
Define and develop AI algorithms, training plan and interface
Create data management plan using FAIR practices
Smart camera setup and associated training strategy
Define initial PoC, validation, and smart protocols plan for next steps
Deliverables
WG1
  • 1.1. Business plan

WG2
  • 2.1. Stem Cells
    • 2.1.1. Proof of concept for pluripotency detection;
    • 2.1.2. Training report
  • 2.2. Bioimaging and AI.
    • 2.2.1. Data management plan.