INSTITUT DE CANCEROLOGIE DE L'OUEST

Their core work

The Institut de Cancérologie de l'Ouest (ICO) is a French comprehensive cancer center combining clinical treatment with applied research. Their H2020 work sits at the intersection of oncology and data science: they have developed and validated machine learning algorithms for predicting breast cancer therapy outcomes, and are now part of a pan-European effort to optimize treatment for solid tumor patients using artificial intelligence. ICO brings real clinical data and oncology domain expertise — not just computational methods — making them a credible bridge between hospital practice and AI research. Their work is grounded in real-world evidence, meaning they work with data from actual patient populations rather than controlled laboratory settings.

What they specialise in

How they've shifted over time

ICO's H2020 trajectory shows a clear progression from narrow, fellowship-driven machine learning work to large-scale, multi-country AI for oncology. Their first project (PredAlgoBC, 2019) had no tagged domain keywords in the system — it was focused and bilateral, with ICO hosting a single MSCA fellow to explore ML prediction for breast cancer. By their second project (OPTIMA, 2021–2026), the framing has shifted to big data, real-world evidence, and AI at the European level, embedded in a large RIA consortium. The direction is unmistakable: from a proof-of-concept ML experiment to contributing clinical data and oncology expertise within a coordinated European AI health initiative.

ICO is moving from hosting individual AI researchers toward being a clinical data and expertise node within large European oncology AI consortia — a trajectory that positions them as a go-to partner for any health AI project needing real cancer patient data and clinical credibility.

How they like to work

ICO's coordinator role was an MSCA Individual Fellowship — meaning they hosted a visiting researcher rather than managing a multi-partner consortium in the traditional sense. Their participant role in OPTIMA places them inside what appears to be a large European consortium (39 unique partners, 14 countries across both projects). This pattern suggests ICO functions as a specialized clinical contributor: they bring patient data, oncology know-how, and validation capacity, rather than driving project administration. For a potential partner, this means ICO is likely a reliable, domain-focused collaborator rather than a project management-heavy lead.

Across just two projects, ICO has connected with 39 unique partners in 14 countries — a figure dominated by OPTIMA's large pan-European consortium structure. Their network is broad in geography but shallow in depth, reflecting one very large collaboration rather than repeated ties to the same partners.

What sets them apart

ICO is one of the few cancer centers in France combining active clinical care with hands-on AI and machine learning research — they are not a university department applying generic ML to health data, but a working hospital center validating algorithms against real patient outcomes. This gives them something most pure tech or academic partners cannot offer: legitimate clinical credibility and access to real-world oncology datasets. For consortia building AI tools for oncology, ICO provides the clinical anchor that makes results publishable and regulatorily relevant.

Highlights from their portfolio

• OPTIMA
  A large ongoing RIA project (2021–2026) on AI-optimized treatment for solid tumors across Europe — ICO's most significant collaboration by scale, connecting them with a pan-European network of 39 partners in 14 countries.
• PredAlgoBC
  ICO's only coordinator role in H2020, hosting an MSCA fellow to develop machine learning algorithms specifically for breast cancer therapy prediction — demonstrating early in-house AI research capacity.