AI and Oncology - Loick Menvielle: “Will Europe build the regulatory, economic, and organisational conditions needed to move from pilots to full-scale deployment?”

• Find out about the full programme for this event

• Find out more about EDHEC’s Management in Innovative Health Chair

What is the current state of play regarding the application of artificial intelligence in oncology?

Loick Menvielle: We have moved beyond the proof-of-concept stage. The results are now tangible and well-documented: early detection of lesions, support for clinical decision-making, optimisation of care pathways, and a reduction in the administrative burden on healthcare teams… to name but a few examples.

AI has been used in oncology for years and has enabled medicine to shift towards precision medicine, allowing for earlier detection of the disease and, consequently, less invasive procedures. This is a clinical reality, not a promise, and the standard of care is evolving alongside it. However, what is currently lacking in this changing medical landscape is scale.

The question, therefore, is no longer whether these tools work, but under what conditions they can be deployed safely, on a large scale, and across all European healthcare systems. This means addressing issues that go beyond mere technical performance: at what point is a tool clinically mature enough? What evidence should be required before integrating it into protocols? Who revises the guidelines for good practice when the algorithm detects a condition before the clinician does? How should responsibilities be defined and documented when an algorithm is involved in a medical decision? These are questions concerning the organisation of care as much as they are about innovation.

Added to this is a paradox of adoption: technology is advancing faster than the ability of teams to take it on board, whether they be doctors, paramedical staff or administrative staff. The bottleneck is human rather than technological. Finally, trust is a matter of architecture: explicit human oversight of algorithmic decisions, coordinated between the national framework, the AI Act and the European Health Data Space (EHDS), is key to gaining the support of both healthcare professionals and patients.

On 5 October, you are co-organising an event that brings together various partners from different backgrounds. What does this mean?

Loick Menvielle: It is precisely this diversity, in my view, that makes this day so unique. Under the leadership of our Chair in Management in Innovative Health and Unicancer, we are bringing together IRCAD, Ethik-IA and the Ligue contre le cancer. In other words: the academic world, cancer centres, surgical and technological excellence, ethical reflection on artificial intelligence, and the voice of patients.

This diversity is by no means merely a stylistic exercise or a public relations exercise; it stems from the very nature of the subject. One cannot seriously envisage the deployment of AI in oncology without involving those who provide care, those who set the regulations, those who produce the technology and those who are directly affected by it. It is a convergence of public, private and not-for-profit stakeholders that one rarely sees gathered around the same table.

In practical terms, how will the day unfold?

Loick Menvielle: We have opted for a format we call a ‘Strategy Lab’, which is deliberately more demanding than a series of round-table discussions. The day is divided into three stages: first, mapping the obstacles to roll-out in order to identify and prioritise them; second, defining a realistic yet ambitious target for 2030; and finally, formulating concrete and actionable recommendations.

Participants work in working groups led by experts, in French and English, comparing their respective experiences and constraints – those of a cancer centre are not the same as those of an industrial partner, nor those of a researcher.

It is precisely from this methodical comparison that operational recommendations emerge, rather than mere wishful thinking.

The day will conclude with a round-table discussion open to the public. What are your expectations for it?

Loick Menvielle: It is entitled ‘AI and cancer: does Europe have the resources to match its ambitions?’. It is a question worth asking, with no ready-made answer: AI in oncology now has both the evidence and the potential; the question remains as to whether Europe is equipping itself with the regulatory, economic and organisational resources needed to move from pilot schemes to full-scale roll-out.

Training all healthcare professionals by 2030: should we be questioning the financial, evaluation and certification models? How can we avoid a two-tier AI system between pioneering teams and the rest of the hospital? Who is funding - and will continue to fund in the long term - AI in hospitals once the calls for proposals have ended? Is European regulation a competitive advantage or a hindrance when compared with American and Chinese players? Do we need European models, or is European oversight of models trained elsewhere sufficient? Europe has real strengths, whether in terms of sovereign access to data, clinical trial networks or its influence through training, but it also runs the risk of becoming dependent on models trained elsewhere.

What makes this session so enriching is the diversity of perspectives brought together: Sophie Beaupère (Unicancer) on training and AI governance in cancer centres, Pauline Choné (Ministry of Health) on the implementation of the 2025–2028 national strategy on AI and health data, David Gruson (Ethik-IA) on human safeguards and regulation as a driver of European competitiveness, and Antonello Forgione (IRCAD) on AI in surgery, from early detection to de-escalation of surgical procedures. I myself will speak on demonstrating medical and economic value, which is a prerequisite for moving from pilot projects to wider roll-out: whilst pilot projects are funded through innovation, roll-out requires a sustainable funding model, which the 2025–2028 national strategy is beginning to outline. Sébastien Marguerès (AI for Health & Artefact Adopt AI) will moderate this exchange of deliberately contrasting perspectives rather than a consensus-based panel. The round table, which is open to participants beyond the Strategy Lab alone, will be followed by a social gathering to continue these discussions.

You mention the publication of a white paper. What is its aim?

Loick Menvielle: All the work produced by the groups - the obstacles identified, the priorities established and the recommendations made - will form the basis of a white paper which we plan to publish in early 2027. Our aim is not simply to produce yet another document to gather dust on the shelves, but to create a tool that cancer centres, industry stakeholders and public decision-makers can actually use to develop their roadmaps leading up to 2030.

To achieve this, we are doing what we do best: working, bringing people together and taking a step back. This is a subject that warrants such an approach, as it will shape European oncology for the coming decade, and I believe we have a collective responsibility to set out the terms of the debate clearly before rushing to find solutions.